Merge remote-tracking branch 'origin/master' into codex/truncated-design

# Conflicts:
#	docs/event-producer-consumer.md
#	docs/module-graph.md
#	docs/rfc/INDEX.md
#	knip.json
#	packages/README.md
#	scripts/gen-module-graph.ts
#	tsconfig.base.json
#	tsconfig.build.json
#	tsconfig.json

packages/README.md

@@ -11,12 +11,15 @@ Packages are grouped by modular role at `packages/<group>/<pkg>/`. The group dir
 | [`core/`](core/README.md) | Product API spine: session, system-prompt, tools, agent, and the concrete loop | Product — stable surface |
 | [`llm/`](llm/README.md) | LLM capability family: the abstract service + provider adapters | Product — stable surface |
 | [`bash/`](bash/README.md) | Bash capability family: the executor seam, a local impl, and the model-facing tool | Product — stable surface |
+| [`code-runtime/`](code-runtime/README.md) | Code-execution capability family: the abstract runtime seam for model-written programs + a worker-thread backend | Product — stable surface |
 | [`fs/`](fs/README.md) | Filesystem capability family: the abstract seam, a local impl, and the model-facing file tools | Product — stable surface |
 | [`compact/`](compact/README.md) | Compaction capability family: the abstract seam + a basic backend (tool deferred) | Product — stable surface |
 | [`subagent/`](subagent/README.md) | Subagent capability family: the provider-registry seam and the model-facing delegation tool | Product — stable surface |
 | [`web/`](web/README.md) | Web capability family: the abstract seam, search/fetch provider impls, and the model-facing web tools | Product — stable surface |
 | [`spill/`](spill/README.md) | Spill capability family: the storage seam, a local impl, and the tool-result spill policy | Product — stable surface |
 | [`todo/`](todo/README.md) | Todo/planning family: the model-facing `todo_write` tool (whole-list task tracking on the session log) | Product — stable surface |
+| [`timeout/`](timeout/README.md) | Tool-call timeout policy: the `tools/execute` deadline enforcer | Product — stable surface |
+| [`guard/`](guard/README.md) | Loop-hygiene guards: advisory repeat-call reminders | Product — stable surface |
 | [`hooks/`](hooks/README.md) | Hook bridges + the shared Claude Code / Codex wire-protocol library | Product — stable surface |
 | [`session-persistence/`](session-persistence/README.md) | Persistence capability family: the seam + JSONL/SQLite backends | Product — stable surface |
 | [`ui/`](ui/README.md) | Editor/client integration surfaces (the ACP bridge) + the app packages | Product — stable surface |
@@ -29,6 +32,6 @@ The split is the point: a package's group says whether it is part of the product
 
 The inter-package dependency graph is generated: [docs/module-graph.md](../docs/module-graph.md) (`pnpm run gen-module-graph`, freshness-gated in CI).
 
-The rule it must obey: **extension plugins depend on interfaces, never on the concrete loop.** `dsh-agent-loop` is swappable — UI/hook/tool plugins keep working against the `dsh-agent` vocabulary if the loop is replaced. The sanctioned exception is a **composition/bundle** package like `dsh-agent-core`, whose whole job is to assemble the concrete spine: it depends on `dsh-agent-loop` (and the other concrete spine plugins) on purpose. The rule constrains plugins that EXTEND the system, not the bundle that COMPOSES it — swapping the loop means shipping a different bundle, not rewiring every extension. A swappable capability splits into interface / implementation / consumer packages (the bash trio is the template — see [capability seams](../docs/rfc/implemented/architecture/2026-06-13-capability-seams.md)).
+The rule it must obey: **extension plugins depend on interfaces, never on the concrete loop.** `dsh-agent-loop` is swappable — UI/hook/tool plugins keep working against the `dsh-agent` vocabulary if the loop is replaced. The sanctioned exception is a **composition/bundle** package like `dsh-agent-core`, whose whole job is to assemble the concrete spine: it depends on `dsh-agent-loop` (and the other concrete spine plugins) on purpose. The rule constrains plugins that EXTEND the system, not the bundle that COMPOSES it. A swappable capability splits into interface / implementation / consumer packages (the bash trio is the template — see [capability seams](../docs/rfc/implemented/architecture/2026-06-13-capability-seams.md)).
 
 Each package has its own `README.md` with purpose, service API, events, extension points, and deliberate non-goals (TODOs).

packages/bash/bash-local/src/run.ts

@@ -62,6 +62,8 @@ export const SENSITIVE_ENV_PATTERN = /KEY|SECRET|TOKEN/i
  * by in-process plugins (the hooks bridges), not the model — `dsh-tool-bash`
  * builds its request from named fields only and does not forward model input
  * here (see its README, § "The tool builds its request from named args only").
+ * @param extra - caller-supplied entries merged last; an explicit entry wins even against the scrub and the overrides.
+ * @returns the environment to hand to `spawn` for the child process.
  */
 export function childEnv(extra?: Record<string, string>): NodeJS.ProcessEnv {
   const env: NodeJS.ProcessEnv = {}
@@ -160,6 +162,14 @@ export class OutputCollector {
     private readonly spillDir: string,
   ) {}
 
+  /**
+   * Ingest one stream chunk, counting it toward the whole-stream total. On
+   * first overflow of the in-memory cap a spill file is opened and every chunk
+   * (already-collected ones included) is appended there from then on; the
+   * in-memory tail then drops whole chunks from its head (or the head of a
+   * single over-cap chunk) until it fits the cap again.
+   * @param chunk - the raw bytes from one stream 'data' event.
+   */
   push(chunk: Buffer): void {
     this.total += chunk.length
     const overflows = this.bytes + chunk.length > this.maxBytes
@@ -203,7 +213,10 @@ export class OutputCollector {
   // the bottom of this file) and `totalBytes` is read only by a test. The live
   // background-poll path goes through `readFrom()`, so inline snapshot() into
   // finalize() and drop or privatize the totalBytes getter.
-  /** Read the collected tail without finalizing (the final-result snapshot). */
+  /**
+   * Read the collected tail without finalizing (the final-result snapshot).
+   * @returns the retained tail text, the truncation flag, and the spill path when one was created.
+   */
   snapshot(): CollectedOutput {
     return {
       text: Buffer.concat(this.chunks).toString('utf8'),
@@ -222,6 +235,8 @@ export class OutputCollector {
    * pushed since `fromByte`. When `fromByte` has already slid out of the
    * in-memory tail window, the read is `lossy` — it returns the whole
    * retained tail and the gap is only recoverable from the spill file.
+   * @param fromByte - whole-stream offset to resume from (a prior read's `nextOffset`; 0 for the first read).
+   * @returns the delta text, the offset for the next read, the `lossy` flag, and the spill path when one was created.
    */
   readFrom(fromByte: number): { text: string; nextOffset: number; lossy: boolean; spillPath?: string } {
     const windowStart = this.total - this.bytes
@@ -236,7 +251,12 @@ export class OutputCollector {
     }
   }
 
-  /** Close the spill file (if any) and return the final output. */
+  /**
+   * Close the spill file (if any) and return the final output. A failed close
+   * (delayed writeback fault) stops advertising the spill path — the file may
+   * be missing its tail — but still returns the in-memory result.
+   * @returns the final collected output: tail text, truncation flag, and the spill path when intact.
+   */
   finalize(): CollectedOutput {
     if (this.spillFd !== undefined) {
       try {
@@ -262,6 +282,8 @@ export class OutputCollector {
  * host process — a kill that cannot be delivered is reported by the process
  * NOT dying, which callers already handle via escalation/timeouts. No-op for
  * non-positive pids (spawn never started a process).
+ * @param pid - the group leader's pid; non-positive means the spawn failed and the call is a no-op.
+ * @param sig - the signal to deliver to the whole group.
  */
 export function killGroup(pid: number, sig: NodeJS.Signals): void {
   if (pid <= 0) return
@@ -303,6 +325,9 @@ export interface RunningBash {
  * exec sessions addressable via session ids + stdin writes. We deliberately
  * spawn a fresh non-login `bash -c` per call for determinism (no rc files,
  * no inherited shell state); revisit when real workflows demand it.
+ * @param spec - the fully-resolved run (command, cwd, limits); no defaulting happens here.
+ * @param internals - test-only knobs; omitted fields fall back to the private per-process spill dir.
+ * @returns the live handle: pid, the two live collectors, the outcome promise, and `kill()`.
  */
 export function runBash(spec: SpawnSpec, internals: RunInternals = {}): RunningBash {
   const spillDir = internals.spillDir ?? privateSpillDir()

packages/bash/bash-local/tests/executor.spec.ts

@@ -90,8 +90,8 @@ describe('LocalBashExecutor.run', () => {
 
   it('kill escalation uses the configured graceMs (a TERM-trapping task dies by SIGKILL)', async () => {
     const { bash } = await setup() // setup pins graceMs: 200 via config
-    const task = bash.start(bash.resolve({ command: 'trap \'\' TERM; sleep 60' }))
-    await new Promise(resolve => setTimeout(resolve, 100))
+    const task = bash.start(bash.resolve({ command: 'trap \'\' TERM; echo ready; while :; do sleep 60 & wait $!; done' }))
+    await readUntil(bash, task.id, 'ready\n')
     bash.kill(task.id)
     await task.done
     expect(task.signal).toBe('SIGKILL')

packages/bash/bash-local/tests/run.spec.ts

@@ -55,6 +55,20 @@ async function waitForStdout(running: RunningBash, expected: string, timeoutMs =
   throw new Error(`stdout did not include ${JSON.stringify(expected)} after ${timeoutMs}ms`)
 }
 
+async function waitForPidFile(path: string, timeoutMs = 5_000): Promise<number> {
+  const deadline = Date.now() + timeoutMs
+  while (Date.now() < deadline) {
+    try {
+      const pid = Number(readFileSync(path, 'utf8').trim())
+      if (Number.isSafeInteger(pid) && pid > 0) return pid
+    } catch {
+      // The child shell has not written the pid file yet.
+    }
+    await new Promise(resolve => setTimeout(resolve, 20))
+  }
+  throw new Error(`pid file ${path} was not written after ${timeoutMs}ms`)
+}
+
 describe('runBash', () => {
   it('captures stdout on success', async () => {
     const result = await runBash(spec('echo hello')).done
@@ -109,7 +123,7 @@ describe('runBash', () => {
   })
 
   it('escalates to SIGKILL when SIGTERM is trapped', async () => {
-    const running = runBash(spec('trap \'\' TERM; echo ready; sleep 60', { graceMs: 200 }))
+    const running = runBash(spec('trap \'\' TERM; echo ready; while :; do sleep 60 & wait $!; done', { graceMs: 200 }))
     await waitForStdout(running, 'ready\n')
     running.kill()
     const result = await running.done
@@ -121,8 +135,7 @@ describe('runBash', () => {
     // group must take the sleep down with bash.
     const pidFile = join(spillDir, `grandchild-${Date.now()}.pid`)
     const running = runBash(spec(`sleep 60 & echo $! > ${pidFile}; wait`))
-    await new Promise(resolve => setTimeout(resolve, 300))
-    const grandchild = Number(readFileSync(pidFile, 'utf8').trim())
+    const grandchild = await waitForPidFile(pidFile)
     expect(grandchild).toBeGreaterThan(0)
 
     running.kill()

packages/bash/bash/src/types.ts

@@ -11,7 +11,11 @@ import type { Branded } from '@deepseek-ai/dsh-brand'
 /** Identifies one background task within an executor (generated `bash-N`). */
 export type BashTaskId = Branded<'BashTaskId'>
 
-/** Brand a string as a {@link BashTaskId}. */
+/**
+ * Brand a string as a {@link BashTaskId}.
+ * @param id - the raw task-id string (the executor generates `bash-N`).
+ * @returns the same string, branded; no validation is performed.
+ */
 export function BashTaskId(id: string): BashTaskId {
   return id as BashTaskId
 }
@@ -26,7 +30,12 @@ export function BashTaskId(id: string): BashTaskId {
  */
 export type OwnerToken = Branded<'OwnerToken'>
 
-/** Brand a string as an {@link OwnerToken}. */
+/**
+ * Brand a string as an {@link OwnerToken}. Only the consuming boundary
+ * (`dsh-tool-bash`) should cast its own id vocabulary in — see the type's doc.
+ * @param id - the consumer's raw owner identity (the tool layer passes the owning agent's session id).
+ * @returns the same string, branded; no validation is performed.
+ */
 export function OwnerToken(id: string): OwnerToken {
   return id as OwnerToken
 }

packages/bash/tool-bash/src/index.ts

@@ -99,6 +99,8 @@ function streamText(output: CollectedOutput): string {
  * stderr section, then exit-status markers. Non-zero exits are REPORTED, not
  * errored — the model decides how to react; only infrastructure failures
  * (spawn errors, aborts) surface as isError results.
+ * @param result - the completed foreground run from the executor.
+ * @returns the model-facing text: output body (or `(no output)`), then any timeout/signal/exit markers, each on its own line.
  */
 export function renderResult(result: BashRunResult): string {
   const out = streamText(result.stdout)

packages/code-runtime/README.md

@@ -0,0 +1,10 @@
+# code-runtime/ — code-execution capability family
+
+The code-execution capability seam (see [capability seams](../../docs/rfc/implemented/architecture/2026-06-13-capability-seams.md)): an abstract runtime interface for executing one model-written program against host-provided async bindings, capturing what it printed and returned. The consumer is the tool registry's Code Mode, specified alongside the seam in the [Code Mode RFC](../../docs/rfc/proposed/feature/2026-06-15-code-mode.md). **Product** packages.
+
+| Package | Role | ctx key |
+|---|---|---|
+| `code-runtime/` | Abstract code-execution seam (interface + vocabulary) | `ctx.codeRuntime` |
+| [`code-runtime-worker/`](code-runtime-worker/README.md) | Worker-thread backend: fresh worker per run, TypeScript via host-side type-strip (annotations advisory, never type-checked), port-bridged bindings, budget/heap containment | registers `ctx.codeRuntime` |
+
+The interface lives at `code-runtime/code-runtime/`; the shipped backend at `code-runtime/code-runtime-worker/`. Backends differ by execution substrate (worker thread, process, container) and by source language — both readonly descriptors on the service — and register `ctx.codeRuntime` without touching the interface or its consumer; that split is what makes a hardened backend a drop-in later.

packages/code-runtime/code-runtime-worker/README.md

@@ -0,0 +1,32 @@
+# @deepseek-ai/dsh-code-runtime-worker
+
+Worker-thread implementation of the [`@deepseek-ai/dsh-code-runtime`](../code-runtime/README.md) seam: `WorkerCodeRuntime` runs each program in ONE fresh Node `worker_threads.Worker` — TypeScript in, type-stripped host-side, bindings bridged over the message port, `{ value, logs, error? }` out. **Containment, not a security boundary**: trust posture is bash-equivalent by design (the [Code Mode RFC](../../../docs/rfc/proposed/feature/2026-06-15-code-mode.md) § Trust posture), with containment bash does not have — separate isolate, empty environment, heap cap, hard termination.
+
+## Config
+
+```yaml
+- id: code-runtime
+  name: '@deepseek-ai/dsh-code-runtime-worker'
+  config:
+    computeMs: 60000              # busy-time budget (measured event-loop active time)
+    maxWallMs: 600000             # wall-clock ceiling; never pauses for anything
+    maxLogBytes: 65536            # shared byte budget for captured log text
+    maxValueBytes: 32768          # rendered-completion-value cap
+    maxOldGenerationSizeMb: 512   # worker heap cap (resourceLimits)
+```
+
+Every field is validated (positive numbers) and defaulted; there are no other tunables.
+
+## Design
+
+- **One fresh worker per run, no pooling** — a program's world dies with its worker: no cross-run state to log, state bleed unrepresentable, runs reconstructable from the session log alone.
+- **Type-strip host-side, in execution context** — the program is wrapped in an async-function shell, stripped with `node:module`'s `stripTypeScriptTypes` (erasable syntax only — `enum`/namespaces are rejected as a program `exception` and no worker spawns), and sliced back out byte-positioned; it then executes as the body of an `AsyncFunction`, so top-level `await`/`return` work.
+- **The port assumes a hostile peer** — model code can reach `parentPort` and forge traffic, so every inbound message is shape-validated and REBUILT before anything reads it (`null`, primitives, junk types, and malformed payloads drop without a throw; forged extra fields never ride along), the host answers each call id at most once, resolves binding names as OWN properties only (a forged `constructor` cannot walk a prototype chain), drops post-settlement replies, and converts a non-cloneable binding resolution into an error reply. Forged `log`/`done` messages cannot bypass the caps: one host-side ledger bounds everything that lands in `logs`, and the completion value is re-capped host-side. Worker-side namespaces are null-prototype with `defineProperty`, so `__proto__`-shaped binding names are ordinary keys.
+- **Two independent budgets, because the peer is hostile** — `computeMs` meters the worker's MEASURED busy time (`worker.performance.eventLoopUtilization()` polling): a hot loop cannot hide behind a pending decoy dispatch, and a program awaiting a slow tool accrues nothing. `maxWallMs` backstops what busy time cannot see (awaiting a promise nobody resolves). Both funnel into `worker.terminate()`, which ends hot synchronous loops too; heap overflow surfaces as the worker's OOM exit (`kind: 'worker-exit'`).
+- **Logs stream eagerly** — console/stdout/stderr entries cross the port as they happen, so a timed-out or killed program still shows what it printed. ONE shared `maxLogBytes` ledger bounds everything: streamed entries, forged port traffic, and pipe bytes that bypass the patched streams (appended after), with the overflow marked in-band once.
+- **Empty environment** — the worker gets `env: {}` and `execArgv: []`: no ambient credentials (stronger than the scrubbed-env rule for spawned commands) and no inherited loader flags.
+- **Dispose to quiescence** — teardown fails in-flight runs as `abort` and AWAITS each worker's exit before resolving.
+
+## The worker entry, unbuilt and built
+
+`worker.ts` is deliberately erasable-only TypeScript with type-only cross-package imports: unbuilt (vitest/tsx), the host spawns `src/worker.ts` directly and Node's native type stripping loads it; built, the entry ships as the sibling bundle `lib/worker.js` (its own tsdown entry). The built path is pinned by `tests/built-lib.e2e.ts`, the real-load-path guard from [docs/testing.md](../../../docs/testing.md).

packages/code-runtime/code-runtime-worker/package.json

@@ -0,0 +1,36 @@
+{
+  "name": "@deepseek-ai/dsh-code-runtime-worker",
+  "description": "Worker-thread implementation of the DeepSeek Harness code-execution seam",
+  "version": "0.0.1",
+  "private": true,
+  "type": "module",
+  "main": "lib/index.js",
+  "types": "lib/types/index.d.ts",
+  "exports": {
+    ".": {
+      "types": "./lib/types/index.d.ts",
+      "default": "./lib/index.js"
+    },
+    "./src/*": "./src/*",
+    "./package.json": "./package.json"
+  },
+  "files": [
+    "lib/index.js",
+    "lib/worker.js",
+    "lib/types/**/*.d.ts",
+    "lib/types/**/*.d.ts.map",
+    "src"
+  ],
+  "license": "BSD-3-Clause",
+  "peerDependencies": {
+    "@deepseek-ai/dsh-code-runtime": "^0.0.1",
+    "cordis": "^4.0.0-rc.6"
+  },
+  "dependencies": {
+    "schemastery": "^3.18.0"
+  },
+  "devDependencies": {
+    "@deepseek-ai/dsh-code-runtime": "workspace:^",
+    "cordis": "^4.0.0-rc.6"
+  }
+}

packages/code-runtime/code-runtime-worker/src/bootstrap.ts

@@ -0,0 +1,301 @@
+/**
+ * Worker-side execution logic, written as plain functions over an injected
+ * port so the unit suite can run every line IN-PROCESS against a fake port
+ * (a real worker thread is a separate V8 isolate the coverage provider
+ * cannot observe). The real worker entry (`worker.ts`) is a thin
+ * self-executing glue file over {@link runWorkerMain}, excluded from
+ * coverage the same way `bin.ts` entrypoints are, and exercised end-to-end
+ * by the integration tests that spawn real workers.
+ *
+ * @module @deepseek-ai/dsh-code-runtime-worker/src/bootstrap
+ */
+
+import { inspect } from 'node:util'
+import { serialize } from 'node:v8'
+import type { CodeLogEntry } from '@deepseek-ai/dsh-code-runtime'
+import { logTruncationMarker } from './protocol.ts'
+import type { DoneMessage, ReplyMessage, WorkerBootData, WorkerToHost } from './protocol.ts'
+
+/** The port surface the bootstrap needs — satisfied by `parentPort` and by the tests' fake. */
+export interface BootstrapPort {
+  postMessage(message: WorkerToHost): void
+  on(event: 'message', listener: (message: ReplyMessage) => void): void
+}
+
+/**
+ * A writable stream's `write` slot, as the bootstrap patches it (see
+ * {@link captureStreamWrites}). Method-typed so the real
+ * `process.stdout`/`process.stderr` (narrower chunk parameters) remain
+ * assignable.
+ */
+export interface PatchableStream {
+  write(chunk: unknown, ...rest: unknown[]): boolean
+}
+
+/**
+ * Ordered log capture under one shared byte budget, delivered to a sink as
+ * each entry lands (the real sink streams entries over the port eagerly, so
+ * captured output survives a mid-run termination). Once the budget is
+ * exhausted it emits exactly one in-band marker entry (on the `stderr`
+ * diagnostics channel) and silently drops everything after — the cap is a
+ * blast-radius bound, so "how much was lost" intentionally stays unmeasured.
+ */
+export class LogBuffer {
+  private remaining: number
+  private truncated = false
+  // Explicit fields, not constructor parameter properties: this module loads
+  // under Node's native strip-only mode, which rejects non-erasable syntax —
+  // and parameter properties are non-erasable.
+  private readonly maxBytes: number
+  private readonly sink: (entry: CodeLogEntry) => void
+
+  constructor(maxBytes: number, sink: (entry: CodeLogEntry) => void) {
+    this.maxBytes = maxBytes
+    this.sink = sink
+    this.remaining = maxBytes
+  }
+
+  /**
+   * Emit one entry to the sink, charging its text against the budget (drops + marks once exhausted).
+   * @param entry - the log entry to deliver.
+   */
+  push(entry: CodeLogEntry): void {
+    if (this.truncated) return
+    const cost = Buffer.byteLength(entry.text, 'utf8')
+    if (cost > this.remaining) {
+      this.truncated = true
+      this.sink({ source: 'stderr', text: logTruncationMarker(this.maxBytes) })
+      return
+    }
+    this.remaining -= cost
+    this.sink(entry)
+  }
+}
+
+/** The five console methods the shim captures, in the seam's level vocabulary. */
+const CONSOLE_LEVELS = ['log', 'info', 'warn', 'error', 'debug'] as const
+
+/**
+ * A `console` replacement whose five leveled methods render their arguments
+ * `util.inspect`-style (matching real console formatting closely enough for
+ * a model to recognize its own output) into the buffer. Only these five
+ * exist — the program gets a deliberately small console, not Node's full
+ * surface.
+ * @param logs - the buffer every rendered line is pushed into.
+ * @returns the five-method console object handed to the program.
+ */
+export function makeConsoleShim(logs: LogBuffer): Record<(typeof CONSOLE_LEVELS)[number], (...args: unknown[]) => void> {
+  const render = (args: unknown[]): string =>
+    args.map(arg => typeof arg === 'string' ? arg : inspect(arg, INSPECT_OPTIONS)).join(' ')
+  const shim = Object.create(null) as Record<(typeof CONSOLE_LEVELS)[number], (...args: unknown[]) => void>
+  for (const level of CONSOLE_LEVELS) {
+    shim[level] = (...args: unknown[]) => { logs.push({ source: 'console', level, text: render(args) }) }
+  }
+  return shim
+}
+
+/**
+ * Redirect a stream's `write` into the log buffer (the program-visible
+ * `process.stdout`/`process.stderr` in the real worker), so raw writes land
+ * in emission order alongside console output instead of racing down a pipe.
+ * The shim keeps Node's `write(chunk[, encoding][, callback])` contract: the
+ * callback fires asynchronously once the chunk is admitted (a program
+ * awaiting flush completion must complete, not sit until the wall timeout),
+ * even for writes the exhausted budget drops.
+ * @param logs - the buffer captured writes are pushed into.
+ * @param stream - the stream whose `write` slot is patched.
+ * @param source - the log source the captured writes are attributed to.
+ * @returns the restore function (the in-process tests un-patch; the real
+ *   worker never needs to).
+ */
+export function captureStreamWrites(logs: LogBuffer, stream: PatchableStream, source: 'stdout' | 'stderr'): () => void {
+  // The slot's VALUE is stored for restore and reassigned — never invoked
+  // detached, so the unbound-method concern does not apply.
+  // eslint-disable-next-line @typescript-eslint/unbound-method
+  const original = stream.write
+  stream.write = (chunk: unknown, ...rest: unknown[]): boolean => {
+    logs.push({ source, text: typeof chunk === 'string' ? chunk : String(chunk) })
+    // Node's optional-encoding shape: the callback is whichever of the next
+    // two positions holds a function (a non-function there is the encoding).
+    const callback = [rest[0], rest[1]].find(
+      (arg): arg is (error?: Error | null) => void => typeof arg === 'function',
+    )
+    if (callback) queueMicrotask(() => { callback(null) })
+    return true
+  }
+  return () => { stream.write = original }
+}
+
+/** Bounded inspect options: deep enough to be useful, bounded so a pathological value cannot explode the rendering. */
+const INSPECT_OPTIONS = { depth: 4, maxArrayLength: 100, maxStringLength: 10_000 } as const
+
+/**
+ * The longest prefix of `text` whose UTF-8 encoding fits `maxBytes`, cut at
+ * a code-point boundary (never mid-surrogate-pair). The byte caps are BYTE
+ * caps — `String.prototype.slice` counts UTF-16 code units, up to 3× smaller
+ * than what a multibyte string actually costs across the boundary.
+ * @param text - the string to bound.
+ * @param maxBytes - the UTF-8 byte budget the prefix must fit.
+ * @returns the prefix (all of `text` when it already fits).
+ */
+export function truncateUtf8Bytes(text: string, maxBytes: number): string {
+  if (Buffer.byteLength(text, 'utf8') <= maxBytes) return text
+  let bytes = 0
+  let end = 0
+  for (const char of text) {
+    const cost = Buffer.byteLength(char, 'utf8')
+    if (bytes + cost > maxBytes) break
+    bytes += cost
+    end += char.length
+  }
+  return text.slice(0, end)
+}
+
+/**
+ * Prepare the program's completion value for the done message: a value whose
+ * MEASURED cross-boundary size fits `maxValueBytes` crosses raw — exact
+ * bytes for a string, the structured-clone wire size (`v8.serialize`) for
+ * everything else, so a huge container whose BOUNDED inspect rendering
+ * happens to be small cannot smuggle itself past the cap. Anything else
+ * (non-cloneable, or oversized) is REPLACED by its bounded `util.inspect`
+ * rendering, byte-truncated ({@link truncateUtf8Bytes}) with an in-band
+ * marker — the seam contract's "a non-transferable value is replaced by a
+ * string rendering", extended to oversized ones so a huge return cannot
+ * flood the host.
+ * @param value - the program's completion value.
+ * @param maxValueBytes - the byte cap for the value.
+ * @returns the done-message fragment: `{}` for `undefined`, else `{ value }`.
+ */
+export function prepareValue(value: unknown, maxValueBytes: number): { value?: unknown } {
+  if (value === undefined) return {}
+  if (typeof value === 'string') {
+    if (Buffer.byteLength(value, 'utf8') <= maxValueBytes) return { value }
+  } else {
+    let size: number | undefined
+    try {
+      size = serialize(value).byteLength
+    } catch {
+      // Only the verdict matters: the value has parts the structured-clone
+      // algorithm rejects (functions, classes, …) and must cross as its
+      // rendering instead.
+      size = undefined
+    }
+    if (size !== undefined && size <= maxValueBytes) return { value }
+  }
+  const rendered = typeof value === 'string' ? value : inspect(value, INSPECT_OPTIONS)
+  const capped = Buffer.byteLength(rendered, 'utf8') > maxValueBytes
+    ? `${truncateUtf8Bytes(rendered, maxValueBytes)}… [truncated]`
+    : rendered
+  return { value: capped }
+}
+
+/** One awaited binding call's settlement handles, keyed by call id in the pending map. */
+export interface PendingCall {
+  resolve(value: unknown): void
+  reject(error: Error): void
+}
+
+/**
+ * Route host replies into the pending-call map: each reply settles its call
+ * at most once, and a reply for an unknown id (stray, or a duplicate answer
+ * to an id already settled) is ignored. Shared wiring between
+ * {@link runWorkerMain} and the tests that exercise {@link makeNamespaces}
+ * standalone.
+ * @param port - the port whose `message` events carry the replies.
+ * @param pending - the id-keyed map of unsettled binding calls.
+ */
+export function wireReplies(port: BootstrapPort, pending: Map<number, PendingCall>): void {
+  port.on('message', (message: ReplyMessage) => {
+    const entry = pending.get(message.id)
+    if (!entry) return
+    pending.delete(message.id)
+    if (message.ok) entry.resolve(message.value)
+    else entry.reject(new Error(message.message))
+  })
+}
+
+/**
+ * Build the binding namespace objects the program sees: one null-prototype
+ * global per namespace, each declared name an own enumerable async function
+ * that bridges over the port (`__proto__`/`constructor`/`toString` are
+ * ordinary keys, never prototype collisions). A non-cloneable argument
+ * rejects that one call with a descriptive error; the host's reply (`ok`
+ * false) rejects it likewise, so a failed tool call surfaces in the program
+ * as an ordinary promise rejection.
+ * @param data - the boot payload's namespace declarations (globals + names).
+ * @param port - the port binding calls are posted to.
+ * @param pending - the id-keyed map each posted call parks its handles in.
+ * @param nextId - the shared mutable id counter (worker-issued correlation ids).
+ * @returns one namespace object per declaration, in declaration order.
+ */
+export function makeNamespaces(
+  data: Pick<WorkerBootData, 'namespaces'>,
+  port: BootstrapPort,
+  pending: Map<number, PendingCall>,
+  nextId: { value: number },
+): Record<string, unknown>[] {
+  return data.namespaces.map(({ global, names }) => {
+    const namespace = Object.create(null) as Record<string, unknown>
+    for (const name of names) {
+      Object.defineProperty(namespace, name, {
+        enumerable: true,
+        value: (args: unknown): Promise<unknown> => new Promise((resolve, reject) => {
+          const id = nextId.value++
+          pending.set(id, { resolve, reject })
+          try {
+            port.postMessage({ type: 'call', id, global, name, args })
+          } catch (error: unknown) {
+            pending.delete(id)
+            reject(new Error(`binding arguments must be structured-cloneable: ${error instanceof Error ? error.message : String(error)}`))
+          }
+        }),
+      })
+    }
+    return namespace
+  })
+}
+
+/**
+ * Run one program to settlement and post the {@link DoneMessage}: wires the
+ * reply handler, materializes the namespaces and console shim, compiles the
+ * type-stripped body as an async function (top-level `await`/`return`
+ * work), and reports a thrown program error as the done message's `error`
+ * field. Exactly one done message is ever posted.
+ * @param port - the message port to the host (the real `parentPort`, or the tests' fake).
+ * @param data - the boot payload the host sent.
+ * @param streams - the stream objects whose `write` is captured (the real
+ *   `process.stdout`/`process.stderr` in the worker; fakes in tests).
+ * @returns resolves after the done message is posted (the tests await it;
+ *   the real entry lets the worker exit naturally).
+ */
+export async function runWorkerMain(
+  port: BootstrapPort,
+  data: WorkerBootData,
+  streams: { stdout: PatchableStream; stderr: PatchableStream },
+): Promise<void> {
+  const logs = new LogBuffer(data.maxLogBytes, (entry) => { port.postMessage({ type: 'log', entry }) })
+  captureStreamWrites(logs, streams.stdout, 'stdout')
+  captureStreamWrites(logs, streams.stderr, 'stderr')
+
+  const pending = new Map<number, PendingCall>()
+  wireReplies(port, pending)
+
+  const nextId = { value: 1 }
+  const namespaces = makeNamespaces(data, port, pending, nextId)
+  const consoleShim = makeConsoleShim(logs)
+
+  let done: DoneMessage
+  try {
+    // The async function constructor, reached through an instance because
+    // `AsyncFunction` is not a global. The program body is strict-mode.
+    /* v8 ignore next -- the arrow exists only to reach the AsyncFunction constructor; it is never invoked. */
+    const AsyncFunction = (async () => {}).constructor as new (...args: string[]) => (...fnArgs: unknown[]) => Promise<unknown>
+    const fn = new AsyncFunction(...data.namespaces.map(namespace => namespace.global), 'console', `'use strict';\n${data.code}`)
+    const value = await fn(...namespaces, consoleShim)
+    done = { type: 'done', ...prepareValue(value, data.maxValueBytes) }
+  } catch (error: unknown) {
+    const message = error instanceof Error ? error.stack ?? error.message : String(error)
+    done = { type: 'done', error: { message } }
+  }
+  port.postMessage(done)
+}

packages/code-runtime/code-runtime-worker/src/index.ts

@@ -0,0 +1,441 @@
+/**
+ * Worker-thread implementation of the code-execution seam: one fresh Node
+ * worker per run, executing the model's TypeScript after a host-side
+ * type-strip, with bindings bridged over the message port. Containment, not
+ * a security boundary (bash-equivalent trust — see the Code Mode RFC's
+ * trust-posture section): the worker gets an EMPTY environment, a heap cap,
+ * and two independent budgets — `computeMs` metered on the worker's
+ * measured event-loop busy time (a hot loop cannot hide behind a pending
+ * binding call) and a never-pausing `maxWallMs` ceiling — all funneling
+ * into `worker.terminate()`, which ends hot synchronous loops too.
+ *
+ * @module @deepseek-ai/dsh-code-runtime-worker
+ */
+
+import { Worker } from 'node:worker_threads'
+import { stripTypeScriptTypes } from 'node:module'
+import { Context } from 'cordis'
+import z from 'schemastery'
+import { CodeRuntime } from '@deepseek-ai/dsh-code-runtime'
+import type { CodeBindingFunction, CodeLogEntry, CodeRunFailure, CodeRunRequest, CodeRunResult } from '@deepseek-ai/dsh-code-runtime'
+import { prepareValue, truncateUtf8Bytes } from './bootstrap.ts'
+import { logTruncationMarker } from './protocol.ts'
+import type { ReplyMessage, WorkerBootData, WorkerToHost } from './protocol.ts'
+
+export type { BootstrapPort, PatchableStream } from './bootstrap.ts'
+export type { CallMessage, DoneMessage, ReplyMessage, WorkerBootData, WorkerToHost } from './protocol.ts'
+
+/** Plugin config: every execution cap, changeable from `cordis.yml` (no hardcoded tunables). */
+export interface Config {
+  /**
+   * Busy-time budget in milliseconds: the run fails with kind `'timeout'`
+   * once the worker's MEASURED event-loop active time
+   * (`worker.performance.eventLoopUtilization()`) exceeds this. Metering
+   * measured busy time — not wall time, not host-side pending-call
+   * bookkeeping — is what makes the budget both fair (a program awaiting a
+   * slow tool accrues nothing) and ungameable (a hot loop accrues whether
+   * or not a decoy dispatch is in flight).
+   */
+  computeMs?: number
+  /**
+   * Wall-clock ceiling in milliseconds; never pauses for anything. The
+   * backstop for what busy-time cannot see (a program awaiting a promise
+   * nobody will resolve).
+   */
+  maxWallMs?: number
+  /** Shared byte budget for captured log text (console + raw stream writes), truncation marked in-band. */
+  maxLogBytes?: number
+  /**
+   * Byte cap for the completion value, measured by its real cross-boundary
+   * size (string bytes, or structured-clone wire size); an oversized or
+   * non-cloneable value crosses as a capped string rendering.
+   */
+  maxValueBytes?: number
+  /** The worker's max old-generation heap in MiB (`resourceLimits`); overflow kills the worker, surfacing as kind `'worker-exit'`. */
+  maxOldGenerationSizeMb?: number
+}
+
+/** {@link Config} after schemastery fills the defaults (every field present). */
+type ResolvedConfig = Required<Config>
+
+/**
+ * How often the host samples the worker's event-loop utilization for the
+ * `computeMs` budget. An internal cadence, not config: the only effect of
+ * the interval is budget-expiry granularity (a run can overshoot by up to
+ * one interval), and nothing a deployment could tune here improves that
+ * without burning host CPU.
+ */
+const ELU_POLL_INTERVAL_MS = 25
+
+/** ECMAScript reserved words that cannot be async-function parameter names — rejected as binding globals. */
+const RESERVED_WORDS = new Set([
+  'await', 'break', 'case', 'catch', 'class', 'const', 'continue', 'debugger', 'default', 'delete', 'do',
+  'else', 'enum', 'export', 'extends', 'false', 'finally', 'for', 'function', 'if', 'import', 'in',
+  'instanceof', 'new', 'null', 'return', 'super', 'switch', 'this', 'throw', 'true', 'try', 'typeof',
+  'var', 'void', 'while', 'with', 'yield', 'let', 'static', 'implements', 'interface', 'package',
+  'private', 'protected', 'public', 'arguments', 'eval',
+])
+
+/** Valid async-function parameter name (the binding global becomes one). */
+const IDENTIFIER = /^[A-Za-z_$][A-Za-z0-9_$]*$/
+
+/**
+ * The shell a program is wrapped in for the type-strip, matching the
+ * grammatical context it will execute in (an async function body, where
+ * top-level `return` and `await` are legal — a bare module parse would
+ * reject the `return`). Strip mode is position-preserving (removed syntax
+ * becomes whitespace, nothing shifts), so the wrapper survives the strip
+ * byte-identical and the body slices back out with the model's own
+ * line/column positions intact.
+ */
+const STRIP_WRAP = { prefix: 'async function __dsh_program__() {\n', suffix: '\n}' } as const
+
+/** One in-flight run's host-side state, tracked for disposal. */
+interface LiveRun {
+  worker: Worker
+  settle(failure: CodeRunFailure): void
+  finished: Promise<void>
+}
+
+/**
+ * The worker entry module. Source runs unbuilt (`src/worker.ts`, loadable
+ * directly on this repo's Node range via native type stripping — the file
+ * is erasable-only with type-only relative imports); the built package
+ * ships it as a sibling bundle (`lib/worker.js`, its own tsdown entry).
+ * The URL *pathname*'s extension says which world this module is in —
+ * pathname, because dev-time module runners (vitest) may suffix
+ * `import.meta.url` with a query string; relative resolution drops it.
+ */
+/* v8 ignore next -- the './worker.js' arm is the built-lib world, unreachable unbuilt by construction; the built-lib e2e pins it. */
+const WORKER_URL = new URL(new URL(import.meta.url).pathname.endsWith('.ts') ? './worker.ts' : './worker.js', import.meta.url)
+
+/** Render an unknown thrown value as a message, `Error` or not. */
+function messageOf(error: unknown): string {
+  return error instanceof Error ? error.message : String(error)
+}
+
+/** The log sources / console levels the seam vocabulary admits, as runtime sets for inbound-message validation. */
+const LOG_SOURCES = new Set<string>(['console', 'stdout', 'stderr'])
+const LOG_LEVELS = new Set<string>(['log', 'info', 'warn', 'error', 'debug'])
+
+/**
+ * Runtime shape gate for inbound port traffic. The peer runs MODEL CODE and
+ * can post anything — `null`, primitives, objects with poisoned fields — so
+ * the compile-time `WorkerToHost` type means nothing here: everything is
+ * re-validated and REBUILT field by field (a forged extra field never rides
+ * along; a non-number call id can never be echoed into a reply). Junk returns
+ * `undefined` and is dropped — a throw in the host's `message` listener would
+ * crash the host process.
+ */
+function parseWorkerMessage(raw: unknown): WorkerToHost | undefined {
+  if (typeof raw !== 'object' || raw === null) return undefined
+  const m = raw as Record<string, unknown>
+  switch (m.type) {
+    case 'call': {
+      if (typeof m.id !== 'number' || typeof m.global !== 'string' || typeof m.name !== 'string') return undefined
+      return { type: 'call', id: m.id, global: m.global, name: m.name, args: m.args }
+    }
+    case 'log': {
+      const entry = m.entry
+      if (typeof entry !== 'object' || entry === null) return undefined
+      const e = entry as Record<string, unknown>
+      if (typeof e.text !== 'string') return undefined
+      if (typeof e.source !== 'string' || !LOG_SOURCES.has(e.source)) return undefined
+      if (e.level !== undefined && (typeof e.level !== 'string' || !LOG_LEVELS.has(e.level))) return undefined
+      return {
+        type: 'log',
+        entry: {
+          source: e.source as CodeLogEntry['source'],
+          ...e.level !== undefined ? { level: e.level as Exclude<CodeLogEntry['level'], undefined> } : {},
+          text: e.text,
+        },
+      }
+    }
+    case 'done': {
+      if (m.error === undefined) return { type: 'done', ...m.value !== undefined ? { value: m.value } : {} }
+      const error = m.error
+      if (typeof error !== 'object' || error === null) return undefined
+      const message = (error as Record<string, unknown>).message
+      if (typeof message !== 'string') return undefined
+      return { type: 'done', ...m.value !== undefined ? { value: m.value } : {}, error: { message } }
+    }
+    default: return undefined
+  }
+}
+
+/**
+ * Headroom the host's value re-cap grants over `maxValueBytes`: exactly the
+ * truncation suffix {@link prepareValue} appends, so a value the WORKER
+ * already capped (byte-exact prefix + this marker) passes through unchanged
+ * instead of being marked twice.
+ */
+const VALUE_RENDER_SLACK = Buffer.byteLength('… [truncated]', 'utf8')
+
+/**
+ * The shipped {@link CodeRuntime} backend (`ctx.codeRuntime`). Registers as
+ * the `codeRuntime` service; every cap comes from validated config. See the
+ * module doc for the containment model and the class JSDoc on the seam for
+ * the contract this implements (error-as-field, hostile-peer port,
+ * no cross-run state, dispose to quiescence).
+ */
+export class WorkerCodeRuntime extends CodeRuntime {
+  static Config: z<Config> = z.object({
+    computeMs: z.number().default(60_000),
+    maxWallMs: z.number().default(600_000),
+    maxLogBytes: z.number().default(65_536),
+    maxValueBytes: z.number().default(32_768),
+    maxOldGenerationSizeMb: z.number().default(512),
+  })
+
+  readonly language = 'typescript'
+  readonly isolation = 'worker-thread'
+
+  private readonly config: ResolvedConfig
+  private readonly live = new Set<LiveRun>()
+  private disposed = false
+
+  constructor(ctx: Context, config: Config) {
+    super(ctx)
+    // Schemastery filled the defaults; the cast records that. Positivity is a
+    // semantic check the schema's plain number type does not carry.
+    this.config = config as ResolvedConfig
+    for (const [key, value] of Object.entries(this.config)) {
+      if (!(Number.isFinite(value) && value > 0)) throw new Error(`dsh-code-runtime-worker: config.${key} must be a positive number, got ${String(value)}`)
+    }
+    ctx.effect(() => () => this.teardown(), 'worker code-runtime teardown')
+  }
+
+  /**
+   * Dispose to quiescence: mark the service unusable, fail every in-flight
+   * run as aborted, and AWAIT each worker's exit so no worker outlives the
+   * fiber.
+   */
+  private async teardown(): Promise<void> {
+    this.disposed = true
+    const runs = [...this.live]
+    for (const run of runs) run.settle({ kind: 'abort', message: 'runtime disposed' })
+    await Promise.all(runs.map(run => run.finished))
+  }
+
+  /**
+   * Execute one program in a fresh worker. Program outcomes — including a
+   * type-strip syntax error, which never spawns a worker — resolve with
+   * `result.error`; the method rejects only for seam misuse (a disposed
+   * runtime, an invalid binding namespace).
+   * @param request - the program, its bindings, and the abort signal.
+   * @returns the run's outcome per the seam contract.
+   */
+  async run(request: CodeRunRequest): Promise<CodeRunResult> {
+    if (this.disposed) throw new Error('dsh-code-runtime-worker: run() after disposal')
+    const bindings = this.validateBindings(request)
+    if (request.signal?.aborted) {
+      return { logs: [], error: { kind: 'abort', message: String(request.signal.reason) } }
+    }
+
+    let code: string
+    try {
+      const stripped = stripTypeScriptTypes(STRIP_WRAP.prefix + request.program + STRIP_WRAP.suffix)
+      code = stripped.slice(STRIP_WRAP.prefix.length, stripped.length - STRIP_WRAP.suffix.length)
+    } catch (error: unknown) {
+      // A program that does not survive the type-strip (syntax error,
+      // non-erasable syntax like `enum`) is a program failure, reported the
+      // same way a thrown exception would be — and no worker ever spawns.
+      return { logs: [], error: { kind: 'exception', message: messageOf(error) } }
+    }
+
+    return await this.execute(request, code, bindings)
+  }
+
+  /** Reject (seam misuse) malformed binding namespaces: non-identifier or reserved globals, duplicates, and the `console` collision. */
+  private validateBindings(request: CodeRunRequest): Map<string, Record<string, CodeBindingFunction>> {
+    const bindings = new Map<string, Record<string, CodeBindingFunction>>()
+    for (const namespace of request.bindings) {
+      if (!IDENTIFIER.test(namespace.global) || RESERVED_WORDS.has(namespace.global)) {
+        throw new Error(`dsh-code-runtime-worker: binding global ${JSON.stringify(namespace.global)} is not a usable identifier`)
+      }
+      if (namespace.global === 'console' || bindings.has(namespace.global)) {
+        throw new Error(`dsh-code-runtime-worker: duplicate binding global ${JSON.stringify(namespace.global)}`)
+      }
+      bindings.set(namespace.global, namespace.functions)
+    }
+    return bindings
+  }
+
+  /** Spawn the worker for one validated, type-stripped run and drive it to settlement. */
+  private execute(
+    request: CodeRunRequest,
+    code: string,
+    bindings: Map<string, Record<string, CodeBindingFunction>>,
+  ): Promise<CodeRunResult> {
+    const bootData: WorkerBootData = {
+      code,
+      namespaces: [...bindings].map(([global, functions]) => ({ global, names: Object.keys(functions) })),
+      maxLogBytes: this.config.maxLogBytes,
+      maxValueBytes: this.config.maxValueBytes,
+    }
+    const worker = new Worker(WORKER_URL, {
+      workerData: bootData,
+      // Model code gets NO ambient environment — stronger than the scrubbed
+      // env the defensive-patterns rule requires for spawned commands.
+      env: {},
+      // Hermetic flags too: without this the worker inherits the host
+      // process's execArgv (a test runner's or tsx's loader hooks), which a
+      // bare isolate with an empty environment cannot satisfy. The entry
+      // needs nothing beyond native type stripping, on this repo's whole
+      // Node range.
+      execArgv: [],
+      resourceLimits: { maxOldGenerationSizeMb: this.config.maxOldGenerationSizeMb },
+      // Backstop capture: the bootstrap patches JS-level writes into its own
+      // ordered buffer, so these pipes normally stay silent; anything that
+      // still arrives (native-level writes) is appended after the done logs.
+      stdout: true,
+      stderr: true,
+    })
+
+    return new Promise<CodeRunResult>((resolve) => {
+      let settled = false
+      const answered = new Set<number>()
+      const logs: CodeLogEntry[] = []
+      const strayLogs: CodeLogEntry[] = []
+
+      // ONE host-side ledger for everything that lands in `logs`/`strayLogs`,
+      // whatever the path: honest port entries, FORGED port entries (model
+      // code posting `log` messages directly, bypassing the worker-side
+      // LogBuffer), and stray pipe bytes. On the first overflow it emits the
+      // same in-band marker the worker's LogBuffer would and drops the rest,
+      // so the documented cap is one shared `maxLogBytes` however it is hit.
+      let logBudget = this.config.maxLogBytes
+      let logsTruncated = false
+      const admit = (entry: CodeLogEntry, sink: CodeLogEntry[]): void => {
+        if (logsTruncated) return
+        const cost = Buffer.byteLength(entry.text, 'utf8')
+        if (cost > logBudget) {
+          logsTruncated = true
+          sink.push({ source: 'stderr', text: logTruncationMarker(this.config.maxLogBytes) })
+          return
+        }
+        logBudget -= cost
+        sink.push(entry)
+      }
+
+      // No settled guard: `finish` snapshots the arrays when it resolves, so
+      // a chunk flushing after settlement mutates only the discarded buffers,
+      // and the ledger bounds that growth until the pipes close.
+      const captureStray = (source: 'stdout' | 'stderr') => (chunk: Buffer) => {
+        admit({ source, text: chunk.toString('utf8') }, strayLogs)
+      }
+      worker.stdout.on('data', captureStray('stdout'))
+      worker.stderr.on('data', captureStray('stderr'))
+
+      // Settlement: exactly one outcome wins; every path funnels through
+      // here, cleans up the timers/listeners, terminates the worker, and
+      // resolves only after the worker actually exited (quiescence). Logs
+      // streamed eagerly before the settlement are kept — a timed-out or
+      // killed program still shows the model what it printed.
+      let finishResolve!: () => void
+      const finished = new Promise<void>((done) => { finishResolve = done })
+      const finish = (result: Omit<CodeRunResult, 'logs'>): void => {
+        if (settled) return
+        settled = true
+        clearInterval(eluTimer)
+        clearTimeout(wallTimer)
+        request.signal?.removeEventListener('abort', onAbort)
+        this.live.delete(live)
+        void worker.terminate().then(() => {
+          finishResolve()
+          resolve({ ...result, logs: [...logs, ...strayLogs] })
+        })
+      }
+
+      const onDone = (message: WorkerToHost): void => {
+        if (message.type !== 'done') return
+        // Re-cap the completion value HOST-side: the honest path already
+        // capped it in the worker (prepareValue there), but a forged done
+        // message bypasses the bootstrap entirely — without this, model code
+        // could flood the host past maxValueBytes. Honest values pass
+        // unchanged (see VALUE_RENDER_SLACK); the error text is bounded too.
+        finish({
+          ...prepareValue(message.value, this.config.maxValueBytes + VALUE_RENDER_SLACK),
+          ...message.error ? { error: { kind: 'exception' as const, message: truncateUtf8Bytes(message.error.message, this.config.maxValueBytes) } } : {},
+        })
+      }
+
+      const onCall = (message: WorkerToHost): void => {
+        if (message.type !== 'call' || settled) return
+        // Hostile-peer rules: a duplicate id is ignored, an unknown name is
+        // answered with a failure, and a binding throw/reject becomes the
+        // program-side rejection — contained here, never a host crash.
+        if (answered.has(message.id)) return
+        answered.add(message.id)
+        const reply = (payload: ReplyMessage): void => {
+          if (settled) return
+          try {
+            worker.postMessage(payload)
+          } catch {
+            // The reply value failed structured clone; renegotiate as an error
+            // reply, which is always clone-plain. Nothing else throws here.
+            worker.postMessage({ type: 'reply', id: message.id, ok: false, message: 'binding resolution is not structured-cloneable' })
+          }
+        }
+        const record = bindings.get(message.global)
+        // Own-property lookup only: a forged name like 'constructor' or
+        // 'hasOwnProperty' must not walk the record's prototype chain and
+        // reach a callable the consumer never declared.
+        const fn = record && Object.hasOwn(record, message.name) ? record[message.name] : undefined
+        if (typeof fn !== 'function') {
+          reply({ type: 'reply', id: message.id, ok: false, message: `unknown binding ${JSON.stringify(`${message.global}.${message.name}`)}` })
+          return
+        }
+        void (async () => {
+          try {
+            reply({ type: 'reply', id: message.id, ok: true, value: await fn(message.args) })
+          } catch (error: unknown) {
+            reply({ type: 'reply', id: message.id, ok: false, message: messageOf(error) })
+          }
+        })()
+      }
+
+      worker.on('message', (raw: unknown) => {
+        // Parse before touching: the peer can post ANY shape, and a throw in
+        // this listener would crash the host process. Junk drops silently.
+        const message = parseWorkerMessage(raw)
+        if (!message) return
+        if (message.type === 'log' && !settled) admit(message.entry, logs)
+        onCall(message)
+        onDone(message)
+      })
+      worker.on('error', (error: Error) => {
+        finish({ error: { kind: 'worker-exit', message: `worker error: ${error.message}` } })
+      })
+      worker.on('exit', (exitCode: number) => {
+        finish({ error: { kind: 'worker-exit', message: `worker exited with code ${exitCode} before completing` } })
+      })
+
+      // The compute budget reads the worker's own measured busy time, so a
+      // hot loop expires it no matter what dispatches are in flight, while a
+      // program idling on a slow binding accrues nothing.
+      const eluTimer = setInterval(() => {
+        const elu = worker.performance.eventLoopUtilization()
+        if (elu.active > this.config.computeMs) {
+          finish({ error: { kind: 'timeout', message: `compute budget exhausted (${this.config.computeMs}ms busy)` } })
+        }
+      }, ELU_POLL_INTERVAL_MS)
+      const wallTimer = setTimeout(() => {
+        finish({ error: { kind: 'timeout', message: `wall-clock ceiling reached (${this.config.maxWallMs}ms)` } })
+      }, this.config.maxWallMs)
+      const onAbort = (): void => {
+        finish({ error: { kind: 'abort', message: String(request.signal?.reason) } })
+      }
+      request.signal?.addEventListener('abort', onAbort, { once: true })
+
+      const live: LiveRun = {
+        worker,
+        finished,
+        settle: (failure: CodeRunFailure) => { finish({ error: failure }) },
+      }
+      this.live.add(live)
+    })
+  }
+}
+
+export default WorkerCodeRuntime

packages/code-runtime/code-runtime-worker/src/protocol.ts

@@ -0,0 +1,78 @@
+/**
+ * Wire protocol between the host runtime and the worker bootstrap. Everything
+ * crossing the message port is structured-clone-plain and versionless — both
+ * ends ship in this package, always at the same version. The host treats
+ * inbound traffic as HOSTILE (the worker runs model code, which can reach
+ * `parentPort` via `import('node:worker_threads')` and forge any of these
+ * shapes); the worker treats inbound traffic as trusted.
+ *
+ * @module @deepseek-ai/dsh-code-runtime-worker/src/protocol
+ */
+
+import type { CodeLogEntry } from '@deepseek-ai/dsh-code-runtime'
+
+/** What the host hands the worker at spawn, via `workerData`. */
+export interface WorkerBootData {
+  /** The type-stripped (plain JS) program body. */
+  code: string
+  /** Binding namespaces to materialize: the global name plus the function names (functions themselves stay host-side). */
+  namespaces: { global: string; names: string[] }[]
+  /** Shared byte budget for captured log text; exceeding it drops further entries after one in-band marker. */
+  maxLogBytes: number
+  /** Byte cap for the rendered completion value (see the value-preparation contract in bootstrap.ts). */
+  maxValueBytes: number
+}
+
+/** Worker → host: one bridged binding call. */
+export interface CallMessage {
+  type: 'call'
+  /** Worker-issued correlation id; the host answers each id at most once and ignores duplicates. */
+  id: number
+  /** The namespace global the call targets. */
+  global: string
+  /** The function name within the namespace. */
+  name: string
+  /** The single argument, structured-clone-plain. */
+  args: unknown
+}
+
+/** Worker → host: one captured log entry, streamed eagerly so output survives a mid-run termination (timeout, abort, OOM). */
+export interface LogMessage {
+  type: 'log'
+  entry: CodeLogEntry
+}
+
+/**
+ * Worker → host: the program settled. `error` carries a program exception
+ * (the only failure the bootstrap itself can report — budgets, aborts, and
+ * substrate death are observed host-side). `value` is present only on a
+ * clean completion that produced one (already size-capped and
+ * clone-safe per the bootstrap's value preparation). Logs are NOT carried
+ * here — they streamed eagerly as {@link LogMessage}s.
+ */
+export interface DoneMessage {
+  type: 'done'
+  value?: unknown
+  error?: { message: string }
+}
+
+/** Every message the worker sends. */
+export type WorkerToHost = CallMessage | LogMessage | DoneMessage
+
+/** Host → worker: the answer to one {@link CallMessage}. */
+export type ReplyMessage =
+  | { type: 'reply'; id: number; ok: true; value: unknown }
+  | { type: 'reply'; id: number; ok: false; message: string }
+
+/**
+ * The in-band marker entry text announcing that log capture stopped at the
+ * byte budget. Shared wire vocabulary: the worker's LogBuffer emits it when
+ * ITS budget exhausts, and the host emits the identical text when its own
+ * ledger drops an entry first (forged port traffic, stray pipe bytes) — so
+ * a truncated run reads the same however the cap was hit.
+ * @param maxBytes - the configured `maxLogBytes` the marker names.
+ * @returns the marker line.
+ */
+export function logTruncationMarker(maxBytes: number): string {
+  return `[dsh-code-runtime-worker] log capture truncated at ${maxBytes} bytes`
+}

packages/code-runtime/code-runtime-worker/src/worker.ts

@@ -0,0 +1,20 @@
+/**
+ * The worker-thread entrypoint: self-executing glue over
+ * `bootstrap.ts`'s {@link runWorkerMain}, kept to the spawn wiring alone.
+ * Like `bin.ts` CLI entrypoints, this file executes only inside a spawned
+ * worker isolate — a place the coverage provider cannot observe — so it is
+ * excluded from the coverage gate while every line of actual logic lives in
+ * `bootstrap.ts`, unit-tested in-process; the real spawn path is pinned by
+ * the integration tests that run genuine workers.
+ *
+ * @module @deepseek-ai/dsh-code-runtime-worker/src/worker
+ */
+
+import { parentPort, workerData } from 'node:worker_threads'
+import { runWorkerMain } from './bootstrap.ts'
+import type { WorkerBootData } from './protocol.ts'
+
+// A worker always has a parent port; guard loudly rather than run detached.
+if (!parentPort) throw new Error('dsh-code-runtime-worker: worker entry loaded outside a worker thread')
+
+await runWorkerMain(parentPort, workerData as WorkerBootData, { stdout: process.stdout, stderr: process.stderr })

packages/code-runtime/code-runtime-worker/tests/bootstrap.spec.ts

@@ -0,0 +1,273 @@
+import { describe, expect, it } from 'vitest'
+import { EventEmitter } from 'node:events'
+import { LogBuffer, makeConsoleShim, makeNamespaces, captureStreamWrites, prepareValue, runWorkerMain, truncateUtf8Bytes, wireReplies } from '@deepseek-ai/dsh-code-runtime-worker/src/bootstrap.ts'
+import type { BootstrapPort, PatchableStream, PendingCall } from '@deepseek-ai/dsh-code-runtime-worker/src/bootstrap.ts'
+import type { ReplyMessage, WorkerToHost } from '@deepseek-ai/dsh-code-runtime-worker/src/protocol.ts'
+import type { CodeLogEntry } from '@deepseek-ai/dsh-code-runtime'
+
+/**
+ * An in-process stand-in for the worker's parentPort: the test plays the
+ * HOST side — inspect what the bootstrap posted, feed replies back — so
+ * every line of worker-side logic runs under coverage without spawning an
+ * isolate (real-worker behavior is pinned by runtime.spec.ts).
+ */
+class FakePort implements BootstrapPort {
+  sent: WorkerToHost[] = []
+  private readonly emitter = new EventEmitter()
+  /** Host-scripted responder; return undefined to leave the call pending. */
+  respond: (message: WorkerToHost) => ReplyMessage | undefined = () => undefined
+
+  postMessage(message: WorkerToHost): void {
+    this.sent.push(message)
+    const reply = this.respond(message)
+    if (reply) queueMicrotask(() => this.emitter.emit('message', reply))
+  }
+
+  on(event: 'message', listener: (message: ReplyMessage) => void): void {
+    this.emitter.on(event, listener)
+  }
+
+  deliver(message: ReplyMessage): void {
+    this.emitter.emit('message', message)
+  }
+
+  logs(): CodeLogEntry[] {
+    return this.sent.filter(message => message.type === 'log').map(message => message.entry)
+  }
+
+  done(): WorkerToHost | undefined {
+    return this.sent.find(message => message.type === 'done')
+  }
+}
+
+function fakeStreams(): { stdout: PatchableStream; stderr: PatchableStream } {
+  return { stdout: { write: () => true }, stderr: { write: () => true } }
+}
+
+const BOOT = { maxLogBytes: 65_536, maxValueBytes: 32_768 }
+
+describe('LogBuffer', () => {
+  it('streams entries to the sink until the byte budget, then emits one marker and drops the rest', () => {
+    const seen: CodeLogEntry[] = []
+    const buffer = new LogBuffer(10, entry => seen.push(entry))
+    buffer.push({ source: 'console', level: 'log', text: '12345' })
+    buffer.push({ source: 'console', level: 'log', text: '123456' })
+    buffer.push({ source: 'console', level: 'log', text: 'dropped' })
+    expect(seen.map(entry => entry.text)).toEqual([
+      '12345',
+      '[dsh-code-runtime-worker] log capture truncated at 10 bytes',
+    ])
+  })
+})
+
+describe('makeConsoleShim', () => {
+  it('captures the five levels and renders non-strings inspect-style', () => {
+    const seen: CodeLogEntry[] = []
+    const shim = makeConsoleShim(new LogBuffer(1_000, entry => seen.push(entry)))
+    shim.log('plain', { a: 1 })
+    shim.info('i')
+    shim.warn('w')
+    shim.error('e')
+    shim.debug('d')
+    expect(seen.map(entry => entry.level)).toEqual(['log', 'info', 'warn', 'error', 'debug'])
+    expect(seen[0]?.text).toBe('plain { a: 1 }')
+    expect(seen.every(entry => entry.source === 'console')).toBe(true)
+  })
+})
+
+describe('captureStreamWrites', () => {
+  it('redirects writes into the buffer and restores on request', () => {
+    const seen: CodeLogEntry[] = []
+    const buffer = new LogBuffer(1_000, entry => seen.push(entry))
+    let underlying = ''
+    const stream: PatchableStream = { write: (chunk: unknown) => { underlying += String(chunk); return true } }
+    const restore = captureStreamWrites(buffer, stream, 'stdout')
+    stream.write('captured', 'utf8')
+    stream.write(Buffer.from('bytes'))
+    restore()
+    stream.write('after')
+    expect(seen.map(entry => entry.text)).toEqual(['captured', 'bytes'])
+    expect(seen[0]).toMatchObject({ source: 'stdout' })
+    expect(underlying).toBe('after')
+  })
+
+  it('invokes the write callback asynchronously, in both optional-encoding shapes', async () => {
+    const buffer = new LogBuffer(1_000, () => {})
+    const stream: PatchableStream = { write: () => true }
+    captureStreamWrites(buffer, stream, 'stdout')
+    const calls: (Error | null | undefined)[] = []
+    stream.write('two-arg', (error?: Error | null) => calls.push(error))
+    stream.write('three-arg', 'utf8', (error?: Error | null) => calls.push(error))
+    // Node's contract: the callback fires after the write call returns.
+    expect(calls).toEqual([])
+    await new Promise<void>(resolve => stream.write('awaited flush', resolve))
+    expect(calls).toEqual([null, null])
+  })
+
+  it('still fires the callback for a write the exhausted budget drops', async () => {
+    const buffer = new LogBuffer(4, () => {})
+    const stream: PatchableStream = { write: () => true }
+    captureStreamWrites(buffer, stream, 'stdout')
+    stream.write('this write overflows the budget and is dropped')
+    await new Promise<void>(resolve => stream.write('also dropped', resolve))
+  })
+})
+
+describe('prepareValue', () => {
+  it('omits undefined, passes small cloneable values raw', () => {
+    expect(prepareValue(undefined, 100)).toEqual({})
+    expect(prepareValue({ a: [1, 'two'] }, 100)).toEqual({ value: { a: [1, 'two'] } })
+  })
+
+  it('replaces a non-cloneable value with its rendering', () => {
+    const { value } = prepareValue({ fn: () => 1 }, 1_000)
+    expect(typeof value).toBe('string')
+    expect(value).toContain('fn')
+  })
+
+  it('replaces an oversized value with a truncation-marked capped rendering', () => {
+    const { value } = prepareValue('x'.repeat(50), 10)
+    expect(value).toBe(`${'x'.repeat(10)}… [truncated]`)
+  })
+
+  it('measures a container by its structured-clone wire size, not its bounded rendering', () => {
+    // The bounded inspect rendering of a huge array is tiny ("... N more
+    // items"), but its real cross-boundary size is not — the cap must catch
+    // it, replacing the value with that bounded rendering.
+    const huge = new Array(50_000).fill(7)
+    const { value } = prepareValue(huge, 1_000)
+    expect(typeof value).toBe('string')
+    expect(value).toContain('more items')
+  })
+
+  it('caps a multibyte string by UTF-8 bytes, not UTF-16 length', () => {
+    // 4 code units but 12 UTF-8 bytes: a length-counting cap would pass the
+    // full string through untruncated.
+    expect(prepareValue('€€€€', 4)).toEqual({ value: '€… [truncated]' })
+  })
+
+  it('caps a multibyte rendering by UTF-8 bytes too', () => {
+    // Wire size (24-byte string inside an array) exceeds the cap, so the
+    // value crosses as its rendering — whose truncation must also be
+    // byte-exact: "[ '" (3 bytes) + two € (6 bytes) = 9; a third € would
+    // overflow the 10-byte budget.
+    expect(prepareValue(['€€€€€€€€'], 10)).toEqual({ value: "[ '€€… [truncated]" })
+  })
+})
+
+describe('truncateUtf8Bytes', () => {
+  it('returns a fitting string whole', () => {
+    expect(truncateUtf8Bytes('fits', 4)).toBe('fits')
+  })
+
+  it('cuts at a code-point boundary, never mid-surrogate-pair', () => {
+    // Each 😀 is one code point, two code units, four UTF-8 bytes: a 5-byte
+    // budget fits exactly one — and never leaves a lone surrogate behind.
+    const cut = truncateUtf8Bytes('😀😀', 5)
+    expect(cut).toBe('😀')
+    expect(Buffer.byteLength(truncateUtf8Bytes('😀😀', 3), 'utf8')).toBe(0)
+  })
+})
+
+describe('makeNamespaces', () => {
+  it('exposes prototype-colliding names as ordinary own properties', async () => {
+    const port = new FakePort()
+    port.respond = message => message.type === 'call' ? { type: 'reply', id: message.id, ok: true, value: `${message.name}-ok` } : undefined
+    const pending = new Map<number, PendingCall>()
+    wireReplies(port, pending)
+    const [tools] = makeNamespaces({ namespaces: [{ global: 'tools', names: ['__proto__', 'constructor', 'toString'] }] }, port, pending, { value: 1 }) as [Record<string, (args: unknown) => Promise<unknown>>]
+    expect(Object.getPrototypeOf(tools)).toBeNull()
+    await expect(tools['__proto__']?.({})).resolves.toBe('__proto__-ok')
+    await expect(tools['constructor']?.({})).resolves.toBe('constructor-ok')
+    await expect(tools['toString']?.({})).resolves.toBe('toString-ok')
+  })
+
+  it('rejects a non-cloneable argument without leaking the pending entry', async () => {
+    let firstCall = true
+    const throwingPort: BootstrapPort = {
+      // First call throws an Error (the real DataCloneError shape), the
+      // second a bare string — the rejection renders both.
+      postMessage: () => {
+        if (firstCall) { firstCall = false; throw new Error('DataCloneError-ish') }
+        throw 'raw-clone-failure'
+      },
+      on: () => {},
+    }
+    const pending = new Map<number, PendingCall>()
+    const [tools] = makeNamespaces({ namespaces: [{ global: 'tools', names: ['x'] }] }, throwingPort, pending, { value: 1 }) as [Record<string, (args: unknown) => Promise<unknown>>]
+    await expect(tools.x?.(() => 1)).rejects.toThrow(/structured-cloneable: DataCloneError-ish/)
+    await expect(tools.x?.(() => 1)).rejects.toThrow(/structured-cloneable: raw-clone-failure/)
+    expect(pending.size).toBe(0)
+  })
+})
+
+describe('runWorkerMain', () => {
+  it('runs a program end-to-end: bindings, console, return value', async () => {
+    const port = new FakePort()
+    port.respond = message => message.type === 'call' ? { type: 'reply', id: message.id, ok: true, value: (message.args as { n: number }).n * 2 } : undefined
+    await runWorkerMain(port, {
+      ...BOOT,
+      code: 'const doubled = await tools.double({ n: 21 }); console.log("got", doubled); return { doubled };',
+      namespaces: [{ global: 'tools', names: ['double'] }],
+    }, fakeStreams())
+    expect(port.logs()).toEqual([{ source: 'console', level: 'log', text: 'got 42' }])
+    expect(port.done()).toEqual({ type: 'done', value: { doubled: 42 } })
+  })
+
+  it('reports a thrown program error on the done message', async () => {
+    const port = new FakePort()
+    await runWorkerMain(port, { ...BOOT, code: 'throw new Error("boom")', namespaces: [] }, fakeStreams())
+    const done = port.done()
+    expect(done?.type).toBe('done')
+    expect(done?.type === 'done' ? done.error?.message : undefined).toContain('boom')
+    expect(done?.type === 'done' ? done.value : undefined).toBeUndefined()
+  })
+
+  it('renders non-Error throws and stack-less Errors on the done message', async () => {
+    const rawPort = new FakePort()
+    await runWorkerMain(rawPort, { ...BOOT, code: 'throw "raw-throw"', namespaces: [] }, fakeStreams())
+    expect(rawPort.done()).toEqual({ type: 'done', error: { message: 'raw-throw' } })
+
+    const barePort = new FakePort()
+    await runWorkerMain(barePort, { ...BOOT, code: 'const e = new Error("bare"); e.stack = undefined; throw e', namespaces: [] }, fakeStreams())
+    expect(barePort.done()).toEqual({ type: 'done', error: { message: 'bare' } })
+  })
+
+  it('surfaces a host failure reply as a program-side rejection it can catch', async () => {
+    const port = new FakePort()
+    port.respond = message => message.type === 'call' ? { type: 'reply', id: message.id, ok: false, message: 'denied by host' } : undefined
+    await runWorkerMain(port, {
+      ...BOOT,
+      code: 'try { await tools.x({}) } catch (error) { return `caught: ${error.message}` }',
+      namespaces: [{ global: 'tools', names: ['x'] }],
+    }, fakeStreams())
+    expect(port.done()).toEqual({ type: 'done', value: 'caught: denied by host' })
+  })
+
+  it('ignores replies for unknown pending ids', async () => {
+    const port = new FakePort()
+    port.respond = (message) => {
+      if (message.type !== 'call') return undefined
+      // Deliver a stray reply first; the real one follows.
+      port.deliver({ type: 'reply', id: 9_999, ok: true, value: 'stray' })
+      return { type: 'reply', id: message.id, ok: true, value: 'real' }
+    }
+    await runWorkerMain(port, {
+      ...BOOT,
+      code: 'return await tools.x({})',
+      namespaces: [{ global: 'tools', names: ['x'] }],
+    }, fakeStreams())
+    expect(port.done()).toEqual({ type: 'done', value: 'real' })
+  })
+
+  it('captures raw stream writes through the patched process streams', async () => {
+    const port = new FakePort()
+    const streams = fakeStreams()
+    await runWorkerMain(port, { ...BOOT, code: 'return 1', namespaces: [] }, streams)
+    streams.stdout.write('never seen — already restored? no: patch persists in worker')
+    // The patch stays installed for the worker's lifetime; writes during the
+    // program landed in order. Here the program wrote nothing via streams, so
+    // only the post-run write above went through the patched slot.
+    expect(port.logs().at(-1)).toMatchObject({ source: 'stdout' })
+  })
+})

packages/code-runtime/code-runtime-worker/tests/built-lib.e2e.ts

@@ -0,0 +1,55 @@
+import { spawn } from 'node:child_process'
+import { existsSync } from 'node:fs'
+import { join } from 'node:path'
+import { fileURLToPath } from 'node:url'
+import { describe, expect, it } from 'vitest'
+
+/**
+ * BUILT-ARTIFACT smoke for the published package (the real-load-path guard
+ * from docs/testing.md): the unit suite runs `src/` under vitest, where the
+ * worker entry resolves to `src/worker.ts` — a consumer runs `lib/index.js`
+ * under plain `node`, where it must resolve the sibling `lib/worker.js`
+ * bundle instead. This spawns plain `node` (NOT tsx) from inside the package
+ * directory and imports the package BY NAME, so resolution flows through the
+ * real `exports` map exactly as it would from a downstream install; the
+ * program exercises the type-strip, the worker spawn, the binding bridge,
+ * and log capture end-to-end through the built bundles.
+ *
+ * It build-gates: SKIPS when the built artifacts are absent (suite run
+ * without `pnpm run build`); CI runs it after the build step. KEYLESS — no
+ * model is involved.
+ */
+
+const pkgDir = fileURLToPath(new URL('..', import.meta.url))
+const built = ['lib/index.js', 'lib/worker.js'].every(file => existsSync(join(pkgDir, file)))
+  && existsSync(join(pkgDir, '../code-runtime/lib/index.js'))
+
+describe.skipIf(!built)('built lib real load path (plain node)', () => {
+  it('runs a TypeScript program with a binding through lib/index.js and its lib/worker.js entry', async () => {
+    const script = `
+      const { Context } = await import('cordis')
+      const { WorkerCodeRuntime } = await import('@deepseek-ai/dsh-code-runtime-worker')
+      const ctx = new Context()
+      await ctx.plugin(WorkerCodeRuntime, {})
+      const result = await ctx.codeRuntime.run({
+        program: 'const doubled: number = await tools.double({ n: 21 }); console.log("halfway", doubled); return doubled;',
+        bindings: [{ global: 'tools', functions: { double: async args => args.n * 2 } }],
+      })
+      console.log(JSON.stringify(result))
+      process.exit(0)
+    `
+    const child = spawn(process.execPath, ['--input-type=module', '-e', script], { cwd: pkgDir, stdio: ['ignore', 'pipe', 'pipe'] })
+    let stdout = ''
+    let stderr = ''
+    child.stdout.on('data', (chunk: Buffer) => { stdout += chunk.toString('utf8') })
+    child.stderr.on('data', (chunk: Buffer) => { stderr += chunk.toString('utf8') })
+    const exitCode = await new Promise<number | null>(resolve => child.on('close', resolve))
+
+    expect(exitCode, `stderr:\n${stderr}`).toBe(0)
+    const lastLine = stdout.trim().split('\n').at(-1) ?? ''
+    const result = JSON.parse(lastLine) as { value?: unknown; logs: { source: string; level?: string; text: string }[]; error?: unknown }
+    expect(result.error).toBeUndefined()
+    expect(result.value).toBe(42)
+    expect(result.logs).toContainEqual({ source: 'console', level: 'log', text: 'halfway 42' })
+  })
+})

packages/code-runtime/code-runtime-worker/tests/runtime.spec.ts

@@ -0,0 +1,451 @@
+import { describe, expect, it } from 'vitest'
+import { Context } from 'cordis'
+import { WorkerCodeRuntime } from '@deepseek-ai/dsh-code-runtime-worker'
+import type { Config } from '@deepseek-ai/dsh-code-runtime-worker'
+import type { CodeRunResult } from '@deepseek-ai/dsh-code-runtime'
+
+/**
+ * Integration suite over REAL worker threads (no mocks — workers are cheap
+ * and local, per docs/testing.md's real-over-mock policy). Each test builds
+ * a fresh context so budgets can be tuned per case.
+ */
+async function setup(config: Config = {}) {
+  const ctx = new Context()
+  await ctx.plugin(WorkerCodeRuntime, config)
+  const runtime = ctx.codeRuntime as WorkerCodeRuntime
+  return { ctx, runtime }
+}
+
+/** Convenience: one namespace `tools` with the given functions. */
+function tools(functions: Record<string, (args: unknown) => Promise<unknown>>) {
+  return [{ global: 'tools', functions }]
+}
+
+describe('WorkerCodeRuntime — programs and bindings (real workers)', () => {
+  it('registers with the seam descriptors', async () => {
+    const { runtime } = await setup()
+    expect(runtime.language).toBe('typescript')
+    expect(runtime.isolation).toBe('worker-thread')
+  })
+
+  it('runs TypeScript (erasable syntax), captures console/stdout in order, returns the value', async () => {
+    const { runtime } = await setup()
+    const result = await runtime.run({
+      program: `
+        interface Point { x: number; y: number }
+        const p: Point = { x: 1, y: 2 } as Point;
+        console.log('point', p);
+        process.stdout.write('raw-out\\n');
+        console.warn('careful');
+        return p.x + p.y;
+      `,
+      bindings: [],
+    })
+    expect(result.error).toBeUndefined()
+    expect(result.value).toBe(3)
+    expect(result.logs.map(entry => [entry.source, entry.level ?? null])).toEqual([
+      ['console', 'log'],
+      ['stdout', null],
+      ['console', 'warn'],
+    ])
+    expect(result.logs[0]?.text).toBe('point { x: 1, y: 2 }')
+  })
+
+  it('bridges binding calls both ways and rejects the program-side call on a host rejection', async () => {
+    const { runtime } = await setup()
+    const calls: unknown[] = []
+    const result = await runtime.run({
+      program: `
+        const first = await tools.echo({ n: 1 });
+        let caught = '';
+        try { await tools.fail({}) } catch (error) { caught = error.message }
+        let caughtRaw = '';
+        try { await tools.failRaw({}) } catch (error) { caughtRaw = error.message }
+        return { first, caught, caughtRaw };
+      `,
+      bindings: tools({
+        echo: async (args) => { calls.push(args); return { echoed: args } },
+        fail: async () => { throw new Error('nope') },
+        // A non-Error throw: the host renders it, the program still catches.
+        failRaw: async () => { throw 'raw-nope' },
+      }),
+    })
+    expect(result.error).toBeUndefined()
+    expect(result.value).toEqual({ first: { echoed: { n: 1 } }, caught: 'nope', caughtRaw: 'raw-nope' })
+    expect(calls).toEqual([{ n: 1 }])
+  })
+
+  it('reports non-erasable syntax as an exception without spawning a worker', async () => {
+    const { runtime } = await setup()
+    const result = await runtime.run({ program: 'enum E { A }\nreturn 1', bindings: [] })
+    expect(result.error?.kind).toBe('exception')
+    expect(result.error?.message).toMatch(/enum|strip/i)
+  })
+
+  it('reports a runtime throw as an exception with the message', async () => {
+    const { runtime } = await setup()
+    const result = await runtime.run({ program: 'throw new Error("kaboom")', bindings: [] })
+    expect(result.error?.kind).toBe('exception')
+    expect(result.error?.message).toContain('kaboom')
+  })
+
+  it('gives the program an EMPTY environment', async () => {
+    const { runtime } = await setup()
+    const result = await runtime.run({ program: 'return JSON.stringify(process.env)', bindings: [] })
+    expect(result.value).toBe('{}')
+  })
+
+  it('replaces a non-cloneable return value with a string rendering', async () => {
+    const { runtime } = await setup()
+    const result = await runtime.run({ program: 'return { f: () => 1 }', bindings: [] })
+    expect(typeof result.value).toBe('string')
+  })
+
+  it('completes a program that returns nothing with no value at all', async () => {
+    const { runtime } = await setup()
+    const result = await runtime.run({ program: 'const x = 1', bindings: [] })
+    expect(result.error).toBeUndefined()
+    expect('value' in result).toBe(false)
+  })
+
+  it('keeps logs streamed before a failure', async () => {
+    const { runtime } = await setup()
+    const result = await runtime.run({
+      program: 'console.log("before"); throw new Error("after-log")',
+      bindings: [],
+    })
+    expect(result.error?.kind).toBe('exception')
+    expect(result.logs.map(entry => entry.text)).toContain('before')
+  })
+})
+
+describe('WorkerCodeRuntime — budgets and containment (real workers)', () => {
+  it('ends a hot loop at the compute budget — including behind a pending decoy dispatch', async () => {
+    const { runtime } = await setup({ computeMs: 300, maxWallMs: 30_000 })
+    const result = await runtime.run({
+      // The decoy: fire a call at a never-resolving binding WITHOUT awaiting,
+      // then spin. Host-side pending-call bookkeeping would pause a naive
+      // budget here; measured busy time cannot be fooled.
+      program: 'void tools.slow({}); for (;;) {}',
+      bindings: tools({ slow: () => new Promise(() => {}) }),
+    })
+    expect(result.error?.kind).toBe('timeout')
+    expect(result.error?.message).toContain('compute budget')
+  }, 15_000)
+
+  it('does not charge time spent awaiting a slow binding against the compute budget', async () => {
+    const { runtime } = await setup({ computeMs: 250, maxWallMs: 30_000 })
+    const result = await runtime.run({
+      program: 'return await tools.slow({})',
+      bindings: tools({ slow: () => new Promise(resolve => setTimeout(() => { resolve('slow-done') }, 700)) }),
+    })
+    expect(result.error).toBeUndefined()
+    expect(result.value).toBe('slow-done')
+  }, 15_000)
+
+  it('ends an idle-forever run at the wall-clock ceiling', async () => {
+    const { runtime } = await setup({ computeMs: 30_000, maxWallMs: 400 })
+    const result = await runtime.run({
+      program: 'await tools.never({}); return 1',
+      bindings: tools({ never: () => new Promise(() => {}) }),
+    })
+    expect(result.error?.kind).toBe('timeout')
+    expect(result.error?.message).toContain('wall-clock ceiling')
+  }, 15_000)
+
+  it('reports an abort mid-run and stops the worker', async () => {
+    const { runtime } = await setup()
+    const controller = new AbortController()
+    setTimeout(() => { controller.abort('user-cancel') }, 150)
+    const result = await runtime.run({ program: 'for (;;) {}', bindings: [], signal: controller.signal })
+    expect(result.error).toEqual({ kind: 'abort', message: 'user-cancel' })
+  }, 15_000)
+
+  it('reports a pre-aborted signal without spawning', async () => {
+    const { runtime } = await setup()
+    const controller = new AbortController()
+    controller.abort('too-late')
+    const result = await runtime.run({ program: 'return 1', bindings: [], signal: controller.signal })
+    expect(result.error).toEqual({ kind: 'abort', message: 'too-late' })
+  })
+
+  it('drops a binding resolution that lands after the run settled', async () => {
+    const { runtime } = await setup()
+    const controller = new AbortController()
+    let replyDelivered!: Promise<void>
+    const result = await runtime.run({
+      program: 'void tools.late({}); for (;;) {}',
+      bindings: tools({
+        // Anchored on invocation: abort 100ms after the call reaches the
+        // host, resolve 400ms after — by then the run has settled, so the
+        // resolution's reply hits the post-settlement drop.
+        late: () => new Promise((resolve) => {
+          setTimeout(() => { controller.abort('cancel-now') }, 100)
+          replyDelivered = new Promise(done => setTimeout(() => { resolve('too-late'); done() }, 400))
+        }),
+      }),
+      signal: controller.signal,
+    })
+    expect(result.error).toEqual({ kind: 'abort', message: 'cancel-now' })
+    // Let the late resolution actually fire so its reply executes instead of
+    // being cancelled with the test.
+    await replyDelivered
+  }, 15_000)
+
+  it('contains an OOM under resourceLimits as worker-exit, host process healthy', async () => {
+    const { runtime } = await setup({ maxOldGenerationSizeMb: 32 })
+    const result = await runtime.run({
+      program: 'const hog = []; for (;;) hog.push(new Array(1e6).fill(1));',
+      bindings: [],
+    })
+    expect(result.error?.kind).toBe('worker-exit')
+    // And the host is fine: run something else.
+    const after = await runtime.run({ program: 'return "alive"', bindings: [] })
+    expect(after.value).toBe('alive')
+  }, 30_000)
+
+  it('truncates runaway log output at the byte budget with an in-band marker', async () => {
+    const { runtime } = await setup({ maxLogBytes: 300 })
+    const result = await runtime.run({
+      program: 'for (let i = 0; i < 1000; i++) console.log("spam line", i); return 1',
+      bindings: [],
+    })
+    expect(result.logs.at(-1)?.text).toContain('truncated at 300 bytes')
+    const total = result.logs.reduce((sum, entry) => sum + Buffer.byteLength(entry.text, 'utf8'), 0)
+    expect(total).toBeLessThan(1_000)
+  })
+
+  it('caps an oversized return value with a truncation marker', async () => {
+    const { runtime } = await setup({ maxValueBytes: 64 })
+    const result = await runtime.run({ program: 'return "y".repeat(10_000)', bindings: [] })
+    expect(result.value).toBe(`${'y'.repeat(64)}… [truncated]`)
+  })
+
+  it('caps a multibyte return value by UTF-8 bytes, not string length', async () => {
+    // 4 code units, 12 UTF-8 bytes: a length-counting cap would let the full
+    // string cross. The worker's byte-exact capped rendering then passes the
+    // host re-cap unchanged (cap + marker is exactly the granted slack).
+    const { runtime } = await setup({ maxValueBytes: 4 })
+    const result = await runtime.run({ program: 'return "€€€€"', bindings: [] })
+    expect(result.value).toBe('€… [truncated]')
+  })
+
+  it('completes a program that awaits its write callback, capturing the chunk', async () => {
+    // Node's write(chunk[, encoding][, callback]) contract: dropping the
+    // callback would leave this promise pending until the wall ceiling and
+    // misreport a completed program as a timeout.
+    const { runtime } = await setup({ maxWallMs: 2_000 })
+    const result = await runtime.run({
+      program: 'await new Promise(resolve => process.stdout.write("flushed", resolve)); return "done"',
+      bindings: [],
+    })
+    expect(result.error).toBeUndefined()
+    expect(result.value).toBe('done')
+    expect(result.logs).toContainEqual({ source: 'stdout', text: 'flushed' })
+  })
+
+  it('caps a huge container whose bounded rendering is small (wire size, not rendering, is what counts)', async () => {
+    const { runtime } = await setup()
+    const result = await runtime.run({ program: 'return new Array(50_000).fill(7)', bindings: [] })
+    expect(result.error).toBeUndefined()
+    expect(typeof result.value).toBe('string')
+    expect(result.value).toContain('more items')
+  })
+
+  it('captures pipe writes that bypass the patched write slot as stray logs, capped by the same budget', async () => {
+    const { runtime } = await setup({ maxLogBytes: 4 })
+    const result = await runtime.run({
+      // The bootstrap patches the stream instance's own `write`; going
+      // through the prototype's slot reaches the real pipe underneath, so
+      // the bytes arrive host-side as stray data. The pauses keep the two
+      // writes in separate pipe chunks and let them land before settlement.
+      program: `
+        const write = (text) => Object.getPrototypeOf(process.stdout).write.call(process.stdout, text);
+        write('abcd');
+        await new Promise(resolve => setTimeout(resolve, 150));
+        write('ef');
+        await new Promise(resolve => setTimeout(resolve, 100));
+        return 1;
+      `,
+      bindings: [],
+    })
+    expect(result.error).toBeUndefined()
+    expect(result.logs).toContainEqual({ source: 'stdout', text: 'abcd' })
+    expect(result.logs.map(entry => entry.text)).not.toContain('ef')
+  }, 15_000)
+})
+
+describe('WorkerCodeRuntime — hostile programs (real workers)', () => {
+  it('survives forged port traffic: unknown binding names, duplicate ids, junk shapes', async () => {
+    const { runtime } = await setup()
+    const result = await runtime.run({
+      program: `
+        const { parentPort } = await import('node:worker_threads');
+        parentPort.postMessage({ type: 'call', id: 7777, global: 'tools', name: 'missing', args: {} });
+        parentPort.postMessage({ type: 'call', id: 7777, global: 'tools', name: 'missing', args: {} });
+        parentPort.postMessage({ type: 'call', id: 7778, global: 'tools', name: 'constructor', args: {} });
+        parentPort.postMessage({ type: 'junk' });
+        return await tools.real({});
+      `,
+      bindings: tools({ real: async () => 'still-works' }),
+    })
+    expect(result.error).toBeUndefined()
+    expect(result.value).toBe('still-works')
+  })
+
+  it('survives arbitrary junk on the port: non-objects, junk types, malformed calls, logs, and dones', async () => {
+    const { runtime } = await setup()
+    const result = await runtime.run({
+      program: `
+        const { parentPort } = await import('node:worker_threads');
+        for (const junk of [
+          null, 42, 'junk', [],
+          { type: 'nope' },
+          { type: 'call' },
+          { type: 'call', id: 'x', global: 'tools', name: 'real', args: {} },
+          { type: 'call', id: 1e9, global: 7, name: 'real', args: {} },
+          { type: 'call', id: 1e9, global: 'tools', name: 7, args: {} },
+          { type: 'log' },
+          { type: 'log', entry: null },
+          { type: 'log', entry: { source: 'stdout', text: 7 } },
+          { type: 'log', entry: { source: 'nope', text: 'x' } },
+          { type: 'log', entry: { source: 'console', level: 'nope', text: 'x' } },
+          { type: 'log', entry: { source: 'console', level: 7, text: 'x' } },
+          { type: 'done', error: 5 },
+          { type: 'done', error: { message: 5 } },
+        ]) parentPort.postMessage(junk);
+        return await tools.real({});
+      `,
+      bindings: tools({ real: async () => 'still-works' }),
+    })
+    expect(result.error).toBeUndefined()
+    expect(result.value).toBe('still-works')
+    expect(result.logs).toEqual([])
+  })
+
+  it('caps forged log floods and forged done values at the configured budgets, dropping forged extra fields', async () => {
+    const { runtime } = await setup({ maxLogBytes: 200, maxValueBytes: 64 })
+    const result = await runtime.run({
+      // Forged messages bypass the worker-side LogBuffer and prepareValue
+      // entirely — only the host-side ledger and re-cap stand between model
+      // code and an unbounded result.
+      program: `
+        const { parentPort } = await import('node:worker_threads');
+        for (let i = 0; i < 50; i++) parentPort.postMessage({ type: 'log', entry: { source: 'stdout', text: 'F'.repeat(100), forged: true } });
+        parentPort.postMessage({ type: 'done', value: 'V'.repeat(100000) });
+        for (;;) {}
+      `,
+      bindings: [],
+    })
+    expect(typeof result.value).toBe('string')
+    const value = result.value as string
+    expect(value.startsWith('V'.repeat(64))).toBe(true)
+    expect(value.endsWith('… [truncated]')).toBe(true)
+    expect(value.length).toBeLessThan(120)
+    const marker = '[dsh-code-runtime-worker] log capture truncated at 200 bytes'
+    const total = result.logs.reduce((sum, entry) => sum + Buffer.byteLength(entry.text, 'utf8'), 0)
+    expect(total).toBeLessThanOrEqual(200 + Buffer.byteLength(marker, 'utf8'))
+    expect(result.logs.at(-1)?.text).toBe(marker)
+    expect(result.logs.every(entry => !('forged' in entry))).toBe(true)
+  })
+
+  it('accepts a forged done carrying both value and error (self-sabotage, contained)', async () => {
+    const { runtime } = await setup()
+    const result = await runtime.run({
+      program: `
+        const { parentPort } = await import('node:worker_threads');
+        parentPort.postMessage({ type: 'done', value: 'lied', error: { message: 'fake failure' } });
+        for (;;) {}
+      `,
+      bindings: [],
+    })
+    expect(result.value).toBe('lied')
+    expect(result.error).toEqual({ kind: 'exception', message: 'fake failure' })
+  })
+
+  it('byte-bounds forged multibyte error text at the host', async () => {
+    // Forged error text bypasses the worker entirely; the host bound is a
+    // BYTE bound (two € = 6 bytes fit an 8-byte cap, a third would not).
+    const { runtime } = await setup({ maxValueBytes: 8 })
+    const result = await runtime.run({
+      program: `
+        const { parentPort } = await import('node:worker_threads');
+        parentPort.postMessage({ type: 'done', error: { message: '€'.repeat(1000) } });
+        for (;;) {}
+      `,
+      bindings: [],
+    })
+    expect(result.error).toEqual({ kind: 'exception', message: '€€' })
+  })
+
+  it('answers a binding whose resolution cannot be cloned with a failure reply', async () => {
+    const { runtime } = await setup()
+    const result = await runtime.run({
+      program: 'try { await tools.bad({}) } catch (error) { return error.message }',
+      bindings: tools({ bad: async () => (() => 1) }),
+    })
+    expect(result.value).toContain('not structured-cloneable')
+  })
+
+  it('exposes binding names that collide with Object.prototype as ordinary functions', async () => {
+    const { runtime } = await setup()
+    const result = await runtime.run({
+      program: 'return [await tools["__proto__"]({}), await tools["constructor"]({}), typeof tools["hasOwnProperty"]]',
+      // Computed keys: a literal `'__proto__': …` entry would SET the record's
+      // prototype instead of declaring a binding of that name.
+      bindings: tools({ ['__proto__']: async () => 'proto-ok', ['constructor']: async () => 'ctor-ok' }),
+    })
+    expect(result.value).toEqual(['proto-ok', 'ctor-ok', 'undefined'])
+  })
+})
+
+describe('WorkerCodeRuntime — seam misuse and lifecycle', () => {
+  it('rejects invalid binding globals loudly (identifier, reserved word, duplicate, console)', async () => {
+    const { runtime } = await setup()
+    const cases: [string, RegExp][] = [
+      ['not valid!', /not a usable identifier/],
+      ['await', /not a usable identifier/],
+      ['console', /duplicate binding global/],
+    ]
+    for (const [global, message] of cases) {
+      await expect(runtime.run({ program: 'return 1', bindings: [{ global, functions: {} }] })).rejects.toThrow(message)
+    }
+    await expect(runtime.run({
+      program: 'return 1',
+      bindings: [{ global: 'tools', functions: {} }, { global: 'tools', functions: {} }],
+    })).rejects.toThrow(/duplicate binding global/)
+  })
+
+  it('rejects config values that are not positive numbers', async () => {
+    const ctx = new Context()
+    await expect(ctx.plugin(WorkerCodeRuntime, { computeMs: -1 })).rejects.toThrow(/positive number/)
+  })
+
+  it('keeps runs isolated: no state survives from one run to the next', async () => {
+    const { runtime } = await setup()
+    await runtime.run({ program: 'globalThis.leak = "value"; return 1', bindings: [] })
+    const second = await runtime.run({ program: 'return typeof globalThis.leak', bindings: [] })
+    expect(second.value).toBe('undefined')
+  })
+
+  it('disposal aborts in-flight runs, awaits worker exit, and rejects later runs', async () => {
+    const ctx = new Context()
+    const fiber = await ctx.plugin(WorkerCodeRuntime)
+    const runtime = ctx.codeRuntime as WorkerCodeRuntime
+    const inflight: Promise<CodeRunResult> = runtime.run({ program: 'for (;;) {}', bindings: [] })
+    // Give the worker a moment to actually start spinning.
+    await new Promise(resolve => setTimeout(resolve, 200))
+    await fiber.dispose()
+    const result = await inflight
+    expect(result.error).toEqual({ kind: 'abort', message: 'runtime disposed' })
+    await expect(runtime.run({ program: 'return 1', bindings: [] })).rejects.toThrow(/after disposal/)
+  }, 15_000)
+
+  it('removes ctx.codeRuntime when the providing fiber disposes (HMR safety)', async () => {
+    const ctx = new Context()
+    const fiber = await ctx.plugin(WorkerCodeRuntime)
+    expect(ctx.get('codeRuntime')).toBeInstanceOf(WorkerCodeRuntime)
+    await fiber.dispose()
+    expect(ctx.get('codeRuntime')).toBeUndefined()
+  })
+})

packages/code-runtime/code-runtime-worker/tsconfig.json

@@ -0,0 +1,24 @@
+{
+  "extends": "../../../tsconfig.base.json",
+  "compilerOptions": {
+    "rootDir": "src",
+    "outDir": "lib/types"
+  },
+  "include": [
+    "src"
+  ],
+  "references": [
+    {
+      "path": "../../../vendor/cosmokit"
+    },
+    {
+      "path": "../../../vendor/cordis"
+    },
+    {
+      "path": "../../../vendor/schemastery"
+    },
+    {
+      "path": "../code-runtime"
+    }
+  ]
+}

packages/code-runtime/code-runtime-worker/tsdown.config.ts

@@ -0,0 +1,35 @@
+import { defineConfig } from 'tsdown'
+
+/**
+ * Package-shape override (see the root tsdown.config.ts): besides the
+ * default lib/index.js bundle, the worker BOOTSTRAP ships as its own
+ * sibling entry — `new Worker(new URL('./worker.js', import.meta.url))`
+ * loads it as a file, so it cannot be part of the index bundle. TWO
+ * single-entry builds, not one two-entry build: a multi-entry build emits
+ * the shared bootstrap module as a `lib/bootstrap-*.js` chunk both bundles
+ * import, which the package.json `files` whitelist (deliberately exact)
+ * would omit from the packed artifact — each single-entry build inlines its
+ * own bootstrap copy instead, keeping every shipped file self-contained.
+ */
+export default defineConfig([
+  {
+    entry: ['lib/types/index.js'],
+    outDir: 'lib',
+    format: ['esm'],
+    platform: 'node',
+    target: 'es2024',
+    fixedExtension: false,
+    dts: false,
+    clean: false,
+  },
+  {
+    entry: ['lib/types/worker.js'],
+    outDir: 'lib',
+    format: ['esm'],
+    platform: 'node',
+    target: 'es2024',
+    fixedExtension: false,
+    dts: false,
+    clean: false,
+  },
+])

packages/code-runtime/code-runtime/README.md

@@ -0,0 +1,19 @@
+# @deepseek-ai/dsh-code-runtime
+
+The **code-execution seam**: an abstract `CodeRuntime` service (`ctx.codeRuntime`) defining WHAT a code runtime does — run one model-written program against a set of host-provided async bindings and report `{ value, logs, error? }` — without saying HOW.
+
+This package is the interface third of the capability (the bash trio is the template — see [capability seams](../../../docs/rfc/implemented/architecture/2026-06-13-capability-seams.md)): implementations subclass `CodeRuntime` and register the service; the consumer is the tool registry's Code Mode, which generates the model-facing SDK and bridges tool dispatch — both specified in the [Code Mode RFC](../../../docs/rfc/proposed/feature/2026-06-15-code-mode.md), whose first implementation is a Node worker-thread backend. The runtime knows nothing about tools or sessions: it is handed named async functions and a program string, and everything tool-shaped stays with the consumer.
+
+## Service API (`ctx.codeRuntime`)
+
+| Member | Semantics |
+|---|---|
+| `run(request)` | Execute one program against the request's bindings. **Resolves with an error FIELD for every program outcome** — parse/transform failure, thrown exception, budget expiry, abort, substrate death (`CodeRunFailure`'s orthogonal `kind` taxonomy); it rejects only for caller misuse of the seam itself (e.g. a run submitted after disposal). The program runs as the body of an async function: top-level `await`/`return` work, and the completion value becomes `result.value` when it survives the serialization boundary. |
+| `language` | Readonly descriptor: the source language `run` expects (`'typescript'` is the well-known value). Informational, not gating — a consumer that generates language-specific presentation switches on it and fails loud on a language it cannot present. |
+| `isolation` | Readonly descriptor: the execution substrate (`'worker-thread'`, `'process'`, `'container'`). A label for deployments and diagnostics, **not a security claim**. |
+
+Semantics every implementation must honor (contract details in the class JSDoc): binding calls bridge to the caller's functions verbatim with structured-cloneable arguments/resolutions; the program is treated as a hostile peer (arbitrary binding names are own properties, malformed traffic never crashes the host); no state survives between runs; disposal terminates in-flight runs AND awaits their exit before completing.
+
+## Vocabulary
+
+`CodeRunRequest` (`program`, `bindings`, `signal?`) carries everything the runtime acts on — defaulting (time budgets, output caps) is the implementation's validated config, never a hidden `??` inside `run()`. `bindings` is a list of `CodeBindingNamespace`s (`global` + `functions`), each exposed to the program as one global object of async callables. `CodeRunResult` reports the completion `value?`, the ordered `logs` (`CodeLogEntry`: `console`/`stdout`/`stderr` source, console `level`, capped text), and the `error?` (`CodeRunFailure`: `kind` + model-feedable `message`). See `src/types.ts` for the full contracts.

packages/code-runtime/code-runtime/package.json

@@ -0,0 +1,30 @@
+{
+  "name": "@deepseek-ai/dsh-code-runtime",
+  "description": "Abstract code-execution seam (ctx.codeRuntime) for the DeepSeek Harness",
+  "version": "0.0.1",
+  "private": true,
+  "type": "module",
+  "main": "lib/index.js",
+  "types": "lib/types/index.d.ts",
+  "exports": {
+    ".": {
+      "types": "./lib/types/index.d.ts",
+      "default": "./lib/index.js"
+    },
+    "./src/*": "./src/*",
+    "./package.json": "./package.json"
+  },
+  "files": [
+    "lib/index.js",
+    "lib/types/**/*.d.ts",
+    "lib/types/**/*.d.ts.map",
+    "src"
+  ],
+  "license": "BSD-3-Clause",
+  "peerDependencies": {
+    "cordis": "^4.0.0-rc.6"
+  },
+  "devDependencies": {
+    "cordis": "^4.0.0-rc.6"
+  }
+}

packages/code-runtime/code-runtime/src/index.ts

@@ -0,0 +1,93 @@
+/**
+ * The code-execution seam (`ctx.codeRuntime`): an abstract service defining
+ * WHAT a code runtime does — run one model-written program against a set of
+ * host-provided async bindings and report `{ value, logs, error? }` — without
+ * saying HOW. Implementations subclass {@link CodeRuntime} and register
+ * themselves as the `codeRuntime` service; backends may differ by execution
+ * substrate (worker thread, separate process, container) and by source
+ * language, both declared as readonly descriptors. The design and its
+ * consumer (the tool registry's Code Mode) are specified in the Code Mode RFC
+ * (docs/rfc/proposed/feature/2026-06-15-code-mode.md).
+ *
+ * The split mirrors the bash seam (`BashExecutor`): the runtime knows nothing
+ * about tools or sessions — it is handed named async functions and a program,
+ * and everything tool-shaped stays with the consumer.
+ *
+ * @module @deepseek-ai/dsh-code-runtime
+ */
+
+import { Context, Service } from 'cordis'
+import type { CodeRunRequest, CodeRunResult } from './types.ts'
+
+export type {
+  CodeBindingFunction,
+  CodeBindingNamespace,
+  CodeLogEntry,
+  CodeRunFailure,
+  CodeRunRequest,
+  CodeRunResult,
+} from './types.ts'
+
+declare module 'cordis' {
+  interface Context {
+    codeRuntime: CodeRuntime
+  }
+}
+
+/**
+ * Abstract code-execution service. Subclass, implement {@link run} and the
+ * two descriptors, and load the subclass as a plugin — it registers as
+ * `ctx.codeRuntime` (one implementation per context; loading a second throws,
+ * cordis' standard duplicate-service behavior).
+ *
+ * Semantics every implementation must honor:
+ * - {@link run} resolves with an error FIELD for every program outcome —
+ *   parse/transform failures, thrown exceptions, budget expiry, abort,
+ *   substrate death ({@link CodeRunFailure}'s taxonomy). It REJECTS only for
+ *   caller misuse of the seam itself (e.g. a run submitted after disposal).
+ * - Binding calls bridge to the caller's {@link CodeBindingFunction}s
+ *   verbatim; arguments and resolutions must be structured-cloneable, and the
+ *   runtime treats the program as a hostile peer (arbitrary binding names are
+ *   own properties, malformed traffic is rejected or ignored, never crashes
+ *   the host).
+ * - Runs are isolated from each other: no state survives from one run to the
+ *   next through the runtime.
+ * - Disposal reaches quiescence: in-flight runs are terminated AND awaited
+ *   before the service's own teardown completes (no orphan substrate survives
+ *   `fiber.dispose()`).
+ */
+export abstract class CodeRuntime extends Service {
+  /**
+   * The source language {@link run} expects `program` to be written in, as a
+   * lowercase identifier. Informational, not gating — a consumer that
+   * generates language-specific presentation (typed SDK stubs, usage
+   * instructions) switches on it and fails loud on a language it cannot
+   * present. Well-known value: `'typescript'`.
+   */
+  abstract readonly language: string
+
+  /**
+   * The execution substrate, as a lowercase identifier. Informational, not
+   * gating — a descriptor so deployments and diagnostics can tell backends
+   * apart, not a security claim. Well-known values: `'worker-thread'`,
+   * `'process'`, `'container'`.
+   */
+  abstract readonly isolation: string
+
+  constructor(ctx: Context) {
+    super(ctx, 'codeRuntime')
+  }
+
+  /**
+   * Execute one program against the request's bindings and capture what it
+   * emitted. See the class doc for the resolution contract (error is a result
+   * field; rejection means seam misuse only).
+   * @param request - the program, its bindings, and the abort signal; the
+   *   request carries everything the runtime acts on, with no hidden defaults.
+   * @returns the run's outcome: completion value (when transferable), the
+   *   ordered log capture, and the failure (if any).
+   */
+  abstract run(request: CodeRunRequest): Promise<CodeRunResult>
+}
+
+export default CodeRuntime

packages/code-runtime/code-runtime/src/types.ts

@@ -0,0 +1,105 @@
+/**
+ * Vocabulary types for the code-execution seam: what a caller hands a
+ * {@link ../index.ts | CodeRuntime} and what it gets back. Pure types — no
+ * runtime code lives here.
+ *
+ * @module @deepseek-ai/dsh-code-runtime/src/types
+ */
+
+/**
+ * One host-side function exposed to the program as an async callable. The
+ * runtime bridges calls to it (possibly across a serialization boundary), so
+ * `args` and the resolution value MUST be structured-cloneable; a runtime
+ * rejects a non-cloneable value with a descriptive error rather than
+ * corrupting the run. A rejection of this function surfaces inside the
+ * program as a rejection of the corresponding call.
+ */
+export type CodeBindingFunction = (args: unknown) => Promise<unknown>
+
+/**
+ * A named group of {@link CodeBindingFunction}s the runtime exposes to the
+ * program as one global object (e.g. `tools`). Function names are arbitrary
+ * strings — a runtime must treat names like `__proto__` or `constructor` as
+ * ordinary own properties (null-prototype construction), never as prototype
+ * collisions.
+ */
+export interface CodeBindingNamespace {
+  /** The global identifier the program sees (must be a valid JS identifier). */
+  global: string
+  /** The callable members, keyed by the exact name the program calls. */
+  functions: Record<string, CodeBindingFunction>
+}
+
+/**
+ * One run: the program source plus everything the runtime acts on. Per the
+ * explicit-over-implicit convention, defaulting (time budgets, output caps)
+ * is the implementation's validated config — a request carries no optional
+ * tuning knobs for a hidden `??` to fill in.
+ */
+export interface CodeRunRequest {
+  /**
+   * The program source, in the runtime's {@link ../index.ts | language}. It
+   * runs as the body of an async function: top-level `await` and `return`
+   * are available, and the completion value becomes
+   * {@link CodeRunResult.value}.
+   */
+  program: string
+  /** Host functions exposed to the program, one global object per namespace. */
+  bindings: CodeBindingNamespace[]
+  /**
+   * Abort the run: the runtime stops the program (hard, even mid-loop) and
+   * resolves with a {@link CodeRunFailure} of kind `'abort'`. In-flight
+   * binding calls are the CALLER's to settle — the runtime only stops asking.
+   */
+  signal?: AbortSignal
+}
+
+/**
+ * One captured output entry, in emission order. `source` says which channel
+ * produced it: the program's `console` (shimmed by the runtime), or a stray
+ * write to the underlying stdout/stderr streams.
+ */
+export interface CodeLogEntry {
+  /** Which channel produced the text. */
+  source: 'console' | 'stdout' | 'stderr'
+  /** The console method used; present only when `source` is `'console'`. */
+  level?: 'log' | 'info' | 'warn' | 'error' | 'debug'
+  /** The captured text (possibly truncated by the implementation's caps, marked in-band). */
+  text: string
+}
+
+/**
+ * Why a run failed. The kinds are orthogonal outcomes reported independently
+ * (per docs/defensive-patterns.md): a budget expiry is not an exception, an
+ * abort is not a timeout, and a substrate death is neither.
+ *
+ * - `'exception'` — the program threw or failed to parse/transform.
+ * - `'timeout'` — an implementation-owned budget expired; the message says which.
+ * - `'abort'` — {@link CodeRunRequest.signal} fired.
+ * - `'worker-exit'` — the execution substrate died without settling (e.g. OOM).
+ */
+export interface CodeRunFailure {
+  /** The failure class (see the interface doc for each kind's meaning). */
+  kind: 'exception' | 'timeout' | 'abort' | 'worker-exit'
+  /** Human-readable detail, suitable for feeding back to a model to self-correct. */
+  message: string
+}
+
+/**
+ * The outcome of one run. An error is a FIELD on a resolved result, never a
+ * rejection of `run()` — reporting a failed program is the caller's job, not
+ * an exception path.
+ */
+export interface CodeRunResult {
+  /**
+   * The program's completion value (its top-level `return`), when it ran to
+   * completion and the value survived the runtime's serialization boundary;
+   * a non-transferable value is replaced by a string rendering, and a failed
+   * or value-less run leaves this absent.
+   */
+  value?: unknown
+  /** Everything the program emitted, in order (capped by the implementation). */
+  logs: CodeLogEntry[]
+  /** Present iff the run failed; see {@link CodeRunFailure} for the taxonomy. */
+  error?: CodeRunFailure
+}

packages/code-runtime/code-runtime/tests/service.spec.ts

@@ -0,0 +1,87 @@
+import { describe, expect, it } from 'vitest'
+import { Context } from 'cordis'
+import { CodeRuntime } from '@deepseek-ai/dsh-code-runtime'
+import type { CodeRunRequest, CodeRunResult } from '@deepseek-ai/dsh-code-runtime'
+
+/**
+ * Minimal concrete runtime: records requests, "executes" by invoking every
+ * binding once in declaration order, and lets tests script the outcome. The
+ * seam package ships no implementation, so the contract is exercised through
+ * the smallest subclass that honors it.
+ */
+class StubRuntime extends CodeRuntime {
+  readonly language = 'typescript'
+  readonly isolation = 'in-process-stub'
+  requests: CodeRunRequest[] = []
+  nextResult: CodeRunResult = { logs: [] }
+
+  async run(request: CodeRunRequest): Promise<CodeRunResult> {
+    this.requests.push(request)
+    if (request.signal?.aborted) {
+      return { logs: [], error: { kind: 'abort', message: String(request.signal.reason) } }
+    }
+    for (const namespace of request.bindings) {
+      for (const fn of Object.values(namespace.functions)) {
+        await fn({ from: 'stub' })
+      }
+    }
+    return this.nextResult
+  }
+}
+
+async function setup() {
+  const ctx = new Context()
+  await ctx.plugin(StubRuntime)
+  const runtime = ctx.codeRuntime as StubRuntime
+  return { ctx, runtime }
+}
+
+describe('CodeRuntime service seam', () => {
+  it('registers as ctx.codeRuntime and serves the abstract API', async () => {
+    const { runtime } = await setup()
+    expect(runtime.language).toBe('typescript')
+    expect(runtime.isolation).toBe('in-process-stub')
+
+    const calls: unknown[] = []
+    const result = await runtime.run({
+      program: 'return 1',
+      bindings: [{ global: 'tools', functions: { probe: async args => void calls.push(args) } }],
+    })
+    expect(result).toEqual({ logs: [] })
+    expect(calls).toEqual([{ from: 'stub' }])
+    expect(runtime.requests).toHaveLength(1)
+  })
+
+  it('reports a failed run as an error field on a resolved result, never a rejection', async () => {
+    const { runtime } = await setup()
+    runtime.nextResult = {
+      logs: [{ source: 'console', level: 'error', text: 'boom' }],
+      error: { kind: 'exception', message: 'boom' },
+    }
+    const result = await runtime.run({ program: 'throw new Error("boom")', bindings: [] })
+    expect(result.error).toEqual({ kind: 'exception', message: 'boom' })
+    expect(result.value).toBeUndefined()
+  })
+
+  it('reports a pre-aborted signal as an abort failure', async () => {
+    const { runtime } = await setup()
+    const controller = new AbortController()
+    controller.abort('cancelled')
+    const result = await runtime.run({ program: 'return 1', bindings: [], signal: controller.signal })
+    expect(result.error).toEqual({ kind: 'abort', message: 'cancelled' })
+  })
+
+  it('is removed from the context when the providing fiber disposes (HMR safety)', async () => {
+    const ctx = new Context()
+    const fiber = await ctx.plugin(StubRuntime)
+    expect(ctx.get('codeRuntime')).toBeInstanceOf(StubRuntime)
+
+    await fiber.dispose()
+    expect(ctx.get('codeRuntime')).toBeUndefined()
+  })
+
+  it('rejects a second implementation in the same context (duplicate service)', async () => {
+    const { ctx } = await setup()
+    await expect(ctx.plugin(StubRuntime)).rejects.toThrow(/registered/)
+  })
+})

packages/code-runtime/code-runtime/tsconfig.json

@@ -0,0 +1,18 @@
+{
+  "extends": "../../../tsconfig.base.json",
+  "compilerOptions": {
+    "rootDir": "src",
+    "outDir": "lib/types"
+  },
+  "include": [
+    "src"
+  ],
+  "references": [
+    {
+      "path": "../../../vendor/cosmokit"
+    },
+    {
+      "path": "../../../vendor/cordis"
+    }
+  ]
+}

packages/compact/compact-basic/README.md

@@ -1,6 +1,6 @@
 # @deepseek-ai/dsh-compact-basic
 
-The **basic compaction backend**: a `BasicCompactService` implementing the `@deepseek-ai/dsh-compact` seam with a chars-per-token heuristic (the `charsPerToken` config, default 4), token-budget retention, and summarization routed through the agent request pipeline.
+The **basic compaction backend**: a `BasicCompactService` implementing the `@deepseek-ai/dsh-compact` seam with a chars-per-token heuristic (the `charsPerToken` config, default 4), token-budget retention, and summarization as a direct one-shot `ctx.llm.stream()` call (interceptable at `llm/stream`).
 
 This is the implementation tier of the compaction capability — see the [interface package](../compact/README.md) for the seam and the [capability-seam RFC](../../../docs/rfc/implemented/feature/2026-06-18-compaction-capability-seam.md) for the design.
 
@@ -11,13 +11,13 @@ The abstract contract states only WHAT compaction does; this backend owns every
 - **Token estimation** — `estimateContentTokens()`: chars divided by the `charsPerToken` config (default 4) with per-block structural overhead (`text`/`reasoning` = `ceil(len/charsPerToken) + 4`, `tool-call` from name + arguments, `tool-result` recursive, unknown blocks via JSON length).
 - **Retention policy** — `compactIfNeeded()` walks the surface nodes tail→head summing per-node token estimates, and retains the smallest tail-run of WHOLE units (a closed step, or a single no-step node such as a pre-step `user/message` or inter-step `steering/message`) whose total reaches `retainTokens`; everything older is compacted. Retention is **turn-agnostic** — turn boundaries play no role, so a single runaway turn that alone exceeds the window compacts its OWN early closed steps rather than being retained verbatim (the failure mode that motivated dropping turn-protection: a tool-heavy turn must stay compactable or the harness dies exactly when compaction is needed). The only structural guard is **tool-pairing balance**: the compacted region's edges are balanced cuts on the surface (no unanswered tool-call crosses either edge), so it never splits a step's `assistant/message` tool-calls from their `tool/result`s. When the only compactable content left is an un-splittable open tail step, it declines (returns `null`) and retries once an older step closes. **Single-unit overflow is out of scope, by design**: if one retained unit (a single closed step, or a large pasted `user/message`) ALONE exceeds the budget, compaction cannot help and the call may go out over-budget — bounding an individual unit's size is a separate concern. `compactRegion()` enforces tool-pairing balance strictly, throwing on a boundary that would split a step. `dsh-session` exports `isToolPairingBalanced` for the check.
 - **Dynamic convergence** — no static summary-length config pretends to bound what the model will write. If framing/estimator/system overhead leaves the compacted surface above threshold, `compactIfNeeded()` re-compacts the head checkpoint up to `compactionRetries` extra times; if it still cannot get below threshold, it throws. A summary whose estimated stored size is not smaller than the shadowed content fails closed before it mutates the surface.
-- **Summarization** — `summarize()`: a `GenerateOptions` request assembled via `BlockAssembler` with a fixed system prompt that asks for a structured checkpoint (Primary Request and Intent · Key Technical Concepts · Files and Code · Errors and Fixes · Pending Tasks · Current Work · Next Step · Critical Context), every section mandatory, exact paths/commands/identifiers preserved. The request runs through the `agent/request` waterfall before `ctx.llm.stream()`, so router agents that choose the concrete model there also route compaction summaries. `maxTokens` is the provider-side generation cap; only text blocks from the model's reply are kept before the checkpoint is stored (reasoning is dropped so private chain-of-thought never leaks into the durable summary, and a stray `tool-call` is dropped so the synthesized `user/message` summary cannot land an orphaned call with no matching `tool-result`). The compacted region is flattened to a plain-text transcript first: text and reasoning contribute their text, and every non-text block (tool-call, tool-result, plugin-added types) contributes a type-tagged placeholder (`[tool-call: name(args)]`, `[tool-result: …]`, …) so the summarizer is told what existed rather than silently dropping it.
+- **Summarization** — `summarize()`: a `GenerateOptions` request assembled via `BlockAssembler` with a fixed system prompt that asks for a structured checkpoint (Primary Request and Intent · Key Technical Concepts · Files and Code · Errors and Fixes · Pending Tasks · Current Work · Next Step · Critical Context), every section mandatory, exact paths/commands/identifiers preserved. The request is a direct one-shot `ctx.llm.stream()` call — NOT a loop step, so it does not run `agent/request` (that seam shapes the loop's conversation requests); the model comes from `summarizationModel` falling back to the agent's own, and per-call routing happens at `llm/stream` like any other direct call. `maxTokens` is the provider-side generation cap; only text blocks from the model's reply are kept before the checkpoint is stored (reasoning is dropped so private chain-of-thought never leaks into the durable summary, and a stray `tool-call` is dropped so the synthesized `user/message` summary cannot land an orphaned call with no matching `tool-result`). The compacted region is flattened to a plain-text transcript first: text and reasoning contribute their text, and every non-text block (tool-call, tool-result, plugin-added types) contributes a type-tagged placeholder (`[tool-call: name(args)]`, `[tool-result: …]`, …) so the summarizer is told what existed rather than silently dropping it.
 - **Checkpoint framing** — the raw summary is not landed directly. `compactRegion()` wraps it in a checkpoint preamble (so a resuming model reads it as a checkpoint, not a fresh user request, and builds on the captured context rather than restating it) plus `<compacted-summary>…</compacted-summary>` tags. Because region compaction can be invoked manually, a surface may hold several checkpoints, so the framing does not claim everything after it is recent or verbatim. The tags make a prior checkpoint detectable in the transcript on the next compaction cycle: the summarization prompt then instructs the model to merge it in place (preserve still-true facts, drop stale ones) rather than re-summarize it verbatim — a cheap incremental merge that needs no extra log/event machinery. The unframed summary stays on the `compact/summary` provenance event.
 - **Surface mutation** — `compactRegion()` appends the `compact/start` → `compact/summary` → `compact/end` log records and the single `user/message` replace node carrying the framed summary (see the interface README).
 - **Auto-compaction** — an `agent/pre-step` listener delegates to `compactIfNeeded()` before every step (not just a turn's first — a tool-heavy turn grows the surface mid-turn, so a runaway turn still compacts, and per-step firing is the only moment to rescue it before overflow). `agent/pre-step` is a serial (awaited, in-order) surface-mutation checkpoint that fires after `turn/start` and BEFORE the step opens (`step/start`) and its request history is derived, so compaction mutates the surface — with its log-only `compact/*` records landing cleanly outside any step — and the loop derives once from the result: no double-derive, and the listener cannot see (or need to rewrite) an already-assembled `messages` array. The listener owns no threshold logic of its own (the single token-pressure check lives in `compactIfNeeded()`); because Cordis `serial` bails early on non-void return values, the listener returns `void` and does not use the dispatcher's bail channel as a veto surface.
 - **Failure handling** — the `compact/start … compact/end` bracket is a log-recorded lock: it makes a crash mid-summarization a detectable orphan (a `compact/start` with no `compact/end`), records provenance, and prevents a concurrent compaction. Two failure paths: a **crash** (the loop dies mid-summarization) leaves a dangling `compact/start` that is inert — `compact/*` events are log-only, the surface replacement never landed, so the full history derives fine and generic turn-repair closes the turn; a **recoverable** failure (summarization throws but the loop survives) appends `compact/end` with its `error` field set, leaving the surface untouched so the call proceeds with full history. Core session repair stays compaction-agnostic by design — it never learns about `compact/*`.
 
-`estimateContentTokens()` and `summarize()` are overridable hooks: a tokenizer-based or template-based backend can subclass `BasicCompactService` and override just those, reusing the retention walk and surface plumbing.
+`estimateContentTokens()` and `summarize()` are overridable hooks: a tokenizer-based or template-based backend can subclass `BasicCompactService` and override just those, reusing the retention walk and surface plumbing. `summarize()` returns the summary blocks together with the call envelope it actually used (`{ summary, model, maxTokens? }`) — the caller logs that envelope on the `compact/summary` provenance event, so an overriding backend reports its own envelope honestly.
 
 ## Config (`BasicCompactConfig`)
 

packages/compact/compact-basic/src/index.ts

@@ -9,9 +9,10 @@
  *   to the next balanced tool-pairing boundary so a compacted region never
  *   splits a step's tool-call/result pair (an open tail step is never crossed —
  *   compaction declines and retries once it closes).
- * - **Summarization** — `ctx.llm.stream()` assembled via `BlockAssembler`
- *   (the single model-call surface; same path the loop uses) with a fixed
- *   condense-the-history system prompt routed through `agent/request`.
+ * - **Summarization** — a direct one-shot `ctx.llm.stream()` call assembled
+ *   via `BlockAssembler` with a fixed condense-the-history system prompt;
+ *   NOT a loop step, so `agent/request` never fires — interception happens
+ *   at `llm/stream` like any other direct call.
  * - **Surface mutation** — a single `user/message` replace node carries the
  *   summary; `compact/*` events are log-only lock + provenance records.
  * - **Auto-compaction** — an `agent/pre-step` listener delegates to
@@ -182,9 +183,9 @@ export class BasicCompactService extends CompactService {
       // log-only `compact/*` records and the replacement node cleanly outside a
       // step, so a crash mid-compaction leaves an inert orphan the turn-repair
       // closes — never a half-open step.
-      ctx.on('agent/pre-step', async (agent: Agent, turn: number, step: number, fullSystemPrompt: string, signal: AbortSignal) => {
+      ctx.on('agent/pre-step', async (agent: Agent, _turn: number, _step: number, fullSystemPrompt: string, signal: AbortSignal) => {
         try {
-          const result = await this.compactIfNeeded(agent, turn, step, fullSystemPrompt, signal)
+          const result = await this.compactIfNeeded(agent, fullSystemPrompt, signal)
           if (result) {
             const after = this.estimateTokens(agent.session.deriveMessages(), fullSystemPrompt)
             ctx.logger.info(
@@ -215,6 +216,11 @@ export class BasicCompactService extends CompactService {
    * Estimate the token count of content blocks — chars divided by the
    * `charsPerToken` config, with per-block overhead. Override in a subclass to
    * plug in a real tokenizer.
+   *
+   * @param blocks - the blocks to estimate; `tool-result` blocks recurse into
+   *   their nested content, and unknown (merge-extended) types fall back to
+   *   their JSON-stringified length.
+   * @returns the estimated token count.
    */
   estimateContentTokens(blocks: readonly ContentBlock[]): number {
     const { charsPerToken } = this.config
@@ -245,6 +251,11 @@ export class BasicCompactService extends CompactService {
   /**
    * Estimate token count for a single session event. Returns 0 for non-message
    * event types (boundaries, chunks, usage, errors, compact markers).
+   *
+   * @param event - any session event; only the message-bearing types carry
+   *   content to count.
+   * @returns the estimated token count of the event's content, or 0 for a
+   *   non-message event.
    */
   estimateEventTokens(event: SessionEvent): number {
     switch (event.type) {
@@ -259,7 +270,14 @@ export class BasicCompactService extends CompactService {
     }
   }
 
-  /** Estimate total tokens across a list of messages plus optional system prompt. */
+  /**
+   * Estimate total tokens across a list of messages plus optional system prompt.
+   *
+   * @param messages - the derived conversation messages; each adds a fixed
+   *   role-framing overhead on top of its content estimate.
+   * @param systemPrompt - counted at chars / `charsPerToken` when provided.
+   * @returns the estimated token footprint of the whole request.
+   */
   estimateTokens(messages: readonly Message[], systemPrompt?: string): number {
     let total = 0
     for (const msg of messages) {
@@ -271,9 +289,13 @@ export class BasicCompactService extends CompactService {
   }
 
   /**
-   * Summarize conversation text into content blocks via `agent/request` plus
-   * `ctx.llm.stream()` assembled through a `BlockAssembler` (the single
-   * model-call surface).
+   * Summarize conversation text into content blocks via `ctx.llm.stream()`
+   * assembled through a `BlockAssembler`. A direct one-shot model call, NOT a
+   * loop step: it does not run the `agent/request` waterfall (that seam shapes
+   * the loop's conversation requests); per-call
+   * interception happens at `llm/stream` like any other direct call. The model
+   * comes from `BasicCompactConfig.summarizationModel`, falling back to the
+   * agent's own model.
    * Override in a subclass for a template or remote summarizer.
    *
    * Honors the adapter failure contract: an adapter may report a model failure
@@ -283,8 +305,22 @@ export class BasicCompactService extends CompactService {
    *
    * Forwards `signal` into `GenerateOptions.signal` so an abort/dispose tears
    * down the in-flight summarization rather than orphaning the model call.
+   *
+   * Returns the summary blocks TOGETHER with the call envelope it actually
+   * used (`model`, `maxTokens`) — the caller logs the envelope on the
+   * `compact/summary` provenance event, so an overriding subclass (template
+   * or remote summarizer) reports its own envelope honestly.
+   *
+   * @param text - plain-text rendering of the conversation region to condense.
+   * @param agent - supplies the fallback model and the session id stamped on
+   *   the call; throws when neither it nor the config names a model.
+   * @param signal - optional abort signal, forwarded into the model call.
+   * @returns the text-only summary blocks plus the call envelope used
+   *   (`model`, and `maxTokens` when the summarizer has a cap).
    */
-  async summarize(text: string, agent: Agent, turn: number, step: number, signal?: AbortSignal): Promise<ContentBlock[]> {
+  async summarize(
+    text: string, agent: Agent, signal?: AbortSignal,
+  ): Promise<{ summary: ContentBlock[]; model: string; maxTokens?: number }> {
     const assembler = new BlockAssembler()
     const options: GenerateOptions = {
       model: this.config.summarizationModel || agent.options.model || '',
@@ -299,11 +335,10 @@ export class BasicCompactService extends CompactService {
     // exactOptionalPropertyTypes: only set `signal` when present — assigning
     // `undefined` to an optional `signal?: AbortSignal` is a type error.
     if (signal) options.signal = signal
-    const request = await this.ctx.waterfall('agent/request', agent, turn, step, options, () => Promise.resolve(options))
-    if (!request.model) {
-      throw new Error('no model available for summarization: set BasicCompactConfig.summarizationModel, AgentOptions.model, or supply one via the agent/request waterfall')
+    if (!options.model) {
+      throw new Error('no model available for summarization: set BasicCompactConfig.summarizationModel or AgentOptions.model')
     }
-    for await (const chunk of this.ctx.llm.stream(request)) {
+    for await (const chunk of this.ctx.llm.stream(options)) {
       assembler.push(chunk)
     }
 
@@ -315,7 +350,10 @@ export class BasicCompactService extends CompactService {
       throw new Error('summarization produced no text summary content')
     }
 
-    return summary
+    // config.maxTokens is required and validated positive, so this backend's
+    // envelope always carries the cap; the return type's optionality exists
+    // for overriding subclasses whose summarizer has none.
+    return { summary, model: options.model, maxTokens: this.config.maxTokens }
   }
 
   // ---- Core API (implements the abstract contract) ----
@@ -348,8 +386,6 @@ export class BasicCompactService extends CompactService {
    */
   override async compactIfNeeded(
     agent: Agent,
-    turn: number,
-    step: number,
     fullSystemPrompt: string,
     signal: AbortSignal,
   ): Promise<CompactionResult | null> {
@@ -368,7 +404,7 @@ export class BasicCompactService extends CompactService {
         break
       }
 
-      result = await this.compactRegion(session, range.start, range.end, agent, turn, step, signal)
+      result = await this.compactRegion(session, range.start, range.end, agent, signal)
     }
 
     const totalTokens = this.estimateTokens(session.deriveMessages(), fullSystemPrompt)
@@ -385,8 +421,6 @@ export class BasicCompactService extends CompactService {
     start: number,
     end: number,
     agent: Agent,
-    turn: number,
-    step: number,
     signal?: AbortSignal,
   ): Promise<CompactionResult> {
     // Resolve the range by surface POSITION, not numeric seq interval. A prior
@@ -450,7 +484,7 @@ export class BasicCompactService extends CompactService {
     try {
       // --- Extract text and summarize ---
       const text = this._extractText(session, shadowedSeqs)
-      const summary = await this.summarize(text, agent, turn, step, signal)
+      const { summary, model, maxTokens } = await this.summarize(text, agent, signal)
 
       // Estimate token count of the shadowed content for provenance.
       let shadowedTokenCount = 0
@@ -472,6 +506,8 @@ export class BasicCompactService extends CompactService {
         shadowedRange: { start, end },
         shadowedSeqs,
         shadowedTokenCount,
+        model,
+        ...maxTokens !== undefined ? { maxTokens } : {},
       })
 
       // --- Surface replacement ---

packages/compact/compact-basic/src/types.ts

@@ -54,6 +54,9 @@ export type ResolvedConfig = Required<BasicCompactConfig>
  * each committed summary must be smaller than the content it shadows, and
  * `compactIfNeeded` may re-compact up to `compactionRetries` extra times before
  * throwing if the surface still exceeds the threshold.
+ *
+ * @param config - the raw, unresolved backend config.
+ * @returns the validated config with `auto` and `charsPerToken` defaulted.
  */
 export function resolveConfig(config: BasicCompactConfig): ResolvedConfig {
   const resolved: ResolvedConfig = { auto: true, charsPerToken: 4, ...config }

packages/compact/compact-basic/tests/compact-basic.spec.ts

@@ -58,13 +58,13 @@ class TestCompactService extends BasicCompactService {
     return blocks.length * 10
   }
 
-  override async summarize(text: string, agent: Agent): Promise<ContentBlock[]> {
+  override async summarize(text: string, agent: Agent): Promise<{ summary: ContentBlock[]; model: string; maxTokens?: number }> {
     const model = this.config.summarizationModel || agent.options.model || ''
     this.summarizeCalls.push({ text, model })
     if (this.summarizeError) throw this.summarizeError
     const summary = this.mockSummaryQueue.shift() ?? this.mockSummary
     this.summaryOutputs.add(summary)
-    return summary
+    return { summary, model }
   }
 }
 
@@ -402,6 +402,9 @@ describe('BasicCompactService.compactRegion', () => {
     expect(startEvent).toBeDefined()
     expect(summaryEvent).toBeDefined()
     expect(endEvent).toBeDefined()
+    // The provenance record carries the summarize call's envelope, so "which
+    // model wrote this summary" is answerable from the log alone.
+    expect(summaryEvent?.type === 'compact/summary' && summaryEvent.data.model).toBe('test-model')
 
     // compact/* events are log-only — no surfaceOp (type system enforces this).
     const startRaw = startEvent as unknown as { surfaceOp?: unknown }
@@ -980,7 +983,7 @@ function compactIfNeeded(
   model: string,
   signal: AbortSignal,
 ) {
-  return svc.compactIfNeeded(stubAgent(session, model), 1, 1, fullSystemPrompt, signal)
+  return svc.compactIfNeeded(stubAgent(session, model), fullSystemPrompt, signal)
 }
 
 function compactRegion(
@@ -991,11 +994,11 @@ function compactRegion(
   model: string,
   signal?: AbortSignal,
 ) {
-  return svc.compactRegion(session, start, end, stubAgent(session, model), 1, 1, signal)
+  return svc.compactRegion(session, start, end, stubAgent(session, model), signal)
 }
 
 function summarize(svc: BasicCompactService, text: string, model: string) {
-  return svc.summarize(text, stubAgent(new Session(SessionId('summary')), model), 1, 1)
+  return svc.summarize(text, stubAgent(new Session(SessionId('summary')), model))
 }
 
 describe('BasicCompactService.summarize (real ctx.llm.stream)', () => {
@@ -1003,8 +1006,12 @@ describe('BasicCompactService.summarize (real ctx.llm.stream)', () => {
     const { ctx, adapter } = await ctxWithModel('SUMMARY TEXT')
     const svc = new BasicCompactService(ctx, cfg({ auto: false, maxTokens: 512 }))
 
-    const summary = await summarize(svc, 'User: hi\n\nAssistant: hello', 'test-model')
+    const { summary, model, maxTokens } = await summarize(svc, 'User: hi\n\nAssistant: hello', 'test-model')
     expect(summary).toEqual([{ type: 'text', text: 'SUMMARY TEXT' }])
+    // The returned envelope reports what the call actually used — the caller
+    // logs it on compact/summary (the reconstructability RFC).
+    expect(model).toBe('test-model')
+    expect(maxTokens).toBe(512)
     // The fixed system prompt and maxTokens flow through.
     expect(adapter.lastOptions!.system).toContain('compaction engine')
     expect(adapter.lastOptions!.system).toContain('## Next Step')
@@ -1035,7 +1042,7 @@ describe('BasicCompactService.summarize (real ctx.llm.stream)', () => {
     ])
     const svc = new BasicCompactService(ctx, cfg({ auto: false }))
 
-    const summary = await summarize(svc, 'User: hi', 'test-model')
+    const { summary } = await summarize(svc, 'User: hi', 'test-model')
 
     expect(summary).toEqual([{ type: 'text', text: 'PUBLIC SUMMARY' }])
   })
@@ -1231,15 +1238,20 @@ describe('BasicCompactService auto-compaction (agent/pre-step listener)', () =>
     expect(session.events.some(e => e.type === 'compact/start')).toBe(false)
   })
 
-  it('routes summarization through agent/request so router agents can choose the model', async () => {
+  it('summarization is interceptable at llm/stream (model routing for direct calls)', async () => {
     const { ctx, adapter } = await ctxWithModel('ROUTED SUMMARY', 'routed-model')
-    ctx.on('agent/request', async (_agent, _turn, _step, options, next) => {
+    // The summarize call is a direct one-shot model call, not a loop step: it
+    // does not run agent/request (that seam shapes the loop's conversation
+    // requests). llm/stream is its interception surface, and a hand-built
+    // request is not frozen, so mutate-then-next model routing works — the
+    // adapter resolves AFTER the waterfall, so the rewrite picks the adapter.
+    ctx.on('llm/stream', (options, next) => {
       options.model = 'routed-model'
       return next()
     })
     void new BasicCompactService(ctx, cfg({ contextWindow: 200, thresholdRatio: 0.5, retainTokens: 20 }))
     const session = multiTurnSession(5, 1)
-    const agent = stubAgent(session)
+    const agent = stubAgent(session, 'agent-model')
 
     await ctx.serial('agent/pre-step', agent, 1, 1, '', SIGNAL)
 

packages/compact/compact-basic/tests/compact-loop-repro.spec.ts

@@ -41,8 +41,8 @@ class ReproCompactService extends BasicCompactService {
     return blocks.length * TOKENS_PER_BLOCK
   }
 
-  override async summarize(): Promise<ContentBlock[]> {
-    return [{ type: 'text', text: 'CHECKPOINT SUMMARY' }]
+  override async summarize(): Promise<{ summary: ContentBlock[]; model: string }> {
+    return { summary: [{ type: 'text', text: 'CHECKPOINT SUMMARY' }], model: 'stub' }
   }
 }
 

packages/compact/compact/README.md

@@ -18,8 +18,8 @@ Both methods are **abstract** — the backend owns the entire strategy (token es
 
 | Member | Semantics |
 |---|---|
-| `compactIfNeeded(agent, turn, step, fullSystemPrompt, signal)` | Estimate the surface-derived history size; if over the backend's threshold, compact an older range via `compactRegion`, keeping recent context intact. Returns the `CompactionResult`, or `null` if nothing needed compacting. All parameters required — the loop's `agent/pre-step` checkpoint supplies the agent, lifecycle context, assembled `fullSystemPrompt`, and turn `signal`; router-aware summarizers can use the agent lifecycle context to route their own model call through `agent/request`. |
-| `compactRegion(session, start, end, agent, turn, step, signal?)` | Forcibly summarize surface nodes `[start, end]` (inclusive seqs) into a single replacement node. **Throws** if a compaction is already in progress, if `start`/`end` aren't surface nodes, or if `start` is positioned after `end` on the surface. The range is a SURFACE-POSITION span, not a numeric seq interval — after a prior replace lands a fresh high-seq summary node at the shadowed range's position, surface order no longer tracks seq order. |
+| `compactIfNeeded(agent, fullSystemPrompt, signal)` | Estimate the surface-derived history size; if over the backend's threshold, compact an older range via `compactRegion`, keeping recent context intact. Returns the `CompactionResult`, or `null` if nothing needed compacting. All parameters required — the loop's `agent/pre-step` checkpoint supplies the agent, assembled `fullSystemPrompt`, and turn `signal`. A backend's summarization request is a direct `ctx.llm.stream()` call (not a loop step), so per-call interception happens at `llm/stream`. |
+| `compactRegion(session, start, end, agent, signal?)` | Forcibly summarize surface nodes `[start, end]` (inclusive seqs) into a single replacement node. **Throws** if a compaction is already in progress, if `start`/`end` aren't surface nodes, or if `start` is positioned after `end` on the surface. The range is a SURFACE-POSITION span, not a numeric seq interval — after a prior replace lands a fresh high-seq summary node at the shadowed range's position, surface order no longer tracks seq order. |
 
 `compactIfNeeded` takes a required `signal`; `compactRegion`'s is optional. A backend that summarizes via `ctx.llm.stream()` **must** forward it into the call's `GenerateOptions.signal`, so an abort or fiber dispose tears down the in-flight summarization instead of leaving an orphaned model call running past the cancellation. The session being compacted comes from the agent context; the turn that the `compact/*` events belong to is recoverable from the log (the currently-open turn), so the backend stamps it from the log rather than trusting a caller-supplied value.
 
@@ -43,7 +43,7 @@ Compaction is serialized via a log-recorded lock: `compactRegion` refuses to sta
 
 ## Events
 
-The `compact/*` events extend `SessionEventMap` (merge-extensible) via declaration merging — they are session events, not cordis `Events`, and all three are log-only (no `surfaceOp`). Per-event payloads and semantics are in the generated [persistence log event catalog](../../../docs/persistence-catalog/log-events.md).
+The `compact/*` events extend `SessionEventMap` (merge-extensible) via declaration merging — they are session events, not cordis `Events`, and all three are log-only (no `surfaceOp`). Per-event payloads and semantics are in the generated [persistence log event catalog](../../../docs/persistence-catalog.md).
 
 ## Implementing a backend
 

packages/compact/compact/src/index.ts

@@ -91,8 +91,6 @@ export abstract class CompactService extends Service {
    *   over-budget. Bounding an individual unit's size is a separate concern.
    *
    * @param agent - agent context owning the session surface and model options.
-   * @param turn - turn number of the pre-step checkpoint.
-   * @param step - step number about to start.
    * @param fullSystemPrompt - assembled system prompt, counted toward the estimate.
    * @param signal - cancellation signal. A backend summarizing via
    *   `ctx.llm.stream()` MUST forward this into the call's `GenerateOptions.signal`
@@ -102,8 +100,6 @@ export abstract class CompactService extends Service {
    */
   abstract compactIfNeeded(
     agent: CompactAgentContext,
-    turn: number,
-    step: number,
     fullSystemPrompt: string,
     signal: AbortSignal,
   ): Promise<CompactionResult | null>
@@ -129,8 +125,6 @@ export abstract class CompactService extends Service {
    * @param start - inclusive seq of the first surface node to compact.
    * @param end - inclusive seq of the last surface node to compact.
    * @param agent - agent context used by router-aware summarizers.
-   * @param turn - lifecycle turn forwarded to request-routing seams.
-   * @param step - lifecycle step forwarded to request-routing seams.
    * @param signal - optional cancellation signal. A backend that summarizes via
    *   `ctx.llm.stream()` MUST forward this into the call's `GenerateOptions.signal`
    *   so an abort/dispose tears down the in-flight summarization rather than
@@ -148,8 +142,6 @@ export abstract class CompactService extends Service {
     start: number,
     end: number,
     agent: CompactAgentContext,
-    turn: number,
-    step: number,
     signal?: AbortSignal,
   ): Promise<CompactionResult>
 }

packages/compact/compact/src/types.ts

@@ -32,6 +32,15 @@ declare module '@deepseek-ai/dsh-session' {
       shadowedRange: { start: number; end: number }
       shadowedSeqs: number[]
       shadowedTokenCount: number
+      /**
+       * The model that wrote the summary — the summarize call's envelope,
+       * reported by the backend that made the call, logged so the one-shot
+       * request is reconstructable from log + code and "which model wrote
+       * this summary" has a durable answer (the reconstructability RFC).
+       */
+      model: string
+      /** The generation cap the summarize call sent, when one applied. */
+      maxTokens?: number
     }
     /** Marks the end of a compaction — log-only, releases the lock. `error` set if summarization failed. */
     'compact/end': { turn: number; error?: string }

packages/compact/compact/tests/compact.spec.ts

@@ -17,8 +17,6 @@ class StubCompactService extends CompactService {
 
   override async compactIfNeeded(
     _agent: CompactAgentContext,
-    _turn: number,
-    _step: number,
     _fullSystemPrompt: string,
     signal: AbortSignal,
   ): Promise<CompactionResult | null> {
@@ -31,8 +29,6 @@ class StubCompactService extends CompactService {
     start: number,
     end: number,
     _agent: CompactAgentContext,
-    _turn: number,
-    _step: number,
     signal?: AbortSignal,
   ): Promise<CompactionResult> {
     this.lastSignal = signal
@@ -43,6 +39,7 @@ class StubCompactService extends CompactService {
       shadowedRange: { start, end },
       shadowedSeqs: [],
       shadowedTokenCount: 0,
+      model: 'stub',
     })
     const endEvent = session.append('compact/end', { turn: 0 })
     return {
@@ -81,7 +78,7 @@ describe('CompactService seam', () => {
     const ctx = new Context()
     const svc = new StubCompactService(ctx)
     const session = new Session(SessionId('s'))
-    expect(await svc.compactIfNeeded(stubAgent(session), 1, 1, '', new AbortController().signal)).toBeNull()
+    expect(await svc.compactIfNeeded(stubAgent(session), '', new AbortController().signal)).toBeNull()
   })
 
   it('compact/* events merge into SessionEventMap and are log-only', async () => {
@@ -89,7 +86,7 @@ describe('CompactService seam', () => {
     const svc = new StubCompactService(ctx)
     const session = new Session(SessionId('s'))
 
-    const result = await svc.compactRegion(session, 0, 0, stubAgent(session, 'm'), 1, 1)
+    const result = await svc.compactRegion(session, 0, 0, stubAgent(session, 'm'))
 
     const startEvent = session.events.find(e => e.type === 'compact/start')
     expect(startEvent).toBeDefined()
@@ -107,10 +104,10 @@ describe('CompactService seam', () => {
     const session = new Session(SessionId('s'))
     const controller = new AbortController()
 
-    await svc.compactRegion(session, 0, 0, stubAgent(session, 'm'), 1, 1, controller.signal)
+    await svc.compactRegion(session, 0, 0, stubAgent(session, 'm'), controller.signal)
     expect(svc.lastSignal).toBe(controller.signal)
 
-    await svc.compactIfNeeded(stubAgent(session), 1, 1, '', controller.signal)
+    await svc.compactIfNeeded(stubAgent(session), '', controller.signal)
     expect(svc.lastSignal).toBe(controller.signal)
   })
 })

packages/core/agent-core/README.md

@@ -35,11 +35,11 @@ This is the [interface/implementation/consumer seam](../../../docs/rfc/implement
 
 ```ts
 import type { Config } from '@deepseek-ai/dsh-agent-core'
-// { agents?, persona? } — the schema is z.intersect([AgentLoop.Config, SystemPrompt.Config]),
+// { agents?, persona?, toolOrder? } — the schema is z.intersect([AgentLoop.Config, SystemPrompt.Config]),
 // so validation and defaulting can never drift from the owners'.
 ```
 
-The bundle FORWARDS each field to the child that owns it: `agents` to `agent-loop` (default `[]`), so each app supplies its own pre-created agents — a stdio app pre-creates a `main`; the ACP app pre-creates none (it creates agents on demand at `session/new`) — and `persona` to `dsh-system-prompt` (default `''`), the deployment's persona section. Forwarding is exactly why the owners can live in the shared spine even though the apps disagree on what to configure.
+The bundle FORWARDS each field to the child that owns it: `agents` to `agent-loop` (default `[]`), so each app supplies its own pre-created agents — a stdio app pre-creates a `main`; the ACP app pre-creates none (it creates agents on demand at `session/new`) — `persona` to `dsh-system-prompt` (default `''`), the deployment's persona section — and `toolOrder` to `dsh-system-prompt` (absent — lexicographic), the explicit model-facing tool order. Forwarding is exactly why the owners can live in the shared spine even though the apps disagree on what to configure.
 
 ## Why a code bundle, not a shared YAML include
 

packages/core/agent-core/src/index.ts

@@ -59,17 +59,20 @@ export const name = 'agent-core'
 /**
  * Bundle config: each field forwarded verbatim to the child that owns it —
  * `agents` to the agent loop (an app that pre-creates no agents, like the ACP
- * bridge, simply omits it), `persona` to the system-prompt plugin (the
- * deployment's persona section). Both are optional INPUT here because each
- * owner's schema supplies the default (`[]` / `''`); the schema is the
- * INTERSECTION of the owners' own schemas, so validation and defaulting can
- * never drift from them.
+ * bridge, simply omits it), `persona` and `toolOrder` to the system-prompt
+ * plugin (the deployment's persona section and the explicit model-facing tool
+ * order). Every field is optional INPUT here because each owner's schema
+ * supplies the default (`[]` / `''` / absent — lexicographic); the schema is
+ * the INTERSECTION of the owners' own schemas, so validation and defaulting
+ * can never drift from them.
  */
 export interface Config {
   /** The agent-loop `agents` list (see dsh-agent-loop's `Config`). */
   agents?: AgentLoopConfig['agents']
   /** The deployment persona (see dsh-system-prompt's `Config`). */
   persona?: SystemPromptConfig['persona']
+  /** The explicit model-facing tool order (see dsh-system-prompt's `Config`). */
+  toolOrder?: SystemPromptConfig['toolOrder']
 }
 
 /** Intersect the owners' schemas so validation + defaulting stay identical. */
@@ -78,11 +81,11 @@ export const Config = z.intersect([AgentLoop.Config, SystemPrompt.Config]) as un
 /**
  * Load the spine. Each `ctx.plugin(...)` mounts one child of the bundle fiber;
  * `agent-loop` receives the forwarded `agents` list and `system-prompt` the
- * forwarded `persona`. Load order is irrelevant (cordis pends each fiber on
- * its `inject` until the services it needs exist), but the listing mirrors the
- * dependency layering for readability: the LLM vocabulary and core registries
- * first, then the dev tripwire and the bash tool consumer, then the loop that
- * drives them.
+ * forwarded `persona` and `toolOrder`. Load order is irrelevant (cordis pends
+ * each fiber on its `inject` until the services it needs exist), but the
+ * listing mirrors the dependency layering for readability: the LLM vocabulary
+ * and core registries first, then the dev tripwire and the bash tool consumer,
+ * then the loop that drives them.
  */
 export function apply(ctx: Context, config: Config): void {
   ctx.plugin(Timer)
@@ -91,8 +94,13 @@ export function apply(ctx: Context, config: Config): void {
   // The forwarded fields are validated + defaulted by this bundle's intersected
   // schema before apply runs, so the ?? fallbacks only narrow the
   // optional-input TYPES — they mirror the owners' schema defaults, never
-  // introduce different ones.
-  ctx.plugin(SystemPrompt, { persona: config.persona ?? '' })
+  // introduce different ones. toolOrder has no owner-supplied default value —
+  // ABSENT means "lexicographic order" — so it is forwarded conditionally
+  // rather than via ??.
+  ctx.plugin(SystemPrompt, {
+    persona: config.persona ?? '',
+    ...config.toolOrder !== undefined ? { toolOrder: config.toolOrder } : {},
+  })
   ctx.plugin(ToolRegistry)
   ctx.plugin(AgentRegistry)
   ctx.plugin(invariants)

packages/core/agent-core/tests/agent-core.spec.ts

@@ -1,6 +1,7 @@
 import { describe, expect, it } from 'vitest'
 import { Context } from 'cordis'
 import Loader from '@cordisjs/plugin-loader'
+import { TOOL_ORDER_REST } from '@deepseek-ai/dsh-system-prompt'
 import * as agentCore from '../src/index.ts'
 import { AgentId } from '@deepseek-ai/dsh-agent'
 
@@ -67,6 +68,23 @@ describe('dsh-agent-core bundle', () => {
     await ctx.fiber.dispose()
   })
 
+  it('forwards toolOrder to the system-prompt assembly', async () => {
+    const ctx = await mount({ toolOrder: ['zulu', TOOL_ORDER_REST] })
+    // The bundle's own bash tools pend on the absent `ctx.bash` executor in
+    // this providerless mount, so register two plain tools to order.
+    for (const name of ['alpha', 'zulu']) {
+      ctx.get('tools')!.register({
+        name,
+        description: name,
+        parameters: {},
+        execute: async () => [],
+      })
+    }
+    const assembly = await ctx.get('systemPrompt')!.assemble()
+    expect(assembly.tools.map(tool => tool.name)).toEqual(['zulu', 'alpha'])
+    await ctx.fiber.dispose()
+  })
+
   it('re-exports the loop config schema as its own', () => {
     expect(agentCore.Config).toBeDefined()
     expect(agentCore.name).toBe('agent-core')

packages/core/agent-core/tests/gen-config-catalog.spec.ts

@@ -0,0 +1,458 @@
+/**
+ * Negative-path tests for the config catalog generator (`scripts/gen-config-catalog.ts`).
+ *
+ * The generated catalog is frozen by a regenerate-and-diff freshness gate, so
+ * the freshness half is exercised by `pnpm run verify-config-catalog` in CI.
+ * What a freshness diff CANNOT prove is that the generator REJECTS malformed
+ * source the way it promises to — an unclassifiable package, an undocumented
+ * config field, a schema key the config type does not declare, or a referenced
+ * type name that resolves nowhere. These tests drive `collectConfigCatalog()`
+ * against synthetic fixture packages to prove each guard fires (and that
+ * well-formed packages classify and extract correctly), mirroring the
+ * negative tests for gen-cordis-catalog. The spec lives in this package
+ * because agent-core is the config-composition plugin (its schema is the
+ * intersection of its children's), the shape the generator's cross-package
+ * folding exists for.
+ */
+
+import { mkdtempSync, mkdirSync, rmSync, writeFileSync } from 'node:fs'
+import { tmpdir } from 'node:os'
+import { join } from 'node:path'
+import { afterEach, describe, expect, it } from 'vitest'
+import { collectConfigCatalog, render } from '../../../../scripts/gen-config-catalog.ts'
+
+/** Write one fixture package (package.json + src files) under a scan root. */
+function writePkg(root: string, dir: string, name: string, files: Record<string, string>): void {
+  const pkgDir = join(root, 'packages', dir)
+  mkdirSync(join(pkgDir, 'src'), { recursive: true })
+  writeFileSync(join(pkgDir, 'package.json'), JSON.stringify({ name }))
+  for (const [rel, text] of Object.entries(files)) writeFileSync(join(pkgDir, rel), text)
+}
+
+const roots: string[] = []
+const makeRoot = (): string => {
+  const root = mkdtempSync(join(tmpdir(), 'config-catalog-'))
+  roots.push(root)
+  return root
+}
+/** One-package fixture: the common case. */
+const make = (files: Record<string, string>, name = '@fix/one'): string => {
+  const root = makeRoot()
+  writePkg(root, 'group/one', name, files)
+  return root
+}
+
+afterEach(() => {
+  while (roots.length) rmSync(roots.pop()!, { recursive: true, force: true })
+})
+
+const DOCUMENTED_CONFIG = `/** Fixture config. */
+export interface Config {
+  /** A knob. */
+  knob?: string
+}
+`
+
+describe('gen-config-catalog classification', () => {
+  it('classifies an apply plugin with a config parameter and extracts the paste', () => {
+    const entries = collectConfigCatalog(make({
+      'src/index.ts': `import type { Context } from 'cordis'
+export const inject = ['tools']
+${DOCUMENTED_CONFIG}
+/** Load. */
+export function apply(ctx: Context, config: Config): void {}
+`,
+    }))
+    expect(entries).toHaveLength(1)
+    expect(entries[0]).toMatchObject({ pkg: '@fix/one', kind: 'config', configTypeName: 'Config', inject: ['tools'] })
+    expect(entries[0]?.pastes?.[0]?.text).toContain('/** A knob. */')
+  })
+
+  it('classifies a default service class, reading its constructor and static inject', () => {
+    const entries = collectConfigCatalog(make({
+      'src/index.ts': `import type { Context } from 'cordis'
+import z from 'schemastery'
+${DOCUMENTED_CONFIG}
+/** Fixture service. */
+export default class Fix {
+  static inject = ['llm']
+  static Config = z.object({ knob: z.string() }) as unknown as z<Config>
+  constructor(ctx: Context, config: Config) {}
+}
+`,
+    }))
+    expect(entries[0]).toMatchObject({ kind: 'config', className: 'Fix', inject: ['llm'], schemaKeys: ['knob'] })
+  })
+
+  it('classifies an abstract default class as a seam', () => {
+    const entries = collectConfigCatalog(make({
+      'src/index.ts': 'export default abstract class FixSeam { abstract run(): void }\n',
+    }))
+    expect(entries[0]).toMatchObject({ kind: 'seam', className: 'FixSeam' })
+  })
+
+  it('classifies a plugin whose apply takes no config as no-config', () => {
+    const entries = collectConfigCatalog(make({
+      'src/index.ts': 'import type { Context } from \'cordis\'\n/** Load. */\nexport function apply(ctx: Context): void {}\n',
+    }))
+    expect(entries[0]?.kind).toBe('no-config')
+  })
+
+  it('classifies a module with neither default export nor apply as a library', () => {
+    const entries = collectConfigCatalog(make({
+      'src/index.ts': 'export const helper = 1\n',
+    }))
+    expect(entries[0]?.kind).toBe('library')
+  })
+
+  it('hard-errors on a package with no entry file', () => {
+    const root = makeRoot()
+    mkdirSync(join(root, 'packages', 'group', 'one'), { recursive: true })
+    writeFileSync(join(root, 'packages', 'group', 'one', 'package.json'), JSON.stringify({ name: '@fix/one' }))
+    expect(() => collectConfigCatalog(root)).toThrow(/entry .* is missing or unreadable/)
+  })
+
+  it('hard-errors on a package.json without a name', () => {
+    const root = makeRoot()
+    mkdirSync(join(root, 'packages', 'group', 'one', 'src'), { recursive: true })
+    writeFileSync(join(root, 'packages', 'group', 'one', 'package.json'), '{}')
+    expect(() => collectConfigCatalog(root)).toThrow(/has no "name"/)
+  })
+})
+
+describe('gen-config-catalog config extraction guards', () => {
+  it('hard-errors on a config field with no JSDoc prose', () => {
+    expect(() => collectConfigCatalog(make({
+      'src/index.ts': `import type { Context } from 'cordis'
+export interface Config {
+  knob?: string
+}
+/** Load. */
+export function apply(ctx: Context, config: Config): void {}
+`,
+    }))).toThrow(/config field 'Config\.knob' .* has no JSDoc prose/)
+  })
+
+  it('hard-errors on an undocumented field nested in a type literal', () => {
+    expect(() => collectConfigCatalog(make({
+      'src/index.ts': `import type { Context } from 'cordis'
+/** Fixture config. */
+export interface Config {
+  /** Entries. */
+  entries: {
+    id: string
+  }[]
+}
+/** Load. */
+export function apply(ctx: Context, config: Config): void {}
+`,
+    }))).toThrow(/config field 'Config\.entries\.id' .* has no JSDoc prose/)
+  })
+
+  it('pastes a package-local type transitively and records external refs', () => {
+    const entries = collectConfigCatalog(make({
+      'src/index.ts': `import type { Context } from 'cordis'
+import type { Mode } from './types.ts'
+import type { Remote } from '@fix/dep'
+/** Fixture config. */
+export interface Config {
+  /** The mode. */
+  mode?: Mode
+  /** The remote. */
+  remote?: Remote
+}
+/** Load. */
+export function apply(ctx: Context, config: Config): void {}
+`,
+      'src/types.ts': '/** Fixture mode. */\nexport type Mode = \'a\' | \'b\'\n',
+    }))
+    expect(entries[0]?.pastes?.map(p => p.source)).toEqual([
+      'packages/group/one/src/index.ts:5',
+      'packages/group/one/src/types.ts:2',
+    ])
+    expect(entries[0]?.refs).toEqual([{ alias: 'Remote', imported: 'Remote', specifier: '@fix/dep' }])
+  })
+
+  it('hard-errors on a referenced type name that resolves nowhere', () => {
+    expect(() => collectConfigCatalog(make({
+      'src/index.ts': `import type { Context } from 'cordis'
+/** Fixture config. */
+export interface Config {
+  /** The ghost. */
+  ghost?: Ghost
+}
+/** Load. */
+export function apply(ctx: Context, config: Config): void {}
+`,
+    }))).toThrow(/references 'Ghost' .* neither declared in the package, imported, nor a known global/)
+  })
+
+  it('hard-errors on a config type imported from another package', () => {
+    expect(() => collectConfigCatalog(make({
+      'src/index.ts': `import type { Context } from 'cordis'
+import type { Config } from '@fix/dep'
+/** Load. */
+export function apply(ctx: Context, config: Config): void {}
+`,
+    }))).toThrow(/config type 'Config' is imported from '@fix\/dep'/)
+  })
+
+  it('hard-errors when one name resolves to two different declarations across the closure', () => {
+    expect(() => collectConfigCatalog(make({
+      'src/index.ts': `import type { Context } from 'cordis'
+import type { A } from './a.ts'
+import type { B } from './b.ts'
+/** Fixture config. */
+export interface Config {
+  /** A. */
+  a?: A
+  /** B. */
+  b?: B
+}
+/** Load. */
+export function apply(ctx: Context, config: Config): void {}
+`,
+      'src/a.ts': '/** First Option. */\nexport interface Option {\n  /** X. */\n  x?: string\n}\n/** A. */\nexport interface A {\n  /** O. */\n  o?: Option\n}\n',
+      'src/b.ts': '/** Second Option. */\nexport interface Option {\n  /** Y. */\n  y?: string\n}\n/** B. */\nexport interface B {\n  /** O. */\n  o?: Option\n}\n',
+    }))).toThrow(/type name 'Option' resolves to two different declarations/)
+  })
+})
+
+describe('gen-config-catalog schema cross-check', () => {
+  it('accepts a chained schema whose keys all appear on the config type', () => {
+    const entries = collectConfigCatalog(make({
+      'src/index.ts': `import type { Context } from 'cordis'
+import z from 'schemastery'
+${DOCUMENTED_CONFIG}
+export const Config: z<Config> = z.object({ knob: z.string() }).default({})
+/** Load. */
+export function apply(ctx: Context, config: Config): void {}
+`,
+    }))
+    expect(entries[0]?.schemaKeys).toEqual(['knob'])
+  })
+
+  it('hard-errors on a schema key the config type does not declare', () => {
+    expect(() => collectConfigCatalog(make({
+      'src/index.ts': `import type { Context } from 'cordis'
+import z from 'schemastery'
+${DOCUMENTED_CONFIG}
+export const Config: z<Config> = z.object({ knob: z.string(), hidden: z.number() })
+/** Load. */
+export function apply(ctx: Context, config: Config): void {}
+`,
+    }))).toThrow(/schema validates key 'hidden' but config type 'Config' declares no such member/)
+  })
+
+  it('hard-errors on a NESTED schema key the config type does not declare', () => {
+    expect(() => collectConfigCatalog(make({
+      'src/index.ts': `import type { Context } from 'cordis'
+import z from 'schemastery'
+/** Fixture config. */
+export interface Config {
+  /** Entries. */
+  entries: {
+    /** Id. */
+    id: string
+  }[]
+}
+export const Config: z<Config> = z.object({ entries: z.array(z.object({ id: z.string(), ghost: z.string() })) })
+/** Load. */
+export function apply(ctx: Context, config: Config): void {}
+`,
+    }))).toThrow(/schema validates key 'entries\[\]\.ghost'/)
+  })
+
+  it('resolves nested keys through a workspace-imported intersection part (re-export chains included)', () => {
+    const root = makeRoot()
+    writePkg(root, 'group/dep', '@fix/dep', {
+      'src/index.ts': 'export * from \'./types.ts\'\n',
+      'src/types.ts': '/** Shared options. */\nexport interface Opts {\n  /** Model. */\n  model?: string\n}\n',
+    })
+    writePkg(root, 'group/one', '@fix/one', {
+      'src/index.ts': `import type { Context } from 'cordis'
+import z from 'schemastery'
+import type { Opts } from '@fix/dep'
+/** Fixture config. */
+export interface Config {
+  /** Entries. */
+  entries: (Opts & {
+    /** Id. */
+    id: string
+  })[]
+}
+export const Config: z<Config> = z.object({ entries: z.array(z.object({ id: z.string(), model: z.string() })) })
+/** Load. */
+export function apply(ctx: Context, config: Config): void {}
+`,
+    })
+    expect(() => collectConfigCatalog(root)).not.toThrow()
+  })
+
+  it('resolves nested keys through a Partial<> wrapper', () => {
+    expect(() => collectConfigCatalog(make({
+      'src/index.ts': `import type { Context } from 'cordis'
+import z from 'schemastery'
+/** Caps. */
+export interface Caps {
+  /** X. */
+  x?: boolean
+}
+/** Fixture config. */
+export interface Config {
+  /** Capabilities. */
+  capabilities?: Partial<Caps>
+}
+export const Config: z<Config> = z.object({ capabilities: z.object({ x: z.boolean() }) })
+/** Load. */
+export function apply(ctx: Context, config: Config): void {}
+`,
+    }))).not.toThrow()
+  })
+
+  it('leaves a nested key under an external (unresolvable) type unreported', () => {
+    expect(() => collectConfigCatalog(make({
+      'src/index.ts': `import type { Context } from 'cordis'
+import z from 'schemastery'
+import type { External } from 'some-external-pkg'
+/** Fixture config. */
+export interface Config {
+  /** Options. */
+  options?: External
+}
+export const Config: z<Config> = z.object({ options: z.object({ whatever: z.string() }) })
+/** Load. */
+export function apply(ctx: Context, config: Config): void {}
+`,
+    }))).not.toThrow()
+  })
+
+  it('folds an intersected workspace schema into the subset check', () => {
+    const root = makeRoot()
+    writePkg(root, 'group/leaf', '@fix/leaf', {
+      'src/index.ts': `import type { Context } from 'cordis'
+import z from 'schemastery'
+/** Leaf config. */
+export interface Config {
+  /** Leaf knob. */
+  leaf?: string
+}
+/** Leaf service. */
+export default class Leaf {
+  static Config = z.object({ leaf: z.string() }) as unknown as z<Config>
+  constructor(ctx: Context, config: Config) {}
+}
+`,
+    })
+    writePkg(root, 'group/bundle', '@fix/bundle', {
+      'src/index.ts': `import type { Context } from 'cordis'
+import z from 'schemastery'
+import Leaf from '@fix/leaf'
+/** Bundle config. */
+export interface Config {
+  /** Forwarded leaf knob. */
+  leaf?: string
+}
+export const Config = z.intersect([Leaf.Config]) as unknown as z<Config>
+/** Load. */
+export function apply(ctx: Context, config: Config): void {}
+`,
+    })
+    const entries = collectConfigCatalog(root)
+    expect(entries.find(e => e.pkg === '@fix/bundle')?.schemaComposes).toEqual(['@fix/leaf'])
+  })
+
+  it('resolves composed nested keys through an indexed-access forwarder', () => {
+    const root = makeRoot()
+    writePkg(root, 'group/leaf', '@fix/leaf', {
+      'src/index.ts': `import type { Context } from 'cordis'
+import z from 'schemastery'
+/** Leaf config. */
+export interface Config {
+  /** Agents. */
+  agents: {
+    /** Id. */
+    id: string
+  }[]
+}
+/** Leaf service. */
+export default class Leaf {
+  static Config = z.object({ agents: z.array(z.object({ id: z.string() })) }) as unknown as z<Config>
+  constructor(ctx: Context, config: Config) {}
+}
+`,
+    })
+    writePkg(root, 'group/bundle', '@fix/bundle', {
+      'src/index.ts': `import type { Context } from 'cordis'
+import z from 'schemastery'
+import Leaf, { type Config as LeafConfig } from '@fix/leaf'
+/** Bundle config forwarding the leaf's agents list. */
+export interface Config {
+  /** Forwarded agents list. */
+  agents?: LeafConfig['agents']
+}
+export const Config = z.intersect([Leaf.Config]) as unknown as z<Config>
+/** Load. */
+export function apply(ctx: Context, config: Config): void {}
+`,
+    })
+    expect(() => collectConfigCatalog(root)).not.toThrow()
+  })
+
+  it('hard-errors when an intersected schema key is missing from the bundle config type', () => {
+    const root = makeRoot()
+    writePkg(root, 'group/leaf', '@fix/leaf', {
+      'src/index.ts': `import type { Context } from 'cordis'
+import z from 'schemastery'
+/** Leaf config. */
+export interface Config {
+  /** Leaf knob. */
+  leaf?: string
+}
+/** Leaf service. */
+export default class Leaf {
+  static Config = z.object({ leaf: z.string() }) as unknown as z<Config>
+  constructor(ctx: Context, config: Config) {}
+}
+`,
+    })
+    writePkg(root, 'group/bundle', '@fix/bundle', {
+      'src/index.ts': `import type { Context } from 'cordis'
+import z from 'schemastery'
+import Leaf from '@fix/leaf'
+/** Bundle config that forgot to declare the forwarded field. */
+export interface Config {
+  /** Unrelated. */
+  other?: string
+}
+export const Config = z.intersect([Leaf.Config]) as unknown as z<Config>
+/** Load. */
+export function apply(ctx: Context, config: Config): void {}
+`,
+    })
+    expect(() => collectConfigCatalog(root)).toThrow(/schema validates key 'leaf' but config type 'Config' declares no such member/)
+  })
+})
+
+describe('gen-config-catalog render', () => {
+  it('renders sections, fences, and the terse classification lists', () => {
+    const root = makeRoot()
+    writePkg(root, 'group/one', '@fix/one', {
+      'src/index.ts': `import type { Context } from 'cordis'
+${DOCUMENTED_CONFIG}
+/** Load. */
+export function apply(ctx: Context, config: Config): void {}
+`,
+    })
+    writePkg(root, 'group/lib', '@fix/lib', { 'src/index.ts': 'export const helper = 1\n' })
+    writePkg(root, 'group/seam', '@fix/seam', {
+      'src/index.ts': 'export default abstract class Seam { abstract run(): void }\n',
+    })
+    const page = render(collectConfigCatalog(root))
+    expect(page).toContain('## `@fix/one`')
+    expect(page).toContain('```ts config-catalog')
+    expect(page).toContain('/** A knob. */')
+    expect(page).toContain('- `@fix/lib` ([`packages/group/lib/src/index.ts`](../packages/group/lib/src/index.ts))')
+    expect(page).toContain('- `@fix/seam` — abstract `Seam`')
+  })
+})

packages/core/agent-loop/README.md

@@ -56,9 +56,11 @@ forever:
       drain steering
       assembly = systemPrompt.assemble({agent})  ⟵ renderPrompt(assembly) IS the full prompt
       await serial agent/pre-step        ⟵ surface mutation (compaction) outside the step
-      session('step/start')
-      request = waterfall agent/request
-      stream llm.stream(request) → session('assistant/chunk')
+      boundary = session.deriveMessages()   ⟵ reconstruction boundary: same sync frame,
+      session('step/start')                     strictly before step/start
+      config = waterfall agent/request       ⟵ frozen seed; return a replacement to switch
+      session('request/header'[-delta])      ⟵ the header event this request owes the log
+      stream llm.stream(freeze({header..., messages: boundary})) → session('assistant/chunk')
       message = waterfall agent/step-result
       session('assistant/message')
       each tool-call: session('tool/call')

packages/core/agent-loop/src/agent.ts

@@ -22,6 +22,10 @@ import { isTurnOpen, lastTurnNumber, runLoop } from './loop.ts'
  * the agent/* event taxonomy — plugins never need this class.
  */
 export class ReactLoopAgent implements Agent {
+  /**
+   * The queued + steering FIFOs behind {@link send}/{@link steer}. Public so
+   * the driver loop can drain it; {@link cancel} clears it wholesale.
+   */
   readonly inbox = new Inbox()
 
   private _status: AgentStatus = 'idle'
@@ -256,6 +260,8 @@ export class ReactLoopAgent implements Agent {
    * promise (unblocking the idle wait), releases any `whenIdle` waiters, and
    * aborts the current request if any. The returned `agent.done` promise
    * resolves once the loop exits.
+   * @returns the disposer — idempotent and infallible (it runs inside the
+   *   fiber's LIFO disposal chain, where a throw would skip later disposers).
    */
   start(): () => void {
     this.done = runLoop(this.ctx, this, {

packages/core/agent-loop/src/inbox.ts

@@ -24,30 +24,47 @@ export class Inbox {
   private steeringMessages: InboxMessage[] = []
   private wakeup: (() => void) | undefined
 
-  /** Resolves when a queued message arrives (used by the idle loop). */
+  /** True while queued messages are pending — read by the idle wait's fast path and the loop's turn-start checks. */
   get hasQueued(): boolean {
     return this.queuedMessages.length > 0
   }
 
+  /** True while steering messages are pending — read by `cancel()`'s arm gate and the loop's stop-override check. */
   get hasSteering(): boolean {
     return this.steeringMessages.length > 0
   }
 
+  /**
+   * Add a message to the queued FIFO and wake a parked {@link waitForQueued}.
+   * @param message - the message to queue for the next turn start.
+   */
   enqueue(message: InboxMessage): void {
     this.queuedMessages.push(message)
     this.wakeup?.()
   }
 
+  /**
+   * Add a message to the steering FIFO. Deliberately no wakeup: steering is
+   * drained between steps of a running turn, never by the idle wait —
+   * `Agent.steer()` on an idle agent falls back to `send()` instead.
+   * @param message - the message to inject between steps of the running turn.
+   */
   steer(message: InboxMessage): void {
     this.steeringMessages.push(message)
   }
 
-  /** Drain all queued messages (turn start). */
+  /**
+   * Drain all queued messages (turn start).
+   * @returns the drained messages in arrival order; the queued FIFO is left empty.
+   */
   drainQueued(): InboxMessage[] {
     return this.queuedMessages.splice(0)
   }
 
-  /** Drain all steering messages (between steps). */
+  /**
+   * Drain all steering messages (between steps).
+   * @returns the drained messages in arrival order; the steering FIFO is left empty.
+   */
   drainSteering(): InboxMessage[] {
     return this.steeringMessages.splice(0)
   }
@@ -62,7 +79,12 @@ export class Inbox {
     this.steeringMessages.length = 0
   }
 
-  /** Wait until a queued message arrives or `cancel` resolves. */
+  /**
+   * Wait until a queued message arrives or `cancel` resolves.
+   * @param cancel - a promise whose resolution abandons the wait without a
+   *   message (the driver loop passes the agent's disposed promise so a parked
+   *   loop can exit).
+   */
   waitForQueued(cancel: Promise<void>): Promise<void> {
     if (this.hasQueued) return Promise.resolve()
     const { promise, resolve } = Promise.withResolvers<void>()

packages/core/agent-loop/src/index.ts

@@ -29,9 +29,14 @@ declare module 'cordis' {
   }
 }
 
+/**
+ * Plugin config: the agents to create — or resume, via `resumeSessionId` —
+ * declaratively at startup, so a cordis.yml deployment needs no code.
+ */
 export interface Config {
   /** Agents created from configuration at startup. */
   agents: (AgentOptions & {
+    /** Agent id to register under; also seeds the fresh per-run session id (`${id}-session-<uuid>`). */
     id: AgentId
     /**
      * If set, the config agent RESUMES this persisted session id instead of

packages/core/agent-loop/src/loop.ts

@@ -8,10 +8,13 @@
  */
 
 import type { Context } from 'cordis'
-import type { FinishReason, GenerateOptions, Message } from '@deepseek-ai/dsh-llm'
-import { BlockAssembler, HarnessError } from '@deepseek-ai/dsh-llm'
+import type { FinishReason, GenerateOptions, LlmCallConfig, Message } from '@deepseek-ai/dsh-llm'
+import { BlockAssembler, HarnessError, deepFreeze } from '@deepseek-ai/dsh-llm'
 import type { ContinuationDecision, HookContext, PromptDecision } from '@deepseek-ai/dsh-agent'
+import { canonicalHeader } from '@deepseek-ai/dsh-session'
 import type { Session, TurnEndReason, TurnTrigger } from '@deepseek-ai/dsh-session'
+import { createTransmissionLog, recordRequestHeader } from './request-log.ts'
+import type { TransmissionLog } from './request-log.ts'
 import { renderPrompt } from '@deepseek-ai/dsh-system-prompt'
 import type { PromptAssembly } from '@deepseek-ai/dsh-system-prompt'
 import type {} from '@deepseek-ai/dsh-tools'
@@ -155,10 +158,13 @@ export interface LoopHandle {
  *       assembly = ctx.systemPrompt.assemble({agent}) ⟵ waterfall system-prompt/assemble; renderPrompt
  *                                                        (persona section + {{variables}}) IS the full prompt
  *       await ctx.serial('agent/pre-step')            ⟵ surface mutation (compaction) OUTSIDE the step
- *       session('step/start')                         ⟵ durable step boundary (no agent/* mirror)
- *       req = {model, system, tools, messages: session.deriveMessages(), signal}
- *       req = waterfall agent/request                 ⟵ hooks/model-switch
- *       stream ctx.llm.stream(req)                    ⟵ waterfall llm/stream (raw chunks)
+ *       boundary = session.deriveMessages()           ⟵ the reconstruction boundary: snapshot in the
+ *       session('step/start')                            same sync frame, strictly before step/start
+ *       config = waterfall agent/request(config)      ⟵ frozen seed; a returned replacement switches
+ *       session('request/header'|'request/header-delta')  ⟵ the header event this request owes the
+ *                                                        log (initial/resume anchor, delta, fallback)
+ *       req = freeze({header..., messages: boundary, sessionId, signal})
+ *       stream ctx.llm.stream(req)                    ⟵ waterfall llm/stream (raw chunks, frozen req)
  *         session('assistant/chunk')
  *       msg = waterfall agent/step-result             ⟵ BEFORE the log append, so the
  *       session('assistant/message' {content, usage?})   session records what actually ran
@@ -179,8 +185,18 @@ export interface LoopHandle {
  *     re-enqueue leftover steering as queued          ⟵ steering is never stranded
  *   idle (emit agent/status) unless more queued
  * ```
+ * @param ctx - the plugin context the loop reaches events (agent/…, session/flush) and services (systemPrompt, llm, tools) through.
+ * @param agent - the agent this invocation drives for its whole lifetime (its inbox, session, and options).
+ * @param handle - the bridge to the agent's mutable state: status/abort setters plus the disposal and cancel-marker reads.
  */
 export async function runLoop(ctx: Context, agent: ReactLoopAgent, handle: LoopHandle): Promise<void> {
+  // Per-instance transmission bookkeeping: whether THIS loop instance has
+  // anchored the log's header fold yet (its first request logs a
+  // 'initial'/'resume' request/header snapshot). Everything else the request
+  // needs is read from the session log itself — the loop holds no
+  // conversation state (the reconstructability RFC).
+  const transmission = createTransmissionLog()
+
   const { session } = agent
 
   while (!handle.isDisposed()) {
@@ -236,7 +252,7 @@ export async function runLoop(ctx: Context, agent: ReactLoopAgent, handle: LoopH
     // turn number is actually last in the log — a stale counter would collide.
     const turn = lastTurnNumber(session) + 1
     try {
-      await runTurn(ctx, agent, handle, turn)
+      await runTurn(ctx, agent, handle, turn, transmission)
     } catch (error: unknown) {
       // Backstop: runTurn rethrows only a PRE-turn throw (the invariant guard
       // before turn/start) — no turn/start was appended, so no turn is open and
@@ -271,7 +287,9 @@ export async function runLoop(ctx: Context, agent: ReactLoopAgent, handle: LoopH
   }
 }
 
-async function runTurn(ctx: Context, agent: ReactLoopAgent, handle: LoopHandle, turn: number): Promise<void> {
+async function runTurn(
+  ctx: Context, agent: ReactLoopAgent, handle: LoopHandle, turn: number, transmission: TransmissionLog,
+): Promise<void> {
   const { session } = agent
 
   // --- Pre-turn. A throw here (the invariant guard) is owed NO turn/end —
@@ -474,6 +492,17 @@ async function runTurn(ctx: Context, agent: ReactLoopAgent, handle: LoopHandle,
         break
       }
 
+      // The reconstruction boundary (the reconstructability RFC): the request's
+      // messages are snapshotted HERE, in the same synchronous frame as the
+      // step/start append directly below — so the snapshot is exactly the
+      // derivation over the log prefix strictly before step/start's seq.
+      // Anything appended later — by a step/start session/event listener, an
+      // agent/request-window inject(), any concurrent task — lands after the
+      // boundary and joins the NEXT request. An external reconstructor
+      // recovers these exact messages by folding the surface over
+      // events[0..stepStartSeq).
+      const boundaryMessages = session.deriveMessages()
+
       // Mark the step open BEFORE the append: Session.append pushes the event
       // to the log before notifying session/event listeners, so a THROWING
       // step/start listener leaves step/start in the log. Setting stepOpen first
@@ -495,7 +524,7 @@ async function runTurn(ctx: Context, agent: ReactLoopAgent, handle: LoopHandle,
 
       let stepOutcome: { hadToolCalls: boolean; finish: FinishReason } | { error: Error }
       try {
-        stepOutcome = await runStep(ctx, agent, turn, step, assembly, fullSystemPrompt, abort.signal)
+        stepOutcome = await runStep(ctx, agent, turn, step, assembly, fullSystemPrompt, boundaryMessages, transmission, abort.signal)
       } catch (error: unknown) {
         stepOutcome = { error: toError(error) }
       } finally {
@@ -644,11 +673,12 @@ function drainSteering(agent: ReactLoopAgent, turn: number): boolean {
   return messages.length > 0
 }
 
-/** One step: derive request from the (already pre-step-mutated) surface →
- * stream model → record → execute tools. The caller assembles the system prompt
- * and fires the `agent/pre-step` seam BEFORE opening the step, then passes the
- * resulting `assembly`/`system` here, so the surface this step derives from
- * already reflects any compaction. */
+/** One step: build the request from the boundary snapshot + the step's
+ * header → log the header event the request owes → stream model → record →
+ * execute tools. The caller assembles the system prompt, fires the
+ * `agent/pre-step` seam, snapshots the derivation, and opens the step BEFORE
+ * calling this, so `boundaryMessages` is exactly the surface prefix at
+ * step/start and already reflects any compaction. */
 async function runStep(
   ctx: Context,
   agent: ReactLoopAgent,
@@ -656,23 +686,63 @@ async function runStep(
   step: number,
   assembly: PromptAssembly,
   system: string,
+  boundaryMessages: Message[],
+  transmission: TransmissionLog,
   signal: AbortSignal,
 ): Promise<{ hadToolCalls: boolean; finish: FinishReason }> {
   const { session, options } = agent
 
-  let request: GenerateOptions = {
-    model: options.model ?? '',
-    messages: session.deriveMessages(),
-    ...system ? { system } : {},
-    ...assembly.tools.length > 0 ? { tools: assembly.tools } : {},
-    sessionId: session.id,
-    signal,
-  }
-  request = await ctx.waterfall('agent/request', agent, turn, step, request, () => Promise.resolve(request))
-  if (!request.model) {
+  // Seed the call config: the first request of THIS loop instance seeds from
+  // current AgentOptions — explicit options always win over the logged
+  // baseline, which is what keeps fork model-overrides and resume-time
+  // reconfiguration correct. Later steps seed from the log's folded header,
+  // which by then is exactly what this instance last logged.
+  // One deep-cloned, frozen seed serves BOTH the listener chain and the
+  // no-listener fallback: structuredClone decouples it from the session's
+  // cached header fold (a raw reference would let a delegating listener
+  // mutate the fold in place and silently skip the delta log), and the freeze
+  // makes in-place shaping unrepresentable — a switch is a RETURNED
+  // replacement, which the header event below records.
+  const seedConfig: LlmCallConfig = deepFreeze(structuredClone(transmission.loggedHeader
+    // eslint-disable-next-line @typescript-eslint/no-non-null-assertion -- loggedHeader ⟹ a snapshot is in the log
+    ? session.requestHeader()!.config
+    : { model: options.model ?? '' }))
+
+  // Shape the call config: listeners return a replacement to switch model or
+  // sampling (the seed is frozen — content shaping is not expressible here;
+  // model-visible content flows through the log channels). The header event
+  // below records whatever the request ACTUALLY uses, so a listener's switch
+  // is a logged, reconstructable fact, never silent drift.
+  const config = await ctx.waterfall('agent/request', agent, turn, step, seedConfig, () => Promise.resolve(seedConfig))
+  if (!config.model) {
     throw new Error(`agent "${agent.id}" has no model: set AgentOptions.model or supply one via the agent/request waterfall`)
   }
 
+  // The request header (the log's request/header* vocabulary): canonical form,
+  // recorded before dispatch so the log always explains the request.
+  const header = canonicalHeader({
+    config,
+    ...system ? { system } : {},
+    ...assembly.tools.length > 0 ? { tools: assembly.tools } : {},
+  })
+  recordRequestHeader(session, transmission, header)
+
+  // Build and freeze: the request is a pure function of (boundary snapshot,
+  // logged header) — llm/stream listeners and adapters read it, mutation
+  // throws. sessionId + frozen is the loop-built marker the dev invariant
+  // keys on.
+  const request: GenerateOptions = deepFreeze({
+    model: header.config.model,
+    messages: boundaryMessages,
+    ...header.system !== undefined ? { system: header.system } : {},
+    ...header.tools !== undefined ? { tools: header.tools } : {},
+    ...header.config.temperature !== undefined ? { temperature: header.config.temperature } : {},
+    ...header.config.maxTokens !== undefined ? { maxTokens: header.config.maxTokens } : {},
+    ...header.config.stop !== undefined ? { stop: header.config.stop } : {},
+    sessionId: session.id,
+    signal,
+  })
+
   // --- Model call (streaming-first; raw chunks are the replay record) ---
   const assembler = new BlockAssembler()
   const chunkSeqs: number[] = []
@@ -808,7 +878,11 @@ function withoutToolCalls(message: Message): Message {
   return { ...message, content: message.content.filter(block => block.type !== 'tool-call') }
 }
 
-/** The last turn number in a (possibly seeded) session log, or 0. */
+/**
+ * The last turn number in a (possibly seeded) session log, or 0.
+ * @param session - the session whose log is scanned for the latest `turn/start`.
+ * @returns the latest `turn/start`'s turn number, or 0 when the log has none (the next turn is this plus one).
+ */
 export function lastTurnNumber(session: Session): number {
   const lastStart = session.events.findLast(event => event.type === 'turn/start')
   return lastStart?.data.turn ?? 0
@@ -822,6 +896,8 @@ export function lastTurnNumber(session: Session): number {
  * returns to idle), so status is not a reliable open-turn signal. Used by
  * `inject()` to choose between appending into an open turn vs. wrapping the
  * injection in its own one-shot turn (the turn-enclosure RFC).
+ * @param session - the session whose log is inspected.
+ * @returns true when the log's last turn boundary is a `turn/start` with no matching `turn/end` yet.
  */
 export function isTurnOpen(session: Session): boolean {
   const last = session.events.findLast(e => e.type === 'turn/start' || e.type === 'turn/end')

packages/core/agent-loop/src/request-log.ts

@@ -0,0 +1,71 @@
+/**
+ * Per-loop-instance transmission bookkeeping for the reconstructability
+ * contract: which header event to append before a request so the session log
+ * always explains the request (the reconstructability RFC). The loop is
+ * otherwise transmission-stateless — the comparison baseline is the log's own
+ * folded header (`Session.requestHeader()`), so resume and fork need no
+ * special path: a fresh loop instance simply logs a `'resume'` snapshot on
+ * its first request and deltas from there.
+ *
+ * @module dsh-agent-loop/request-log
+ */
+
+import { diffHeader, headerEquals, applyHeaderDelta } from '@deepseek-ai/dsh-session'
+import type { EpochHeader, Session } from '@deepseek-ai/dsh-session'
+
+/** Per-loop-instance bookkeeping: whether THIS instance has logged a header yet. */
+export interface TransmissionLog {
+  /** True once this loop instance appended its anchoring `request/header` snapshot. */
+  loggedHeader: boolean
+}
+
+/**
+ * Fresh bookkeeping for a newly-started loop instance.
+ * @returns state with `loggedHeader` false, so the instance's first request appends an anchoring snapshot.
+ */
+export function createTransmissionLog(): TransmissionLog {
+  return { loggedHeader: false }
+}
+
+/**
+ * Append whatever header event this request owes the log, so folding the log
+ * reproduces the header the request was built under. Exactly one of four
+ * things happens:
+ *
+ * 1. This loop instance has not logged a header yet → a full `request/header`
+ *    snapshot anchors the fold: reason `'initial'` when the log has no header
+ *    events at all (a new conversation), `'resume'` when it does (process
+ *    restart, fork seed — the boundary itself is a recorded fact, so the
+ *    snapshot is appended even when nothing changed).
+ * 2. The header equals the folded baseline → nothing; the log already
+ *    explains this request.
+ * 3. It differs and the delta round-trips (`applyHeaderDelta` on the baseline
+ *    reproduces the header exactly) → a `request/header-delta`.
+ * 4. It differs and the delta encoding cannot express the change (a pure tool
+ *    reordering) → a full snapshot with reason `'fallback'`; deltas are an
+ *    encoding optimization, never a correctness dependency.
+ *
+ * @param session - the session whose log explains the request.
+ * @param state - this loop instance's bookkeeping (mutated on first log).
+ * @param header - the canonical header the request will ACTUALLY use
+ *   (post-`agent/request`).
+ */
+export function recordRequestHeader(session: Session, state: TransmissionLog, header: EpochHeader): void {
+  if (!state.loggedHeader) {
+    session.append('request/header', { header, reason: session.requestHeader() === undefined ? 'initial' : 'resume' })
+    state.loggedHeader = true
+    return
+  }
+  // This instance logged a snapshot, so the fold is necessarily defined.
+  // eslint-disable-next-line @typescript-eslint/no-non-null-assertion
+  const baseline = session.requestHeader()!
+  if (headerEquals(baseline, header)) return
+  const delta = diffHeader(baseline, header)
+  /* v8 ignore next -- headerEquals false ⟹ diffHeader defined: both compare the same three parts */
+  if (delta === undefined) return
+  if (headerEquals(applyHeaderDelta(baseline, delta), header)) {
+    session.append('request/header-delta', delta)
+  } else {
+    session.append('request/header', { header, reason: 'fallback' })
+  }
+}

packages/core/agent-loop/tests/loop.spec.ts

@@ -230,9 +230,8 @@ describe('agent loop', () => {
       assembly.variables['model'] = 'mock'
       return next()
     })
-    ctx.on('agent/request', async (_agent, _turn, _step, options, next) => {
-      options.model = 'mock'
-      return next()
+    ctx.on('agent/request', async (_agent, _turn, _step, config, _next) => {
+      return { ...config, model: 'mock' }
     })
     const agent = ctx.agentLoop.create(AgentId('a-late-model'), {})
 
@@ -428,20 +427,27 @@ describe('agent loop', () => {
     expect(agent.session.events.some(e => e.type === 'tool/result')).toBe(true)
   })
 
-  it('agent/request waterfall can rewrite the request (model-switch pattern)', async () => {
+  it('agent/request waterfall switches models by returning a replacement config; the switch is logged', async () => {
     const adapter = new MockAdapter([textResponse('ok')])
     const ctx = await harness(adapter)
     ctx.llm.registerAdapter(['other-model'], adapter)
     const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
 
-    ctx.on('agent/request', async (_agent, _turn, _step, options, next) => {
-      options.model = 'other-model'
-      return next()
+    ctx.on('agent/request', async (_agent, _turn, _step, config, _next) => {
+      // The seed is frozen — config is not a mutable per-call knob; a switch
+      // is proposed by returning a replacement, and the loop logs it.
+      expect(Object.isFrozen(config)).toBe(true)
+      expect(() => { (config as { model: string }).model = 'other-model' }).toThrow(TypeError)
+      return { ...config, model: 'other-model' }
     })
 
     send(agent, 'hi')
     await waitForIdle(ctx, agent)
     expect(adapter.requests[0]!.model).toBe('other-model')
+    // The header event records what the request ACTUALLY used — the switch is
+    // a reconstructable fact, not silent drift.
+    const headerEvent = agent.session.events.find(e => e.type === 'request/header')
+    expect(headerEvent?.type === 'request/header' && headerEvent.data.header.config.model).toBe('other-model')
   })
 
   it('agent/pre-step fires once per step before the step is opened', async () => {

packages/core/agent-loop/tests/request-cache.e2e.ts

@@ -0,0 +1,104 @@
+import { afterEach, describe, expect, it } from 'vitest'
+import { Context } from 'cordis'
+import LlmService from '@deepseek-ai/dsh-llm'
+import SessionStore from '@deepseek-ai/dsh-session'
+import SystemPrompt from '@deepseek-ai/dsh-system-prompt'
+import ToolRegistry, { defineTool } from '@deepseek-ai/dsh-tools'
+import AgentRegistry, { AgentId, type Agent } from '@deepseek-ai/dsh-agent'
+import AgentLoop from '@deepseek-ai/dsh-agent-loop'
+import * as LlmDeepSeek from '@deepseek-ai/dsh-llm-deepseek'
+
+/**
+ * With-key proof that log-derived requests translate into REAL provider cache
+ * hits: a multi-step tool turn (plus a follow-up turn) against the live
+ * DeepSeek API must report `cacheReadTokens > 0` on every request after the
+ * first — the adapter maps the provider's `prompt_cache_hit_tokens`, and the
+ * per-step usage recorded on `assistant/message` events is the production
+ * observable for cache behavior (the reconstructability RFC's measurement
+ * layer: prefix stability is corollary #1). Mocks prove the requests are
+ * append-extensions; only the real API proves those bytes actually hit the
+ * provider cache. Key-gated — skips entirely without $DEEPSEEK_API_KEY.
+ */
+
+// Long enough that the shared request prefix comfortably spans the provider's
+// cache-block granularity (64 tokens) from the very first request.
+const SYSTEM = 'You are a terse coding assistant used in an automated cache test. '
+  + 'Always follow instructions literally and exactly. When the user asks you to look '
+  + 'something up, call the lookup tool with the requested key and wait for its result '
+  + 'before answering. Never invent a value the tool has not returned. After the tool '
+  + 'returns, answer with a single short sentence that repeats the returned value '
+  + 'verbatim. Do not add explanations, do not use markdown, do not ask follow-up '
+  + 'questions. If the user asks anything else, answer in one short sentence.'
+
+let ctx: Context | undefined
+
+afterEach(async () => {
+  await ctx?.fiber.dispose()
+  ctx = undefined
+})
+
+async function loopHarness(): Promise<Context> {
+  const created = new Context()
+  await created.plugin(LlmService)
+  await created.plugin(SessionStore)
+  await created.plugin(SystemPrompt, { persona: SYSTEM })
+  await created.plugin(ToolRegistry)
+  await created.plugin(AgentRegistry)
+  await created.plugin(AgentLoop, { agents: [] })
+  await created.plugin(LlmDeepSeek, { models: ['deepseek-v4-flash'] })
+  created.tools.register(defineTool({
+    name: 'lookup',
+    description: 'Look up the stored value for a key.',
+    parameters: { key: { type: 'string', description: 'The key to look up.' } },
+    async execute(args) {
+      return [{ type: 'text', text: `value(${String(args.key)}) = azure-falcon-42` }]
+    },
+  }))
+  return created
+}
+
+function waitForIdle(context: Context, agent: Agent): Promise<void> {
+  return new Promise((resolve) => {
+    const dispose = context.on('agent/status', (subject, status) => {
+      if (subject === agent && status === 'idle') {
+        dispose()
+        resolve()
+      }
+    })
+  })
+}
+
+describe.skipIf(!process.env.DEEPSEEK_API_KEY)('log-derived request cache hits (real API)', () => {
+  it('every request after the first hits the provider prefix cache', async () => {
+    ctx = await loopHarness()
+    const agent = ctx.agentLoop.create(AgentId('cache-e2e'), { model: 'deepseek-v4-flash' })
+
+    // Turn 1: forces a tool call → at least two steps (two model requests).
+    agent.send([{ type: 'text', text: 'Look up the key "deploy-color" with the lookup tool and tell me the value.' }])
+    await waitForIdle(ctx, agent)
+    // Turn 2: a follow-up over the same (longer) prefix.
+    agent.send([{ type: 'text', text: 'Thanks. Repeat that value one more time.' }])
+    await waitForIdle(ctx, agent)
+
+    const usages = [...agent.session.events]
+      .filter(e => e.type === 'assistant/message')
+      .map(e => e.data.usage)
+    expect(usages.length).toBeGreaterThanOrEqual(3) // 2 steps in turn 1 + ≥1 in turn 2
+    for (const usage of usages) expect(usage).toBeDefined()
+
+    // The first request has nothing to hit; every later one shares its
+    // predecessor as a byte-identical prefix, so the provider must report
+    // cached prompt tokens (prompt_cache_hit_tokens → cacheReadTokens).
+    for (const usage of usages.slice(1)) {
+      expect(usage!.cacheReadTokens ?? 0).toBeGreaterThan(0)
+    }
+
+    // World-verification of the conversation itself: the tool value made it
+    // through the loop into the final answer.
+    const finalText = agent.session.deriveMessages().at(-1)!.content
+      .filter(block => block.type === 'text')
+      .map(block => block.text)
+      .join('')
+    expect(finalText).toContain('azure-falcon-42')
+  }, 180_000)
+})

packages/core/agent-loop/tests/request-log.spec.ts

@@ -0,0 +1,85 @@
+/**
+ * recordRequestHeader unit tests: exactly one of four things per request —
+ * an 'initial' snapshot (log has no header yet), a 'resume' snapshot (fresh
+ * loop instance over a log that has one), nothing (header unchanged), a
+ * round-tripping delta, or a 'fallback' snapshot when the delta encoding
+ * cannot express the change (pure tool reordering).
+ */
+
+import { describe, expect, it } from 'vitest'
+import { Session, SessionId, canonicalHeader } from '@deepseek-ai/dsh-session'
+import type { SessionEvent } from '@deepseek-ai/dsh-session'
+import type { ToolSchema } from '@deepseek-ai/dsh-llm'
+import { createTransmissionLog, recordRequestHeader } from '../src/request-log.ts'
+
+function tool(name: string, description = 'd'): ToolSchema {
+  return { name, description, parameters: { type: 'object' } }
+}
+
+function openSession(id: string): Session {
+  const session = new Session(SessionId(id))
+  session.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
+  return session
+}
+
+function headerEvents(session: Session): SessionEvent[] {
+  return session.events.filter(e => e.type === 'request/header' || e.type === 'request/header-delta')
+}
+
+describe('recordRequestHeader', () => {
+  it("anchors a new conversation with an 'initial' snapshot, then logs nothing while unchanged", () => {
+    const session = openSession('rl-initial')
+    const state = createTransmissionLog()
+    const header = canonicalHeader({ config: { model: 'm' }, system: 's', tools: [tool('t')] })
+
+    recordRequestHeader(session, state, header)
+    const [first] = headerEvents(session)
+    expect(first?.type === 'request/header' && first.data.reason).toBe('initial')
+
+    recordRequestHeader(session, state, header)
+    expect(headerEvents(session)).toHaveLength(1)
+  })
+
+  it("anchors a fresh loop instance over an anchored log with a 'resume' snapshot, even unchanged", () => {
+    const session = openSession('rl-resume')
+    const header = canonicalHeader({ config: { model: 'm' }, system: 's' })
+    recordRequestHeader(session, createTransmissionLog(), header)
+
+    // A second instance (process restart / fork): the boundary itself is a
+    // recorded fact — snapshot appended even though the header is identical.
+    recordRequestHeader(session, createTransmissionLog(), header)
+    const events = headerEvents(session)
+    expect(events).toHaveLength(2)
+    expect(events[1]?.type === 'request/header' && events[1].data.reason).toBe('resume')
+  })
+
+  it('logs a round-tripping delta for a mid-run change, and the fold reproduces the header', () => {
+    const session = openSession('rl-delta')
+    const state = createTransmissionLog()
+    const first = canonicalHeader({ config: { model: 'm' }, system: 'a\nb', tools: [tool('t')] })
+    recordRequestHeader(session, state, first)
+
+    const second = canonicalHeader({ config: { model: 'm' }, system: 'a\nc', tools: [tool('t'), tool('u')] })
+    recordRequestHeader(session, state, second)
+    const events = headerEvents(session)
+    expect(events).toHaveLength(2)
+    expect(events[1]?.type).toBe('request/header-delta')
+    expect(session.requestHeader()).toEqual(second)
+  })
+
+  it("records a change the delta cannot express (pure reordering) as a 'fallback' snapshot", () => {
+    const session = openSession('rl-fallback')
+    const state = createTransmissionLog()
+    const first = canonicalHeader({ config: { model: 'm' }, tools: [tool('a'), tool('b')] })
+    recordRequestHeader(session, state, first)
+
+    const reordered = canonicalHeader({ config: { model: 'm' }, tools: [tool('b'), tool('a')] })
+    recordRequestHeader(session, state, reordered)
+    const events = headerEvents(session)
+    expect(events).toHaveLength(2)
+    expect(events[1]?.type === 'request/header' && events[1].data.reason).toBe('fallback')
+    // The fold still lands on the exact header — deltas are an encoding
+    // optimization, never a correctness dependency.
+    expect(session.requestHeader()).toEqual(reordered)
+  })
+})

packages/core/agent-loop/tests/request-reconstruction.spec.ts

@@ -0,0 +1,313 @@
+/**
+ * Loop-level reconstructability: every request the loop sends is a pure
+ * function of the session log — messages are the derivation at the step/start
+ * boundary, the header is the fold of request/header* events — and every
+ * request is an append-extension of its predecessor unless a logged event
+ * (compaction replace, header change) explains the difference. The requests
+ * recorded by the mock adapter are the observable; the offline-rebuild test
+ * at the bottom is the theorem stated end-to-end.
+ */
+
+import { describe, expect, it } from 'vitest'
+import { Context } from 'cordis'
+import LlmService from '@deepseek-ai/dsh-llm'
+import type { GenerateOptions } from '@deepseek-ai/dsh-llm'
+import SessionStore, { Session, SessionId, foldRequestHeader } from '@deepseek-ai/dsh-session'
+import SystemPrompt from '@deepseek-ai/dsh-system-prompt'
+import ToolRegistry, { defineTool } from '@deepseek-ai/dsh-tools'
+import AgentRegistry, { AgentId } from '@deepseek-ai/dsh-agent'
+import AgentLoop, { ReactLoopAgent } from '@deepseek-ai/dsh-agent-loop'
+import { MockAdapter, textResponse, toolCallResponse } from './mock-adapter.ts'
+
+async function harness(adapter: MockAdapter, persona = 'stable base') {
+  const ctx = new Context()
+  await ctx.plugin(LlmService)
+  await ctx.plugin(SessionStore)
+  await ctx.plugin(SystemPrompt, { persona })
+  await ctx.plugin(ToolRegistry)
+  await ctx.plugin(AgentRegistry)
+  await ctx.plugin(AgentLoop, { agents: [] })
+  ctx.llm.registerAdapter(['mock'], adapter)
+  return ctx
+}
+
+function waitForIdle(ctx: Context, agent: ReactLoopAgent): Promise<void> {
+  return new Promise((resolve) => {
+    const dispose = ctx.on('agent/status', (subject, status) => {
+      if (subject === agent && status === 'idle') {
+        dispose()
+        resolve()
+      }
+    })
+  })
+}
+
+function send(agent: ReactLoopAgent, text: string) {
+  agent.send([{ type: 'text', text }])
+}
+
+/** Assert `previous` is a strict value-prefix of `current`. */
+function expectPrefixExtension(previous: GenerateOptions, current: GenerateOptions) {
+  expect(current.messages.length).toBeGreaterThan(previous.messages.length)
+  expect(current.messages.slice(0, previous.messages.length)).toEqual([...previous.messages])
+  expect(current.system).toEqual(previous.system)
+  expect(current.tools).toEqual(previous.tools)
+}
+
+function registerEcho(ctx: Context) {
+  ctx.tools.register(defineTool({
+    name: 'echo',
+    description: 'echo back',
+    parameters: { text: { type: 'string' } },
+    async execute(args) {
+      return [{ type: 'text', text: `echo: ${String(args.text)}` }]
+    },
+  }))
+}
+
+describe('request stability across the loop', () => {
+  it('each step request within a turn append-extends the previous, frozen end to end', async () => {
+    const adapter = new MockAdapter([
+      toolCallResponse('c1', 'echo', { text: 'one' }, 'first'),
+      toolCallResponse('c2', 'echo', { text: 'two' }, 'second'),
+      textResponse('done'),
+    ])
+    const ctx = await harness(adapter)
+    registerEcho(ctx)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+
+    send(agent, 'go')
+    await waitForIdle(ctx, agent)
+
+    expect(adapter.requests).toHaveLength(3)
+    expectPrefixExtension(adapter.requests[0]!, adapter.requests[1]!)
+    expectPrefixExtension(adapter.requests[1]!, adapter.requests[2]!)
+    for (const request of adapter.requests) {
+      expect(Object.isFrozen(request)).toBe(true)
+      expect(Object.isFrozen(request.messages)).toBe(true)
+    }
+    // One anchoring header snapshot; no further header events (nothing changed).
+    const headerEvents = agent.session.events.filter(e => e.type === 'request/header' || e.type === 'request/header-delta')
+    expect(headerEvents).toHaveLength(1)
+    expect(headerEvents[0]?.type === 'request/header' && headerEvents[0].data.reason).toBe('initial')
+  })
+
+  it('a later turn append-extends the previous turn (one conversation, one log)', async () => {
+    const adapter = new MockAdapter([textResponse('one'), textResponse('two')])
+    const ctx = await harness(adapter)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+
+    send(agent, 'first')
+    await waitForIdle(ctx, agent)
+    send(agent, 'second')
+    await waitForIdle(ctx, agent)
+
+    expect(adapter.requests).toHaveLength(2)
+    expectPrefixExtension(adapter.requests[0]!, adapter.requests[1]!)
+  })
+
+  it('a compaction replace rewrites the resend, and the log explains it', async () => {
+    const adapter = new MockAdapter([textResponse('one'), textResponse('two')])
+    const ctx = await harness(adapter)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+
+    send(agent, 'first')
+    await waitForIdle(ctx, agent)
+
+    // A pre-step listener compacts turn 1's history before turn 2's step —
+    // the sanctioned surface rewrite, landing OUTSIDE the step.
+    const preStep = ctx.on('agent/pre-step', () => {
+      preStep()
+      const session = agent.session
+      const nodes = session.surface.nodes
+      session.append('context/message', {
+        content: [{ type: 'text', text: '[summary of turn 1]' }],
+        source: { kind: 'plugin', plugin: 'test-compact' },
+      }, {
+        surfaceOp: { op: 'replace', start: nodes[0]!.seq, end: nodes[1]!.seq },
+        sourceEventSeqs: [nodes[0]!.seq, nodes[1]!.seq],
+      })
+    })
+
+    send(agent, 'second')
+    await waitForIdle(ctx, agent)
+
+    const second = adapter.requests[1]!
+    // The rewritten history: summary replaces turn 1's user+assistant pair.
+    expect(second.messages[0]!.content.some(b => b.type === 'text' && b.text.includes('[summary of turn 1]'))).toBe(true)
+    // No header event beyond the anchor: the replace is itself in the log.
+    expect(agent.session.events.filter(e => e.type === 'request/header')).toHaveLength(1)
+  })
+
+  it('a real system-prompt change is a logged header delta; a stable prompt logs nothing', async () => {
+    const adapter = new MockAdapter([textResponse('one'), textResponse('two'), textResponse('three')])
+    const ctx = await harness(adapter)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+
+    send(agent, 'first')
+    await waitForIdle(ctx, agent)
+    send(agent, 'second')
+    await waitForIdle(ctx, agent)
+    // Identical assembly re-rendered per step is NOT a change.
+    expect(agent.session.events.filter(e => e.type === 'request/header-delta')).toHaveLength(0)
+
+    ctx.systemPrompt.section({ name: 'extra', order: 2, text: 'new guidance' })
+    send(agent, 'third')
+    await waitForIdle(ctx, agent)
+
+    const deltas = agent.session.events.filter(e => e.type === 'request/header-delta')
+    expect(deltas).toHaveLength(1)
+    expect(adapter.requests[2]!.system).toContain('new guidance')
+    // History is preserved across the change — only the header moved.
+    expect(adapter.requests[2]!.messages.length).toBeGreaterThan(adapter.requests[1]!.messages.length)
+  })
+
+  it('an inject() during the agent/request waterfall joins the NEXT request (the step/start boundary)', async () => {
+    const adapter = new MockAdapter([textResponse('one'), textResponse('two')])
+    const ctx = await harness(adapter)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+
+    let injected = false
+    ctx.on('agent/request', async (_agent, _turn, _step, _config, next) => {
+      if (!injected) {
+        injected = true
+        agent.inject([{ type: 'text', text: '[late context]' }], { source: { kind: 'plugin', plugin: 'test' } })
+      }
+      return next()
+    })
+
+    send(agent, 'first')
+    await waitForIdle(ctx, agent)
+    const first = adapter.requests[0]!
+    // The inject landed in the log after the boundary: not in THIS request…
+    expect(first.messages.some(m => m.content.some(b => b.type === 'text' && b.text.includes('[late context]')))).toBe(false)
+    expect(agent.session.events.some(e => e.type === 'context/message')).toBe(true)
+
+    send(agent, 'second')
+    await waitForIdle(ctx, agent)
+    // …but in the next one, at its logged position.
+    const second = adapter.requests[1]!
+    expect(second.messages.some(m => m.content.some(b => b.type === 'text' && b.text.includes('[late context]')))).toBe(true)
+  })
+
+  it('a mutation attempt on the frozen request content throws into the step (loud, not silent)', async () => {
+    const adapter = new MockAdapter([textResponse('one')])
+    const ctx = await harness(adapter)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+
+    const errors: Error[] = []
+    ctx.on('agent/error', (_agent, _turn, _step, error) => void errors.push(error))
+    ctx.on('llm/stream', (options, next) => {
+      // The historical failure mode this design kills: a listener rewriting
+      // request content in place. The freeze turns it into a loud error.
+      options.messages.push({ role: 'user', content: [{ type: 'text', text: 'sneaky' }] })
+      return next()
+    })
+
+    send(agent, 'go')
+    await waitForIdle(ctx, agent)
+
+    expect(errors).toHaveLength(1)
+    expect(errors[0]!.message).toMatch(/not extensible|frozen|read only|readonly/i)
+  })
+
+  it('a fresh loop instance over a seeded log anchors with a resume snapshot and stays cache-aligned', async () => {
+    const adapter = new MockAdapter([textResponse('one')])
+    const ctx = await harness(adapter)
+    const agent = ctx.agentLoop.create(AgentId('gen1'), { model: 'mock' })
+    send(agent, 'first')
+    await waitForIdle(ctx, agent)
+
+    // Second generation: a new agent whose session is seeded with the first
+    // one's full log (the resume/fork path).
+    const adapter2 = new MockAdapter([textResponse('two')])
+    const ctx2 = await harness(adapter2)
+    const handle = ctx2.agents.create({
+      agentId: AgentId('gen2'),
+      sessionId: SessionId('gen2-session'),
+      seed: [...agent.session.events],
+      agentOptions: { model: 'mock' },
+    })
+    const agent2 = handle.agent as ReactLoopAgent
+    send(agent2, 'second')
+    await waitForIdle(ctx2, agent2)
+
+    const snapshots = agent2.session.events.filter(e => e.type === 'request/header')
+    expect(snapshots).toHaveLength(2)
+    expect(snapshots[1]?.type === 'request/header' && snapshots[1].data.reason).toBe('resume')
+    // Identical header across the restart: byte-identical continuation.
+    expect(adapter2.requests[0]!.system).toEqual(adapter.requests[0]!.system)
+    expectPrefixExtension(adapter.requests[0]!, adapter2.requests[0]!)
+  })
+
+  it('a delegating listener cannot mutate the seed through next() — the fold stays log-true', async () => {
+    const adapter = new MockAdapter([textResponse('one'), textResponse('two')])
+    const ctx = await harness(adapter)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+
+    ctx.on('agent/request', async (_agent, _turn, _step, _config, next) => {
+      const config = await next()
+      // next() resolves the SAME frozen seed — in-place shaping after
+      // delegation is unrepresentable, so a "mutate what next() returned"
+      // listener cannot desync the log from the request (nor reach the
+      // session's cached header fold, which is deep-cloned away and itself
+      // frozen).
+      expect(Object.isFrozen(config)).toBe(true)
+      expect(() => { (config as { temperature?: number }).temperature = 0.9 }).toThrow(TypeError)
+      return config
+    })
+
+    send(agent, 'first')
+    await waitForIdle(ctx, agent)
+    send(agent, 'second')
+    await waitForIdle(ctx, agent)
+
+    // No delta was logged (nothing really changed), and the session's own
+    // fold is immutable state.
+    expect(agent.session.events.filter(e => e.type === 'request/header-delta')).toHaveLength(0)
+    expect(Object.isFrozen(agent.session.requestHeader())).toBe(true)
+    expect(adapter.requests[1]!.temperature).toBeUndefined()
+  })
+
+  it('THEOREM: every request rebuilds byte-equal from the session log alone', async () => {
+    const adapter = new MockAdapter([
+      toolCallResponse('c1', 'echo', { text: 'one' }, 'calling'),
+      textResponse('done'),
+      textResponse('after change'),
+    ])
+    const ctx = await harness(adapter)
+    registerEcho(ctx)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+
+    send(agent, 'go')
+    await waitForIdle(ctx, agent)
+    ctx.systemPrompt.section({ name: 'extra', order: 2, text: 'now with guidance' })
+    ctx.on('agent/request', async (_agent, _turn, _step, config, _next) => ({ ...config, temperature: 0.5, maxTokens: 99, stop: ['<END>'] }))
+    send(agent, 'again')
+    await waitForIdle(ctx, agent)
+
+    expect(adapter.requests).toHaveLength(3)
+    const events = agent.session.events
+    const stepStarts = events.filter(e => e.type === 'step/start')
+    expect(stepStarts).toHaveLength(3)
+
+    adapter.requests.forEach((request, index) => {
+      const stepStart = stepStarts[index]!
+      // Messages: the derivation over the log prefix strictly before this
+      // step's step/start — rebuilt here through a completely fresh Session.
+      const rebuilt = new Session(SessionId(`rebuild-${index}`), structuredClone(events.slice(0, stepStart.seq)))
+      expect(structuredClone(request.messages)).toEqual(rebuilt.deriveMessages())
+
+      // Header: the fold of request/header* events up to this step's dispatch
+      // (its header event sits between step/start and the first chunk).
+      const firstChunk = events.find(e => e.type === 'assistant/chunk' && e.seq > stepStart.seq)!
+      const header = foldRequestHeader(events.slice(0, firstChunk.seq))!
+      expect(request.model).toBe(header.config.model)
+      expect(request.system).toEqual(header.system)
+      expect(structuredClone(request.tools ?? [])).toEqual(structuredClone(header.tools ?? []))
+      expect(request.temperature).toBe(header.config.temperature)
+      expect(request.maxTokens).toBe(header.config.maxTokens)
+      expect(request.stop).toEqual(header.config.stop)
+    })
+  })
+})

packages/core/agent-loop/tests/review-fixes.spec.ts

@@ -399,9 +399,8 @@ describe('MEDIUM: misc registry and config fixes', () => {
     const ctx = await harness(adapter)
     const agent = ctx.agentLoop.create(AgentId('a1'), {}) // no model — router plugin decides
 
-    ctx.on('agent/request', async (_agent, _turn, _step, options, next) => {
-      options.model = 'mock'
-      return next()
+    ctx.on('agent/request', async (_agent, _turn, _step, config, _next) => {
+      return { ...config, model: 'mock' }
     })
 
     send(agent, 'go')

packages/core/agent-loop/tests/tool-order.spec.ts

@@ -0,0 +1,117 @@
+/**
+ * Loop-level tool-order determinism: the request/header event — and therefore
+ * the frozen request the adapter receives — carries the assembly's canonical
+ * tool order (system-prompt's `toolOrder` config, or lexicographic name
+ * order), regardless of the order tool plugins happened to register in.
+ * Registration order is a plugin-load artifact (concurrent dynamic imports
+ * race), so nothing downstream of the registry may depend on it.
+ */
+
+import { describe, expect, it } from 'vitest'
+import { Context } from 'cordis'
+import LlmService from '@deepseek-ai/dsh-llm'
+import SessionStore, { foldRequestHeader } from '@deepseek-ai/dsh-session'
+import SystemPrompt, { TOOL_ORDER_REST } from '@deepseek-ai/dsh-system-prompt'
+import type { Config as SystemPromptConfig } from '@deepseek-ai/dsh-system-prompt'
+import ToolRegistry, { defineTool } from '@deepseek-ai/dsh-tools'
+import AgentRegistry, { AgentId } from '@deepseek-ai/dsh-agent'
+import AgentLoop, { ReactLoopAgent } from '@deepseek-ai/dsh-agent-loop'
+import { MockAdapter, textResponse } from './mock-adapter.ts'
+
+async function harness(adapter: MockAdapter, toolOrder?: SystemPromptConfig['toolOrder']) {
+  const ctx = new Context()
+  await ctx.plugin(LlmService)
+  await ctx.plugin(SessionStore)
+  await ctx.plugin(SystemPrompt, { persona: 'stable base', ...toolOrder !== undefined ? { toolOrder } : {} })
+  await ctx.plugin(ToolRegistry)
+  await ctx.plugin(AgentRegistry)
+  await ctx.plugin(AgentLoop, { agents: [] })
+  ctx.llm.registerAdapter(['mock'], adapter)
+  return ctx
+}
+
+function waitForIdle(ctx: Context, agent: ReactLoopAgent): Promise<void> {
+  return new Promise((resolve) => {
+    const dispose = ctx.on('agent/status', (subject, status) => {
+      if (subject === agent && status === 'idle') {
+        dispose()
+        resolve()
+      }
+    })
+  })
+}
+
+function registerNamed(ctx: Context, name: string) {
+  ctx.tools.register(defineTool({
+    name,
+    description: `the ${name} tool`,
+    parameters: {},
+    async execute() {
+      return [{ type: 'text', text: name }]
+    },
+  }))
+}
+
+/** Run one text-only turn and return the harness context + agent. */
+async function runTurn(registrationOrder: string[], toolOrder?: SystemPromptConfig['toolOrder']) {
+  const adapter = new MockAdapter([textResponse('done')])
+  const ctx = await harness(adapter, toolOrder)
+  for (const name of registrationOrder) registerNamed(ctx, name)
+  const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+  agent.send([{ type: 'text', text: 'go' }])
+  await waitForIdle(ctx, agent)
+  return { ctx, agent, adapter }
+}
+
+describe('loop-level canonical tool order', () => {
+  it('logs the request/header with tools in canonical order, not registration order', async () => {
+    const { agent, adapter } = await runTurn(['zulu', 'alpha', 'mike'])
+    const header = foldRequestHeader(agent.session.events)
+    expect(header?.tools?.map(tool => tool.name)).toEqual(['alpha', 'mike', 'zulu'])
+    // The dispatched request is built FROM the logged header (whose tools the
+    // assembly already canonicalized) and reaches the adapter deep-frozen —
+    // the marker the reconstruction invariant keys on.
+    expect(adapter.requests[0]?.tools?.map(tool => tool.name)).toEqual(['alpha', 'mike', 'zulu'])
+    expect(Object.isFrozen(adapter.requests[0])).toBe(true)
+    expect(adapter.requests[0]?.sessionId).toBe(agent.session.id)
+  })
+
+  it('produces the same header order for any registration order', async () => {
+    const first = await runTurn(['alpha', 'mike', 'zulu'])
+    const second = await runTurn(['zulu', 'mike', 'alpha'])
+    const names = (run: typeof first) => foldRequestHeader(run.agent.session.events)?.tools?.map(tool => tool.name)
+    expect(names(first)).toEqual(['alpha', 'mike', 'zulu'])
+    expect(names(second)).toEqual(names(first))
+  })
+
+  it('honors a configured toolOrder in the logged header and the dispatched request', async () => {
+    const { agent, adapter } = await runTurn(['alpha', 'zulu', 'mike'], ['zulu', TOOL_ORDER_REST])
+    const header = foldRequestHeader(agent.session.events)
+    expect(header?.tools?.map(tool => tool.name)).toEqual(['zulu', 'alpha', 'mike'])
+    expect(adapter.requests[0]?.tools?.map(tool => tool.name)).toEqual(['zulu', 'alpha', 'mike'])
+    expect(Object.isFrozen(adapter.requests[0])).toBe(true)
+  })
+
+  it('fails the turn — no model request — when toolOrder names an unregistered tool', async () => {
+    // The assemble rejection escapes to runTurn's outer catch: the open turn
+    // closes with an `error` reason (agent/error mirrors it), no step opens,
+    // no request/header is logged, the adapter never sees a request, and the
+    // agent returns to idle — a misconfigured deployment fails every turn
+    // deterministically instead of silently reordering nothing.
+    const adapter = new MockAdapter([textResponse('never sent')])
+    const ctx = await harness(adapter, ['ghost', TOOL_ORDER_REST])
+    registerNamed(ctx, 'alpha')
+    const errors: Error[] = []
+    ctx.on('agent/error', (_agent, _turn, _step, error) => void errors.push(error))
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+    agent.send([{ type: 'text', text: 'go' }])
+    await waitForIdle(ctx, agent)
+    expect(adapter.requests).toHaveLength(0)
+    expect(errors.map(e => e.message)).toEqual(['toolOrder lists unregistered tool "ghost"; registered tools: alpha'])
+    expect(foldRequestHeader(agent.session.events)).toBeUndefined()
+    const end = agent.session.events.find(e => e.type === 'turn/end')
+    expect(end?.type === 'turn/end' && end.data.reason).toMatchObject({ kind: 'error', step: 1 })
+    // The turn is balanced (turn/start → turn/end) with no step events inside.
+    expect(agent.session.events.some(e => e.type === 'step/start')).toBe(false)
+  })
+})

packages/core/agent/README.md

@@ -44,7 +44,7 @@ Turn and step boundaries are NOT mirrored as `agent/*` emits: a consumer that ne
 - `agent/session-start` (emit) — fired once before the first turn; a listener seeds context via `agent.inject()` (it cannot veto startup).
 - `agent/prompt-submit` — decide what happens to one drained queued message before it becomes a `user/message`: `PromptDecision` = `allow` (optionally rewriting the prompt `content` or attaching `additionalContext`) or `block` (drop it; a batch whose every prompt is blocked opens a zero-step turn that ends `rejected`). Maps onto Claude Code's `UserPromptSubmit`.
 - `agent/pre-step` (serial) — mutate the session surface before the step opens and history is derived (compaction). Fires after `turn/start` and before `step/start`, so a listener's appended events land outside the step.
-- `agent/request` — mutate `GenerateOptions` before the model call (hooks, model switching, tool filtering)
+- `agent/request` — shape the call config before the model call: a frozen `LlmCallConfig` seed in, a replacement out (model switching, sampling overrides). Content is not shapeable here — every request is a pure function of the session log ([reconstructability RFC](../../../docs/rfc/implemented/architecture/2026-07-05-reconstructable-requests.md)); the loop logs whatever config the request actually uses as a `request/header*` event
 - `agent/step-result` — post-process the assembled assistant message before tool dispatch (validates what the log records)
 - `agent/turn-continuation` — override the continue/stop decision via `ContinuationDecision` = `{action:'stop'}` or `{action:'continue', reason?}` (a `continue` `reason` is recorded as next-step steering in the same turn — the typed `/goal` pattern). Force-continue `/loop`, force-stop budget guard.
 

packages/core/agent/src/types.ts

@@ -44,13 +44,17 @@
  */
 
 import type { Branded } from '@deepseek-ai/dsh-brand'
-import type { ContentBlock, GenerateOptions, Message, MessageSource } from '@deepseek-ai/dsh-llm'
+import type { ContentBlock, LlmCallConfig, Message, MessageSource } from '@deepseek-ai/dsh-llm'
 import type {} from '@deepseek-ai/dsh-system-prompt'
 
 /** Identifies one live agent in the registry. */
 export type AgentId = Branded<'AgentId'>
 
-/** Brand a string as an {@link AgentId}. */
+/**
+ * Brand a string as an {@link AgentId}.
+ * @param id - the raw agent id string.
+ * @returns the same string, branded (a compile-time cast — no runtime cost).
+ */
 export function AgentId(id: string): AgentId {
   return id as AgentId
 }
@@ -80,10 +84,22 @@ export interface AgentOptions {
   model?: string
 }
 
+/**
+ * Options for {@link Agent.send}/{@link Agent.steer}/{@link Agent.inject}. An
+ * absent `source` resolves to `{ kind: 'user' }`, so a plugin supplying content
+ * must label itself here or its message is recorded as a user prompt (see
+ * {@link HookContext} on why that label is load-bearing).
+ */
 export interface SendOptions {
   source?: MessageSource
 }
 
+/**
+ * An agent's lifecycle state, emitted on every transition as `agent/status`:
+ * `idle` (parked, waiting for queued work), `running` (a turn is in progress),
+ * `disposed` (terminal — no transition leaves it, and `send`/`steer`/`inject`
+ * throw).
+ */
 export type AgentStatus = 'idle' | 'running' | 'disposed'
 
 /**
@@ -345,18 +361,28 @@ declare module 'cordis' {
      */
     'agent/prompt-submit'(agent: Agent, content: ContentBlock[], source: MessageSource, next: () => Promise<PromptDecision>): Promise<PromptDecision>
     /**
-     * Waterfall: mutate the fully-assembled {@link GenerateOptions} before the
-     * model call (hooks, model switching, tool filtering, …). Call `next()` to
-     * delegate, or return without it to short-circuit. For surface mutation that
-     * must precede history derivation (compaction), use {@link agent/pre-step}
-     * instead — by the time this fires, `options.messages` is already derived.
+     * Waterfall: shape the step's call configuration — model switching,
+     * sampling overrides — by returning a replacement {@link LlmCallConfig}
+     * (the frozen seed is the config the loop would otherwise use). Config is
+     * ALL a listener shapes here: every request is a pure function of the
+     * session log (the reconstructability RFC), so model-visible content
+     * flows through the log channels — `inject()`, steering, prompt-submit
+     * `additionalContext`, prompt sections via `system-prompt/assemble` —
+     * never through request mutation, and the loop records whatever config
+     * the request actually uses as a `request/header*` event before dispatch.
+     * The step's messages are already snapshotted when this fires (the
+     * `step/start` boundary): an `inject()` from a listener here lands in the
+     * log but joins the NEXT request. For surface mutation that must precede
+     * the snapshot (compaction), use {@link agent/pre-step}. Call `next()` to
+     * delegate, or return an {@link LlmCallConfig} without it to
+     * short-circuit.
      * @param agent - the agent making the model call.
      * @param turn - the open turn number.
      * @param step - the step whose request this is.
-     * @param options - the assembled request; listeners return a transformed copy.
+     * @param config - the config the loop would use (frozen); return a replacement to switch.
      * @mode waterfall
      */
-    'agent/request'(agent: Agent, turn: number, step: number, options: GenerateOptions, next: () => Promise<GenerateOptions>): Promise<GenerateOptions>
+    'agent/request'(agent: Agent, turn: number, step: number, config: LlmCallConfig, next: () => Promise<LlmCallConfig>): Promise<LlmCallConfig>
     /**
      * Waterfall: post-process the assembled assistant {@link Message} before
      * tool dispatch (validation, content rewriting, …).

packages/core/agent/tests/verify-export-jsdoc.spec.ts

@@ -0,0 +1,555 @@
+/**
+ * Negative-path tests for the export-surface JSDoc gate
+ * (`scripts/verify-export-jsdoc.ts`).
+ *
+ * The gate's positive half runs against the real tree in CI (`pnpm run
+ * verify-export-jsdoc`, part of doc-sync). What that run cannot prove is that
+ * the walk REJECTS an undocumented surface the way it promises to — and that
+ * every deliberate exemption (heritage members, plugin-protocol slots,
+ * constructors, overload implementations, augmentation bodies, re-exports)
+ * actually holds. These tests drive `collectExportJsdocViolations()` against
+ * synthetic fixture packages, mirroring the gen-cordis-catalog negative
+ * tests.
+ */
+
+import { mkdirSync, mkdtempSync, rmSync, writeFileSync } from 'node:fs'
+import { tmpdir } from 'node:os'
+import { dirname, join } from 'node:path'
+import { afterEach, describe, expect, it } from 'vitest'
+import { collectExportJsdocViolations } from '../../../../scripts/verify-export-jsdoc.ts'
+
+const roots: string[] = []
+
+afterEach(() => {
+  while (roots.length) rmSync(roots.pop()!, { recursive: true, force: true })
+})
+
+/** Write fixture files under `packages/group/fix/src/` and return the scan root. */
+function fixture(files: Record<string, string>): string {
+  const root = mkdtempSync(join(tmpdir(), 'export-jsdoc-'))
+  roots.push(root)
+  for (const [rel, content] of Object.entries(files)) {
+    const abs = join(root, 'packages', 'group', 'fix', 'src', rel)
+    mkdirSync(dirname(abs), { recursive: true })
+    writeFileSync(abs, content)
+  }
+  return root
+}
+
+/** Single-file fixture shorthand: the content becomes `src/index.ts`. */
+const make = (content: string): string => fixture({ 'index.ts': content })
+
+describe('verify-export-jsdoc functions and consts', () => {
+  it('accepts a fully documented surface', () => {
+    expect(collectExportJsdocViolations(make(`
+/**
+ * Add one to a count.
+ * @param n - the count to bump.
+ * @returns the count plus one.
+ */
+export function bump(n: number): number { return n + 1 }
+
+/**
+ * Fire-and-forget (void needs no @returns).
+ * @param flag - whether to arm.
+ */
+export function poke(flag: boolean): void { void flag }
+
+/** The default retry budget. */
+export const RETRIES = 3
+
+/**
+ * Halve a count.
+ * @param n - the count to halve.
+ * @returns the count halved.
+ */
+export const halve = (n: number): number => n / 2
+`))).toEqual([])
+  })
+
+  it('flags an exported function with no JSDoc at all', () => {
+    expect(collectExportJsdocViolations(make(
+      'export function bare(): void {}\n',
+    ))).toEqual([expect.stringMatching(/exported function 'bare' .* has no JSDoc\./)])
+  })
+
+  it('flags a missing @param and a missing @returns', () => {
+    const violations = collectExportJsdocViolations(make(
+      '/** Docs without tags. */\nexport function f(x: number): number { return x }\n',
+    ))
+    expect(violations).toEqual([
+      expect.stringMatching(/exported function 'f' .* is missing @param x\./),
+      expect.stringMatching(/exported function 'f' .* is missing @returns \(return type: number\)\./),
+    ])
+  })
+
+  it('flags an unannotated (inferred) return type', () => {
+    expect(collectExportJsdocViolations(make(
+      '/**\n * Docs.\n * @param x - value.\n */\nexport function f(x: number) { return x }\n',
+    ))).toEqual([expect.stringMatching(/no return type annotation/)])
+  })
+
+  it('flags tags-only JSDoc with no description prose', () => {
+    expect(collectExportJsdocViolations(make(
+      '/**\n * @param x - value.\n */\nexport function f(x: number): void {}\n',
+    ))).toEqual([expect.stringMatching(/no description prose above its block tags/)])
+  })
+
+  it('flags a stale @param and a binding-pattern parameter', () => {
+    const violations = collectExportJsdocViolations(make(
+      '/**\n * Docs.\n * @param ghost - not real.\n */\nexport function f({ a }: { a: number }): void {}\n',
+    ))
+    expect(violations).toEqual([
+      expect.stringMatching(/parameter '\{ a \}' is a binding pattern; the export surface needs simple identifier parameters/),
+      expect.stringMatching(/@param ghost does not match any parameter \(stale tag\?\)/),
+    ])
+  })
+
+  it('exempts a `this` receiver annotation from @param', () => {
+    expect(collectExportJsdocViolations(make(
+      '/**\n * Docs.\n * @param x - value.\n */\nexport function f(this: object, x: number): void {}\n',
+    ))).toEqual([])
+  })
+
+  it('waives @returns for a declarator-annotated const but not an unannotated one', () => {
+    expect(collectExportJsdocViolations(make(`
+type Fn = (x: number) => number
+/**
+ * Uses the named signature.
+ * @param x - value.
+ */
+export const good: Fn = x => x
+/**
+ * No signature anywhere.
+ * @param x - value.
+ */
+export const bad = (x: number) => x
+`))).toEqual([expect.stringMatching(/exported const 'bad' .* has no return type annotation/)])
+  })
+
+  it('requires description prose on a non-function const', () => {
+    expect(collectExportJsdocViolations(make(
+      'export const LIMIT = 10\n',
+    ))).toEqual([expect.stringMatching(/exported const 'LIMIT' .* has no JSDoc\./)])
+  })
+})
+
+describe('verify-export-jsdoc type-level exports', () => {
+  it('requires description prose on interfaces, type aliases, and enums', () => {
+    const violations = collectExportJsdocViolations(make(
+      'export interface I { a: number }\nexport type T = number\nexport enum E { A }\n',
+    ))
+    expect(violations).toEqual([
+      expect.stringMatching(/exported interface 'I' .* has no JSDoc\./),
+      expect.stringMatching(/exported type 'T' .* has no JSDoc\./),
+      expect.stringMatching(/exported enum 'E' .* has no JSDoc\./),
+    ])
+  })
+
+  it('skips `declare module` augmentation bodies (the cordis gate owns them)', () => {
+    expect(collectExportJsdocViolations(make(
+      "declare module 'cordis' {\n  interface Events {\n    'fix/x'(): void\n  }\n}\nexport {}\n",
+    ))).toEqual([])
+  })
+})
+
+describe('verify-export-jsdoc export forms', () => {
+  it('resolves an `export { … }` list to the local declaration', () => {
+    expect(collectExportJsdocViolations(make(
+      'function f(): void {}\nexport { f }\n',
+    ))).toEqual([expect.stringMatching(/exported function 'f' .* has no JSDoc\./)])
+  })
+
+  it('does not treat a never-exported sibling declarator as surface (review round 2)', () => {
+    // `export { publicValue }` resolves to the whole variable statement; only
+    // the named declarator is surface — the gate must not demand JSDoc for
+    // the private sibling sharing the statement.
+    expect(collectExportJsdocViolations(make(
+      '/** The public knob. */\nconst publicValue = 1, privateHelper = 2\nexport { publicValue }\nvoid privateHelper\n',
+    ))).toEqual([])
+  })
+
+  it('unions declarators across multiple export lists over one statement (review round 2)', () => {
+    // Two lists each name one declarator of the same undocumented statement:
+    // both are surface (deduplicating on first resolution would drop `b`),
+    // while the never-exported `c` stays out.
+    const violations = collectExportJsdocViolations(make(
+      'const a = 1, b = 2, c = 3\nexport { a }\nexport { b }\nvoid c\n',
+    ))
+    expect(violations).toEqual([
+      expect.stringMatching(/exported const 'a' .* has no JSDoc\./),
+      expect.stringMatching(/exported const 'b' .* has no JSDoc\./),
+    ])
+  })
+
+  it('scopes a default-export identifier to its own declarator (review round 2)', () => {
+    // `export default` of an identifier reaches the statement through the
+    // same name lookup as an export list; the sibling stays private.
+    expect(collectExportJsdocViolations(make(
+      '/** The app entry. */\nconst app = 1, scratch = 2\nexport default app\nvoid scratch\n',
+    ))).toEqual([])
+  })
+
+  it('reports a re-exported module once, at its defining file', () => {
+    const violations = collectExportJsdocViolations(fixture({
+      'index.ts': "export * from './other.ts'\n",
+      'other.ts': 'export function f(): void {}\n',
+    }))
+    expect(violations).toEqual([expect.stringMatching(/other\.ts:1\) has no JSDoc\./)])
+  })
+
+  it('exempts overload implementations when the signatures are documented', () => {
+    expect(collectExportJsdocViolations(make(`
+/**
+ * From a number.
+ * @param x - the number.
+ * @returns its text.
+ */
+export function f(x: number): string
+/**
+ * From a flag.
+ * @param x - the flag.
+ * @returns its text.
+ */
+export function f(x: boolean): string
+export function f(x: number | boolean): string { return String(x) }
+`))).toEqual([])
+  })
+})
+
+describe('verify-export-jsdoc classes', () => {
+  it('flags an undocumented class, method, property, and accessor', () => {
+    const violations = collectExportJsdocViolations(make(`
+export class C {
+  state = 1
+  get view(): number { return this.state }
+  run(x: number): number { return x }
+}
+`))
+    expect(violations).toEqual([
+      expect.stringMatching(/exported class 'C' .* has no JSDoc\./),
+      expect.stringMatching(/exported class property 'C.state' .* has no JSDoc\./),
+      expect.stringMatching(/exported class accessor 'C.view' .* has no JSDoc\./),
+      expect.stringMatching(/exported class method 'C.run' .* has no JSDoc\./),
+    ])
+  })
+
+  it('exempts members declared by an extends/implements heritage type', () => {
+    expect(collectExportJsdocViolations(make(`
+/** Seam. */
+export abstract class Base {
+  /**
+   * Do it.
+   * @param x - input.
+   * @returns output.
+   */
+  abstract run(x: number): number
+}
+/** Iface. */
+export interface Sized {
+  /** Byte size. */
+  size: number
+}
+/** Impl. */
+export class Impl extends Base implements Sized {
+  size = 0
+  run(x: number): number { return x }
+}
+`))).toEqual([])
+  })
+
+  it('skips private/protected/#private members and constructors', () => {
+    expect(collectExportJsdocViolations(make(`
+/** Documented. */
+export class C {
+  #secret = 1
+  private hidden(): void {}
+  protected hook(): void {}
+  constructor(x: number) { void x }
+}
+`))).toEqual([])
+  })
+
+  it('exempts plugin-protocol statics but checks other statics', () => {
+    const violations = collectExportJsdocViolations(make(`
+/** Plugin. */
+export class C {
+  static Config = { a: 1 }
+  static inject = ['bash']
+  static reusable = true
+  static other = 1
+}
+`))
+    expect(violations).toEqual([expect.stringMatching(/exported class property 'C.other' .* has no JSDoc\./)])
+  })
+
+  it("covers a set accessor by the getter's doc", () => {
+    expect(collectExportJsdocViolations(make(`
+/** Documented. */
+export class C {
+  /** The current width. */
+  get width(): number { return 1 }
+  set width(_v: number) {}
+}
+`))).toEqual([])
+  })
+})
+
+describe('verify-export-jsdoc plugin protocol and namespaces', () => {
+  it('exempts top-level plugin-protocol exports', () => {
+    expect(collectExportJsdocViolations(make(`
+export const name = 'fix'
+export const inject = ['bash']
+export const reusable = true
+export const Config = { parse: true }
+export function apply(): void {}
+`))).toEqual([])
+  })
+
+  it('recurses into namespaces with qualified names and honors the merge idiom', () => {
+    const violations = collectExportJsdocViolations(make(`
+/** The plugin class. */
+export class Fix {}
+export namespace Fix {
+  export interface Config { a: number }
+}
+export namespace Loose {
+  export const x = 1
+}
+`))
+    expect(violations).toEqual([
+      expect.stringMatching(/exported interface 'Fix.Config' .* has no JSDoc\./),
+      expect.stringMatching(/exported namespace 'Loose' .* has no JSDoc\./),
+      expect.stringMatching(/exported const 'Loose.x' .* has no JSDoc\./),
+    ])
+  })
+})
+
+describe('verify-export-jsdoc fail-closed forms (review round 1)', () => {
+  it('checks the function contract on a non-identifier default export', () => {
+    expect(collectExportJsdocViolations(make(
+      '/** Doubles. */\nexport default (x: number): number => x * 2\n',
+    ))).toEqual([
+      expect.stringMatching(/default export .* is missing @param x\./),
+      expect.stringMatching(/default export .* is missing @returns \(return type: number\)\./),
+    ])
+    expect(collectExportJsdocViolations(make(
+      '/**\n * Doubles.\n * @param x - the input.\n * @returns twice the input.\n */\nexport default (x: number): number => x * 2\n',
+    ))).toEqual([])
+  })
+
+  it('treats an inline function-type annotation as the surface signature', () => {
+    expect(collectExportJsdocViolations(make(
+      '/** Maps a number. */\nexport declare const f: (x: number) => number\n',
+    ))).toEqual([
+      expect.stringMatching(/exported const 'f' .* is missing @param x\./),
+      expect.stringMatching(/exported const 'f' .* is missing @returns \(return type: number\)\./),
+    ])
+    expect(collectExportJsdocViolations(make(
+      '/**\n * Maps a number.\n * @param x - the input.\n * @returns the mapped value.\n */\nexport const f: (x: number) => number = v => v\n',
+    ))).toEqual([])
+  })
+
+  it('recurses into an ambient declare namespace where members export implicitly', () => {
+    expect(collectExportJsdocViolations(make(
+      'export declare namespace N {\n  function f(x: number): number\n}\n',
+    ))).toEqual([
+      expect.stringMatching(/exported namespace 'N' .* has no JSDoc\./),
+      expect.stringMatching(/exported function 'N.f' .* has no JSDoc\./),
+    ])
+  })
+
+  it('requires an export-import alias to document itself (its target may be unwalked)', () => {
+    expect(collectExportJsdocViolations(make(
+      '/** Holder. */\nexport namespace N {\n  /** The value. */\n  export const x = 1\n}\nexport import y = N.x\n',
+    ))).toEqual([expect.stringMatching(/exported alias 'y' .* has no JSDoc\./)])
+    expect(collectExportJsdocViolations(make(
+      'namespace N {\n  export const x = 1\n}\n/** Alias surfacing the internal counter. */\nexport import y = N.x\n',
+    ))).toEqual([])
+  })
+
+  it('refuses an export-import alias to a callable, class, or namespace target', () => {
+    const refusal = /exported alias 'g' .* aliases a callable, class, or namespace target/
+    expect(collectExportJsdocViolations(make(
+      'namespace N {\n  export function f(x: number): number { return x }\n}\n/** Alias. */\nexport import g = N.f\n',
+    ))).toEqual([expect.stringMatching(refusal)])
+    expect(collectExportJsdocViolations(make(
+      'namespace N {\n  export class C {\n    run(x: number): number { return x }\n  }\n}\n/** Alias. */\nexport import g = N.C\n',
+    ))).toEqual([expect.stringMatching(refusal)])
+    expect(collectExportJsdocViolations(make(
+      'namespace N {\n  export namespace Sub {\n    export function f(x: number): number { return x }\n  }\n}\n/** Alias. */\nexport import g = N.Sub\n',
+    ))).toEqual([expect.stringMatching(refusal)])
+  })
+
+  it('classifies wrapped function initializers and default exports (parens, satisfies)', () => {
+    expect(collectExportJsdocViolations(make(
+      'type Fn = (x: number) => number\n/** Wrapped. */\nexport const f = (((x: number): number => x)) satisfies Fn\n',
+    ))).toEqual([
+      expect.stringMatching(/exported const 'f' .* is missing @param x\./),
+      expect.stringMatching(/exported const 'f' .* is missing @returns \(return type: number\)\./),
+    ])
+    expect(collectExportJsdocViolations(make(
+      'type Fn = (x: number) => number\n/** Wrapped. */\nexport default (((x: number): number => x * 2) satisfies Fn)\n',
+    ))).toEqual([
+      expect.stringMatching(/default export .* is missing @param x\./),
+      expect.stringMatching(/default export .* is missing @returns \(return type: number\)\./),
+    ])
+  })
+
+  it('treats a single-call-signature type literal as the surface signature', () => {
+    expect(collectExportJsdocViolations(make(
+      '/** Maps. */\nexport declare const f: { (x: number): number }\n',
+    ))).toEqual([
+      expect.stringMatching(/exported const 'f' .* is missing @param x\./),
+      expect.stringMatching(/exported const 'f' .* is missing @returns \(return type: number\)\./),
+    ])
+  })
+
+  it('refuses a hybrid callable type literal instead of narrowing the check', () => {
+    expect(collectExportJsdocViolations(make(
+      '/** Hybrid. */\nexport declare const f: { (x: number): number; flush: () => void }\n',
+    ))).toEqual([expect.stringMatching(/exported const 'f'.*callable type literal is not gate-classifiable; extract a named type/)])
+  })
+
+  it('refuses an export-equals assignment instead of failing open', () => {
+    expect(collectExportJsdocViolations(make(
+      'const x = 1\nexport = x\n',
+    ))).toEqual([expect.stringMatching(/export-equals assignment .* is not a gate-supported export form/)])
+  })
+})
+
+describe('verify-export-jsdoc heritage refinement (review round 1)', () => {
+  it('requires @param for parameters the base member never names', () => {
+    const violations = collectExportJsdocViolations(make(`
+/** Seam. */
+export abstract class Base {
+  /**
+   * Do it.
+   * @param x - input.
+   * @returns output.
+   */
+  abstract run(x: number): number
+}
+/** Impl. */
+export class Impl extends Base {
+  override run(x: number, verbose?: boolean): number { return verbose ? x : -x }
+}
+`))
+    expect(violations).toEqual([expect.stringMatching(/exported class method 'Impl.run' .* is missing @param verbose\./)])
+  })
+
+  it('does not exempt a public override of a protected-only base member', () => {
+    expect(collectExportJsdocViolations(make(`
+/** Seam. */
+export abstract class Base {
+  /** Subclass hook. */
+  protected hook(): void {}
+}
+/** Impl. */
+export class Impl extends Base {
+  override hook(): void {}
+}
+`))).toEqual([expect.stringMatching(/exported class method 'Impl.hook' .* has no JSDoc\./)])
+  })
+
+  it('treats an underscore-prefixed rename of a base parameter as the same parameter', () => {
+    expect(collectExportJsdocViolations(make(`
+/** Seam. */
+export abstract class Base {
+  /**
+   * Load it.
+   * @param cwd - the working directory to scope the lookup.
+   * @returns the loaded value.
+   */
+  abstract load(cwd: string): number
+}
+/** Impl (ignores cwd). */
+export class Impl extends Base {
+  load(_cwd: string): number { return 1 }
+}
+`))).toEqual([])
+  })
+
+  it('flags a binding-pattern parameter an override adds beyond the base', () => {
+    expect(collectExportJsdocViolations(make(`
+/** Seam. */
+export abstract class Base {
+  /**
+   * Do it.
+   * @param x - input.
+   * @returns output.
+   */
+  abstract run(x: number): number
+}
+/** Impl. */
+export class Impl extends Base {
+  override run(x: number, { verbose }: { verbose?: boolean } = {}): number { return verbose ? x : -x }
+}
+`))).toEqual([expect.stringMatching(/exported class method 'Impl.run' .* is a binding pattern/)])
+  })
+
+  it('revives the @returns duty when an override grows a concrete result over a void base', () => {
+    const voidBase = `
+/** Seam. */
+export abstract class Base {
+  /** Do it (fire-and-forget). */
+  abstract run(): void
+}
+`
+    expect(collectExportJsdocViolations(make(`${voidBase}
+/** Impl. */
+export class Impl extends Base {
+  override run(): number { return 1 }
+}
+`))).toEqual([expect.stringMatching(/exported class method 'Impl.run' .* is missing @returns \(return type: number\)\./)])
+    expect(collectExportJsdocViolations(make(`${voidBase}
+/** Impl. */
+export class Impl extends Base {
+  /**
+   * Do it and count.
+   * @returns how many were done.
+   */
+  override run(): number { return 1 }
+}
+`))).toEqual([])
+  })
+
+  it('classifies an unannotated override return over a void base via the checker', () => {
+    const voidBase = `
+/** Seam. */
+export abstract class Base {
+  /** Do it (fire-and-forget). */
+  abstract run(): void
+}
+`
+    expect(collectExportJsdocViolations(make(`${voidBase}
+/** Impl. */
+export class Impl extends Base {
+  override run() { return 1 }
+}
+`))).toEqual([expect.stringMatching(/exported class method 'Impl.run' .* non-void result its heritage declaration does not document/)])
+    expect(collectExportJsdocViolations(make(`${voidBase}
+/** Impl (faithful void, no annotation needed). */
+export class Impl extends Base {
+  override run() {}
+}
+`))).toEqual([])
+  })
+
+  it('keeps the full exemption when the base return already carries the @returns duty', () => {
+    expect(collectExportJsdocViolations(make(`
+/** Seam. */
+export abstract class Base {
+  /**
+   * Count things.
+   * @returns the count.
+   */
+  abstract run(): number
+}
+/** Impl. */
+export class Impl extends Base {
+  override run(): number { return 1 }
+}
+`))).toEqual([])
+  })
+})

packages/core/session/README.md

@@ -9,6 +9,7 @@ Creates and holds event-sourced `Session` instances. Persistence is intentionall
 ### Public API
 
 - `ctx.sessions.create(id?: SessionId, options?: { seed?: SessionEvent[]; meta?: { cwd?: string; parentSession?: SessionId; createdAt?: number; seedLength?: number } }): Session` — Create a session. `options.seed` replays/forks an existing event log; `options.meta` attaches creation metadata (validated absolute `cwd`, `parentSession` lineage, seed boundary) as the immutable `SessionHeader`. The store fills `version`/`id` and defaults `createdAt` to now; a caller reconstructing a persisted session passes the original `createdAt` and persisted `seedLength` to preserve them. Disposed with the calling fiber.
+- `ctx.sessions.fork(source, boundary?, childSessionId?): Session` — Resolve a live session object or id, select a seed through the inclusive `boundary` event seq (default: current last event), require that boundary to be `turn/end`, and create a live child session with lineage metadata.
 - `ctx.sessions.get(id: SessionId): Session | undefined`
 - `ctx.sessions.list(): Session[]`
 
@@ -35,8 +36,9 @@ Creates and holds event-sourced `Session` instances. Persistence is intentionall
 Plain class (not a Cordis Service). Create via `ctx.sessions.create()`.
 
 - `session.append(type, data, opts?): SessionEvent` — synchronous, never blocks on I/O. **Throws** if `data` is not losslessly JSON-serializable (BigInt, function, symbol, undefined, non-finite number, circular ref, or an exotic object like Map/Set/Date) — the event log is the durable source of truth, so this invariant is enforced at the source (exported as `isJsonValue` for backends to reuse on their replay/fork entry points). A third parameter `opts: SurfaceIntent` carries surface metadata: `surfaceOp` controls how the event enters the surface linked list, and `sourceEventSeqs` records provenance (the seq numbers of events this one derives from). It is **required** for the five `SurfaceEventType` events (every message-producing event must declare how it joins the surface) and rejected by the compiler for non-surface types. The marker requirement is enforced two ways: the typed overload makes `opts` mandatory when `type` is a specific `SurfaceEventType` literal, AND `append` **throws** at runtime if a surface-eligible event arrives with no `surfaceOp` — covering the case where `type` widens to the `SessionEventType` union (a caller iterating raw events, where the conditional overload collapses to optional) so a marker-less message event can never silently land in the log and vanish from `deriveMessages()`.
-- `session.deriveMessages(): Message[]` — derive the LLM message history by walking the surface linked list (skipping non-surface events like chunks and boundaries; a `replace` shadows the nodes it covers). The surface is the single source of derived history — there is no raw-log fallback.
-- `session.surface: SurfaceManager` — the derived surface, lazily rebuilt from `surfaceOp` markers in the log. Processes only new events (delta) on each access — the log is append-only, so prior events never change.
+- `session.deriveMessages(): Message[]` — the LLM message history, CACHED: each surface node is projected exactly once, when first seen (O(new nodes) per call; a surface rewrite rebuilds via `surface.replaceGeneration`). Returns a fresh array snapshot per call over SHARED, deep-frozen `Message` objects — cloned once off the log at projection time, so a consumer can never mutate logged data (mutation throws). The surface is the single source of derived history — there is no raw-log fallback.
+- `session.deriveEventMessage(event): Message | null` — the per-event projection `deriveMessages()` folds: one event's derived message (an unfrozen clone), or `null` when it produces none (a non-surface event, or an empty-content `assistant/message` hosting only usage). External reconstructors and the dev invariant fold the same function over a log prefix's surface, so no two paths can disagree about what a request's messages were (the reconstructability RFC).
+- `session.surface: SurfaceManager` — the derived surface, lazily rebuilt from `surfaceOp` markers in the log. Processes only new events (delta) on each access — the log is append-only, so prior events never change. `surface.replaceGeneration` is the rewrite signal: bumped by every folded `replace` and by `invalidate()`, never reset, so an incremental consumer comparing generations cannot be fooled.
 - `session.events`, `session.seq`, `session.id`
 - `session.header: SessionHeader` — immutable creation metadata (`version`, `id`, `createdAt`, optional `cwd`/`parentSession`/`seedLength`). Kept out of the event log (a storage concern, not replayable state); a minimal header (stamped with the current `SESSION_FORMAT_VERSION`) is synthesized for bare `Session` construction.
 
@@ -47,9 +49,13 @@ Plain class (not a Cordis Service). Create via `ctx.sessions.create()`.
 - `SurfaceNode` — `{ seq: number; prev: number | null; next: number | null }`, one node in the surface linked list.
 - `isSurfaceEvent(event)` / `isSurfaceEligibleType(type)` — the first narrows a `SessionEvent` to a fully-formed surface node (type is surface-eligible AND `surfaceOp` present); the second is the type-only check (is this one of the five `SurfaceEventType` values?), used to detect a surface-eligible event MISSING its marker — e.g. when validating a seed/load log.
 
+### Request-header reconstruction (`request-header.ts`)
+
+The `request/header` (full `EpochHeader` snapshot with a `RequestHeaderReason`) and `request/header-delta` (system line-trim / name-keyed tools delta / whole config) events make the request envelope logged session state, so every conversation request is a pure function of the log. The pure trio reconstructs it: `foldRequestHeader(events)` folds a log (or any prefix) into the header in force; `diffHeader(prev, next)` encodes a change (undefined when equal); `applyHeaderDelta(prev, delta)` replays one. Writer contract: every logged delta is round-trip-verified (`apply(prev, delta)` deep-equals the new header) with a `'fallback'` snapshot when the encoding cannot express the change (a pure tool reordering), so folding never needs error recovery on a well-formed log. `canonicalHeader` pins the one representation of absence (empty system/tools ≡ absent fields).
+
 ### Session event vocabulary (`types.ts`)
 
-The append-only log's event types, enumerated member by member — payloads, surface badges, provenance — in the generated [persistence log event catalog](../../../docs/persistence-catalog/log-events.md). Token usage rides on `assistant/message.usage`; an operational error's step is on `turn/end.reason` for `kind: 'error'`.
+The append-only log's event types, enumerated member by member — payloads, surface badges, provenance — in the generated [persistence log event catalog](../../../docs/persistence-catalog.md). Token usage rides on `assistant/message.usage`; an operational error's step is on `turn/end.reason` for `kind: 'error'`.
 
 Merge-extensible via `SessionEventMap` — a plugin declaration-merges its own types (the compaction seam's `compact/*`, the hook bridges' `hook/*`); merged members appear in the same catalog.
 
@@ -67,9 +73,9 @@ Every `SessionEvent` carries two optional top-level fields (structural metadata)
 ### Extension points
 
 - Persistence plugins: subscribe to `session/event` (write-behind) and drain on `session/flush` (awaited) and fiber dispose. A durable backend reads the log and reloads it into a live session; the metadata seam (`SessionHeader`, `session.header`) is what such a backend stores beside the log.
-- Replay/fork: `ctx.sessions.create(id, { seed })` seeds a new session with an existing event log. The surface rebuilds deterministically from `surfaceOp` markers in the seeded events. The seed is validated to the SAME invariants `append` enforces — including that every surface-eligible event (`SurfaceEventType`) carries a `surfaceOp` marker — so a marker-less message event is rejected at construction rather than silently vanishing from `deriveMessages()` (the surface is the sole derivation path) on resume.
+- Replay/fork: `ctx.sessions.create(id, { seed })` seeds a new session with an existing event log. The surface rebuilds deterministically from `surfaceOp` markers in the seeded events. The seed is validated to the SAME always-on invariants `append` enforces — contiguous seqs, JSON-serializable data, and required `surfaceOp` markers on surface-eligible events — so marker-less message events are rejected at construction rather than silently vanishing from `deriveMessages()`. Broader turn-enclosure checks stay in `dsh-invariants` and persistence repair. Ordinary live-session forks use `ctx.sessions.fork(source, boundary?, childSessionId?)`, where `boundary` is the inclusive source event seq to fork through.
 - Compaction: the `dsh-compact-basic` plugin appends a `user/message` with `surfaceOp: { op: 'replace', start, end }` to shadow old surface nodes behind a summary checkpoint.
 
 ### What is NOT here (TODO)
 
-- **Session branching/tree** (pi-style entry tree) — deferred unless needed beyond seed-based forking.
+- **Session branching/tree** (pi-style entry tree) — deferred unless needed beyond boundary-based `fork()`.

packages/core/session/src/index.ts

@@ -8,11 +8,13 @@
 
 import { Context, Service } from 'cordis'
 import { isAbsolute } from 'node:path'
+import { deepFreeze } from '@deepseek-ai/dsh-llm'
 import type { ContentBlock, Message, MessageSource } from '@deepseek-ai/dsh-llm'
 import { SESSION_FORMAT_VERSION, SessionId } from './types.ts'
-import type { CreateSessionOptions, SessionEvent, SessionEventMap, SessionEventType, SessionHeader, SurfaceIntent, SurfaceEventType } from './types.ts'
+import type { CreateSessionOptions, EpochHeader, SessionEvent, SessionEventMap, SessionEventType, SessionHeader, SurfaceIntent, SurfaceEventType } from './types.ts'
 import { isJsonValue } from './json.ts'
 import { SurfaceManager, isSurfaceEligibleType } from './surface.ts'
+import { foldRequestHeader } from './request-header.ts'
 
 export * from './types.ts'
 export { isJsonValue } from './json.ts'
@@ -21,6 +23,7 @@ export { interruptedTurnClosers } from './repair.ts'
 export type { SurfaceNode } from './surface.ts'
 export { isSurfaceEvent, isSurfaceEligibleType } from './surface.ts'
 export { isToolPairingBalanced } from './tool-pairing.ts'
+export { applyHeaderDelta, canonicalHeader, diffHeader, foldRequestHeader, headerEquals } from './request-header.ts'
 
 declare module 'cordis' {
   interface Context {
@@ -150,10 +153,15 @@ export class Session {
     this.header = header ?? { version: SESSION_FORMAT_VERSION, id, createdAt: Date.now() }
   }
 
+  /**
+   * The append-only event log, exposed live by reference (readonly-typed, not
+   * a snapshot): later appends are visible through the same array.
+   */
   get events(): readonly SessionEvent[] {
     return this.log
   }
 
+  /** The next event's sequence number — always the log length (the `seq = log.length` contiguity contract). */
   get seq(): number {
     return this.log.length
   }
@@ -172,6 +180,9 @@ export class Session {
    *   declare how it joins the surface, the sole source of derived history) and
    *   rejected by the compiler for non-surface types like `turn/start` or
    *   `assistant/chunk`.
+   * @returns the logged event — its assigned `seq`/`time` plus the SNAPSHOT of
+   *   `data` that entered the log, so reading `event.data` back sees the logged
+   *   value, never the caller's still-mutable input.
    * @throws if `data` is not losslessly JSON-serializable (BigInt, function,
    *   symbol, undefined, non-finite number, circular ref, or an exotic object
    *   like Map/Set/Date). The event log is the durable source of truth, so this
@@ -234,53 +245,93 @@ export class Session {
     return event
   }
 
+  /** Cached fold of the request-header events — see {@link requestHeader}. */
+  private headerFold: EpochHeader | undefined
+  /** Log position (events consumed) the header fold has reached. */
+  private headerFoldSeq = 0
+
+  /**
+   * The {@link EpochHeader} in force after the log's last header event — the
+   * header the NEXT request will be compared against — or undefined before
+   * the first `request/header` snapshot. The live, incrementally-maintained
+   * form of `foldRequestHeader(session.events)`: each header event is folded
+   * once, when first seen, so a per-step read costs O(new events).
+   * @returns the folded header, or undefined when no header event exists yet.
+   */
+  requestHeader(): EpochHeader | undefined {
+    if (this.headerFoldSeq < this.log.length) {
+      // Frozen on update: the fold is session state exposed by reference — a
+      // consumer mutating it in place (instead of building a replacement)
+      // would desync every later comparison against the log, so mutation
+      // throws instead.
+      this.headerFold = deepFreeze(foldRequestHeader(this.log.slice(this.headerFoldSeq), this.headerFold))
+      this.headerFoldSeq = this.log.length
+    }
+    return this.headerFold
+  }
+
+  /** The derived-message cache: frozen projections, extended per unseen node. */
+  private derived: Message[] = []
+  /** Surface position (nodes projected) the cache has reached. */
+  private derivedNodes = 0
+  /** {@link SurfaceManager.replaceGeneration} the cache was built under. */
+  private derivedGeneration = 0
+
   /**
    * Derive the LLM message history by walking the session surface — the linked
    * list of message-producing events maintained by `surfaceOp` markers. The
    * surface is the single source of derived history: every message-producing
    * append records its `surfaceOp`, so a raw event with no marker (a chunk, a
    * turn boundary) is correctly absent, and a compaction `replace` deletes the
-   * shadowed nodes from the derivation.
+   * shadowed nodes from the derivation. The projection rules are
+   * {@link deriveEventMessage}, folded per node.
    *
-   * - `user/message` → user message
-   * - `assistant/message` → assistant message (chunks are skipped — they are
-   *   replay/UI data; the assembled message is authoritative for history). An
-   *   EMPTY-content assistant/message is skipped: a max-tokens step cut off with
-   *   no content still records an assistant/message to host its `usage`, but a
-   *   content-less assistant turn must not enter the provider transcript.
-   * - `tool/result` → user message carrying a tool-result block
-   * - `context/message` / `steering/message` → tagged synthetic user messages
-   *   at their chronological position
-   *
-   * The returned `content` is **deep-cloned** off the logged events: the loop
-   * hands these messages into the mutable `agent/request` waterfall and on to
-   * adapters, where mutating the request is sanctioned — but the session log
-   * is append-only by contract. Cloning at this boundary keeps in-flight
-   * mutation from reaching back and rewriting history (which would silently
-   * break replay equivalence). Cost is one structured clone per step,
-   * negligible next to a model call.
+   * CACHED: each surface node is projected exactly once, when first seen — a
+   * call costs O(new nodes), and a surface rewrite (a `replace`;
+   * {@link SurfaceManager.replaceGeneration}) rebuilds. The returned array is
+   * a fresh snapshot per call (later appends never grow an array a caller
+   * already holds); the `Message` objects in it are SHARED and **deep-frozen**
+   * — cloned once off the log at projection time, so consumers can never
+   * mutate logged data, and mutation attempts throw instead of silently
+   * diverging replay from history.
+   * @returns a fresh array of the shared, frozen derived history.
    */
   deriveMessages(): Message[] {
-    const messages: Message[] = []
-    for (const node of this.surface.nodes) {
+    const nodes = this.surface.nodes
+    const generation = this.surface.replaceGeneration
+    if (generation !== this.derivedGeneration) {
+      this.derived = []
+      this.derivedNodes = 0
+      this.derivedGeneration = generation
+    }
+    for (const node of nodes.slice(this.derivedNodes)) {
       // Surface nodes are built from this.log — node.seq is always a valid
       // index by construction. The non-null assertion expresses that invariant.
       // eslint-disable-next-line @typescript-eslint/no-non-null-assertion
-      const msg = this._deriveOneMessage(this.log[node.seq]!)
+      const msg = this.deriveEventMessage(this.log[node.seq]!)
       // A surface node is one of the five message-producing types, but an
       // empty-content assistant/message (a max-tokens step that hosts only
       // usage) derives to null and must not enter the transcript.
-      if (msg) messages.push(msg)
+      if (msg) this.derived.push(deepFreeze(msg))
     }
-    return messages
+    this.derivedNodes = nodes.length
+    return [...this.derived]
   }
 
   /**
-   * Derive a single LLM message from one surface event, or null if it produces
-   * no message (an empty-content assistant/message that exists only to host
-   * usage).
+   * Project a single event into the LLM message it derives to, or null when
+   * it produces none — a non-surface event (chunk, boundary, log-only record)
+   * or an empty-content assistant/message (which exists only to host usage).
+   * The per-node pure function {@link deriveMessages} folds over the surface;
+   * an external reconstructor (or the dev invariant) folds the same function
+   * over a log prefix's surface to rebuild the exact messages any request was
+   * built from (the reconstructability RFC). The returned `content` is
+   * deep-cloned off the logged event: the log is append-only by contract, so
+   * no live reference to logged data leaves this boundary.
+   * @param event - the event to project.
+   * @returns the derived message, or null when the event produces none.
    */
-  private _deriveOneMessage(event: SessionEvent): Message | null {
+  deriveEventMessage(event: SessionEvent): Message | null {
     // Intentionally non-exhaustive: only message-producing events derive
     // history; turn/step boundaries, chunks, usage, and errors are
     // trace/replay data.
@@ -311,13 +362,40 @@ export class Session {
         const { content, source } = event.data
         return { role: 'user', content: renderTagged('steering', structuredClone(content), source) }
       }
-      /* v8 ignore next 2 -- unreachable: only surface nodes (the 5 message-producing types) reach here */
       default:
+        // A non-surface event (boundary, chunk, log-only record) projects to
+        // no message. Merge-extensible union: no assertNever here.
         return null
     }
   }
 }
 
+/** A fork source: either the live session object or its live store id. */
+export type SessionForkSource = Session | SessionId
+
+/**
+ * Rejection codes for session forking: the fork source id is unknown to the
+ * live store (`SESSION_NOT_FOUND`) or names a session object that is not the
+ * store's live instance (`SESSION_NOT_LIVE`); the requested child id is
+ * already taken (`SESSION_ALREADY_EXISTS`); the boundary is not a contiguous
+ * existing seq (`INVALID_BOUNDARY`); or the boundary event is not a
+ * `turn/end` — a fork must cut on a closed turn (`OPEN_TURN`).
+ */
+export type SessionForkErrorCode =
+  | 'SESSION_NOT_FOUND'
+  | 'SESSION_NOT_LIVE'
+  | 'SESSION_ALREADY_EXISTS'
+  | 'INVALID_BOUNDARY'
+  | 'OPEN_TURN'
+
+/** Typed error for session fork rejections. */
+export class SessionForkError extends Error {
+  constructor(message: string, public readonly code: SessionForkErrorCode) {
+    super(message)
+    this.name = 'SessionForkError'
+  }
+}
+
 /**
  * In-memory session store (`ctx.sessions`).
  *
@@ -452,6 +530,92 @@ export class SessionStore extends Service {
   list(): Session[] {
     return [...this.store.values()]
   }
+
+  /**
+   * Create a live child session from a turn-enclosed prefix of a live source.
+   * `boundary` is an inclusive source event seq; omitted means the source's
+   * current last event. A non-empty selected slice must end at `turn/end`.
+   *
+   * @param source - Live source session object or id.
+   * @param boundary - Inclusive source event seq to fork through; omitted means
+   *   the source's current last event, and omitted on an empty source forks an
+   *   empty child.
+   * @param childSessionId - Optional child session id; omitted delegates to
+   *   `SessionStore`'s id policy.
+   * @returns The created live child session.
+   */
+  fork(source: SessionForkSource, boundary?: number, childSessionId?: SessionId): Session {
+    if (childSessionId !== undefined && this.get(childSessionId) !== undefined) {
+      throw new SessionForkError(`session "${childSessionId}" already exists`, 'SESSION_ALREADY_EXISTS')
+    }
+    const liveSource = this._resolveForkSource(source)
+    const seed = this._forkSeed(liveSource, boundary)
+    return this.create(childSessionId, {
+      seed,
+      meta: {
+        ...liveSource.header.cwd !== undefined ? { cwd: liveSource.header.cwd } : {},
+        parentSession: liveSource.id,
+        seedLength: seed.length,
+      },
+    })
+  }
+
+  private _forkSeed(session: Session, requestedBoundary: number | undefined): SessionEvent[] {
+    const events = session.events
+    const lastEvent = events.at(-1)
+    let boundary: number
+    if (requestedBoundary !== undefined) {
+      boundary = requestedBoundary
+    } else {
+      if (lastEvent === undefined) return []
+      boundary = lastEvent.seq
+    }
+    if (!Number.isSafeInteger(boundary) || boundary < 0) {
+      throw new SessionForkError(
+        `fork boundary for session "${session.id}" must be a non-negative safe integer, got ${String(boundary)}`,
+        'INVALID_BOUNDARY',
+      )
+    }
+    if (boundary >= events.length) {
+      const lastSeq = events.at(-1)?.seq
+      throw new SessionForkError(
+        `fork boundary ${boundary} does not exist in session "${session.id}" (last seq: ${lastSeq ?? 'none'})`,
+        'INVALID_BOUNDARY',
+      )
+    }
+
+    const boundaryEvent = events[boundary]
+    if (boundaryEvent === undefined || boundaryEvent.seq !== boundary) {
+      throw new SessionForkError(
+        `fork boundary ${boundary} does not match a contiguous event seq in session "${session.id}"`,
+        'INVALID_BOUNDARY',
+      )
+    }
+    if (boundaryEvent.type !== 'turn/end') {
+      throw new SessionForkError(
+        `fork boundary ${boundary} in session "${session.id}" must be turn/end, got ${boundaryEvent.type}`,
+        'OPEN_TURN',
+      )
+    }
+
+    return events.slice(0, boundary + 1).map(event => structuredClone(event))
+  }
+
+  private _resolveForkSource(source: SessionForkSource): Session {
+    if (typeof source === 'string') {
+      const session = this.get(source)
+      if (session === undefined) throw new SessionForkError(`session "${source}" not found`, 'SESSION_NOT_FOUND')
+      return session
+    }
+
+    const live = this.get(source.id)
+    if (live === undefined) {
+      throw new SessionForkError(`session "${source.id}" not found`, 'SESSION_NOT_FOUND')
+    }
+    if (live !== source) throw new SessionForkError(`session "${source.id}" is not the live store instance`, 'SESSION_NOT_LIVE')
+    return source
+  }
+
 }
 
 export default SessionStore

packages/core/session/src/json.ts

@@ -40,6 +40,10 @@ export type JsonValue = null | boolean | number | string | JsonValue[] | { [key:
  * hiding under a symbol/non-enumerable key cannot make the round-trip lossy.
  * Getters are invoked during the check (again as `JSON.stringify` would), so the
  * contract is for plain data records, not objects with side-effecting accessors.
+ * @param value - the candidate event data to test.
+ * @param seen - objects on the current descent path, for circular-reference
+ *   detection; the recursion threads it — callers omit it.
+ * @returns true when `value` survives a JSON round-trip losslessly.
  */
 export function isJsonValue(value: unknown, seen: Set<object> = new Set()): boolean {
   if (value === null) return true

packages/core/session/src/repair.ts

@@ -54,6 +54,8 @@ import type { SessionEvent } from './types.ts'
  * Only the LAST turn can be open: the invariants plugin guarantees a `turn/end`
  * before any later `turn/start`, so an interior open turn is impossible in a
  * valid committed log. Likewise at most one step is open within that turn.
+ * @param events - the loaded durable log to scan (a valid committed prefix, possibly with a crash tail).
+ * @returns the synthetic closer events to append after `events`, in order; empty when the log is already balanced.
  */
 export function interruptedTurnClosers(events: readonly SessionEvent[]): SessionEvent[] {
   let openTurn: number | null = null

packages/core/session/src/request-header.ts

@@ -0,0 +1,191 @@
+/**
+ * Request-header reconstruction utilities: the pure fold/diff/apply trio over
+ * the `request/header` / `request/header-delta` session events. Anyone
+ * holding a session log reconstructs the {@link EpochHeader} any request was
+ * built under by folding these events in log order; the loop uses the same
+ * functions to decide whether a step's header changed and to encode the
+ * change. Deltas are an encoding optimization with a safety valve — the
+ * writer round-trip-verifies every delta before appending and falls back to
+ * a full snapshot when the encoding cannot express the change — so folding
+ * never needs error recovery on a well-formed log.
+ *
+ * @module dsh-session/request-header
+ */
+
+import { callConfigEquals } from '@deepseek-ai/dsh-llm'
+import type { LlmCallConfig, ToolSchema } from '@deepseek-ai/dsh-llm'
+import type { EpochHeader, SessionEvent, SystemDelta, ToolsDelta } from './types.ts'
+
+/**
+ * Normalize a header to canonical form: an empty system prompt and an empty
+ * tool list become ABSENT fields, matching how requests are built (both
+ * request-build spreads skip empty values). Diff, fold, and comparison all
+ * operate on canonical headers, so "no system prompt" has exactly one
+ * representation.
+ * @param header - the header to normalize (not mutated).
+ * @returns the canonical header.
+ */
+export function canonicalHeader(header: EpochHeader): EpochHeader {
+  return {
+    config: header.config,
+    ...header.system !== undefined && header.system.length > 0 ? { system: header.system } : {},
+    ...header.tools !== undefined && header.tools.length > 0 ? { tools: header.tools } : {},
+  }
+}
+
+/** Split a canonical (possibly absent) system prompt into lines; absence is zero lines. */
+function systemLines(system: string | undefined): string[] {
+  return system === undefined ? [] : system.split('\n')
+}
+
+/** Join lines back into a canonical system value; zero lines is absence. */
+function joinSystem(lines: string[]): string | undefined {
+  return lines.length === 0 ? undefined : lines.join('\n')
+}
+
+/**
+ * Compute the line-level {@link SystemDelta} between two canonical system
+ * prompts: trim the common prefix and (non-overlapping) common suffix, and
+ * carry the replacement lines between them. Deterministic and library-free;
+ * with nothing shared it degenerates to a full replacement.
+ */
+function diffSystem(prev: string | undefined, next: string | undefined): SystemDelta {
+  const a = systemLines(prev)
+  const b = systemLines(next)
+  let keepStart = 0
+  while (keepStart < a.length && keepStart < b.length && a[keepStart] === b[keepStart]) keepStart += 1
+  let keepEnd = 0
+  while (
+    keepEnd < a.length - keepStart &&
+    keepEnd < b.length - keepStart &&
+    a[a.length - 1 - keepEnd] === b[b.length - 1 - keepEnd]
+  ) keepEnd += 1
+  return { keepStart, keepEnd, insert: b.slice(keepStart, b.length - keepEnd) }
+}
+
+/** Apply a {@link SystemDelta} to a canonical system prompt. */
+function applySystem(prev: string | undefined, delta: SystemDelta): string | undefined {
+  const a = systemLines(prev)
+  return joinSystem([...a.slice(0, delta.keepStart), ...delta.insert, ...a.slice(a.length - delta.keepEnd)])
+}
+
+/** Canonical JSON equality for tool schemas — sound because schemas are
+ * JSON-serializable by construction and both sides come from the same
+ * assembly path, so key insertion order matches when the values do. */
+function sameSchema(a: ToolSchema, b: ToolSchema): boolean {
+  return JSON.stringify(a) === JSON.stringify(b)
+}
+
+/**
+ * Compute the name-keyed {@link ToolsDelta} between two canonical tool lists.
+ * A pure reordering produces an empty delta — the writer's round-trip guard
+ * catches that case and records a snapshot instead.
+ */
+function diffTools(prev: readonly ToolSchema[], next: readonly ToolSchema[]): ToolsDelta {
+  const prevByName = new Map(prev.map(tool => [tool.name, tool]))
+  const nextNames = new Set(next.map(tool => tool.name))
+  return {
+    added: next.filter(tool => !prevByName.has(tool.name)),
+    removed: prev.filter(tool => !nextNames.has(tool.name)).map(tool => tool.name),
+    changed: next.filter((tool) => {
+      const before = prevByName.get(tool.name)
+      return before !== undefined && !sameSchema(before, tool)
+    }),
+  }
+}
+
+/** Apply a {@link ToolsDelta} to a canonical tool list: drop removed, replace changed in place, append added. */
+function applyTools(prev: readonly ToolSchema[], delta: ToolsDelta): ToolSchema[] {
+  const removed = new Set(delta.removed)
+  const changedByName = new Map(delta.changed.map(tool => [tool.name, tool]))
+  const kept = prev
+    .filter(tool => !removed.has(tool.name))
+    .map(tool => changedByName.get(tool.name) ?? tool)
+  return [...kept, ...delta.added]
+}
+
+/**
+ * Field-wise equality over canonical headers — the cheap comparison the
+ * writer's round-trip guard runs (`applyHeaderDelta(prev, delta)` must equal
+ * the intended header) and the loop runs to skip logging an unchanged header.
+ * Tools compare per-schema IN ORDER (canonical JSON), so a pure reordering is
+ * correctly unequal.
+ * @param a - one canonical header.
+ * @param b - the other.
+ * @returns whether config, system, and tools (in order) all match.
+ */
+export function headerEquals(a: EpochHeader, b: EpochHeader): boolean {
+  if (!callConfigEquals(a.config, b.config) || a.system !== b.system) return false
+  const at = a.tools ?? []
+  const bt = b.tools ?? []
+  return at.length === bt.length && at.every((tool, i) => sameSchema(tool, bt[i] as ToolSchema))
+}
+
+/**
+ * Compute the `request/header-delta` payload between two canonical headers,
+ * or undefined when they are equal. The caller MUST round-trip the result
+ * ({@link applyHeaderDelta} on `prev` deep-equals `next`) before logging it —
+ * the encoding cannot express every change (a pure tool reordering) — and
+ * fall back to a full `request/header` snapshot when the check fails.
+ * @param prev - the folded header the log currently implies.
+ * @param next - the header the next request will actually use.
+ * @returns the delta payload, or undefined when nothing changed.
+ */
+export function diffHeader(
+  prev: EpochHeader, next: EpochHeader,
+): { system?: SystemDelta; tools?: ToolsDelta; config?: LlmCallConfig } | undefined {
+  const delta: { system?: SystemDelta; tools?: ToolsDelta; config?: LlmCallConfig } = {}
+  if (prev.system !== next.system) delta.system = diffSystem(prev.system, next.system)
+  const prevTools = prev.tools ?? []
+  const nextTools = next.tools ?? []
+  if (JSON.stringify(prevTools) !== JSON.stringify(nextTools)) delta.tools = diffTools(prevTools, nextTools)
+  if (!callConfigEquals(prev.config, next.config)) delta.config = next.config
+  return Object.keys(delta).length > 0 ? delta : undefined
+}
+
+/**
+ * Apply a `request/header-delta` payload to a canonical header, producing the
+ * canonical header it encodes. Total for well-formed logs (the writer only
+ * appends round-trip-verified deltas).
+ * @param prev - the folded header before the delta.
+ * @param delta - the logged delta payload.
+ * @returns the canonical header after the delta.
+ */
+export function applyHeaderDelta(
+  prev: EpochHeader, delta: { system?: SystemDelta; tools?: ToolsDelta; config?: LlmCallConfig },
+): EpochHeader {
+  const system = delta.system !== undefined ? applySystem(prev.system, delta.system) : prev.system
+  const tools = delta.tools !== undefined ? applyTools(prev.tools ?? [], delta.tools) : prev.tools
+  return canonicalHeader({
+    config: delta.config ?? prev.config,
+    ...system !== undefined ? { system } : {},
+    ...tools !== undefined ? { tools } : {},
+  })
+}
+
+/**
+ * Fold the header events of a log (or any prefix of one) into the
+ * {@link EpochHeader} in force after the last of them: each
+ * `request/header` snapshot replaces the state, each `request/header-delta`
+ * amends it. The pure, offline form of reconstruction — external tooling and
+ * the dev invariant both use it; the live session tracks the same fold
+ * incrementally.
+ * @param events - session events in log order (non-header events are skipped).
+ * @param from - a previously folded state to continue from (the live session's
+ *   incremental cursor); omit to fold from nothing.
+ * @returns the folded header, or undefined when no header event exists yet.
+ */
+export function foldRequestHeader(events: readonly SessionEvent[], from?: EpochHeader): EpochHeader | undefined {
+  let state: EpochHeader | undefined = from
+  for (const event of events) {
+    if (event.type === 'request/header') {
+      state = canonicalHeader(event.data.header)
+    } else if (event.type === 'request/header-delta') {
+      if (state === undefined) {
+        throw new Error(`request/header-delta at seq ${event.seq} before any request/header snapshot: corrupt log`)
+      }
+      state = applyHeaderDelta(state, event.data)
+    }
+  }
+  return state
+}

packages/core/session/src/surface.ts

@@ -29,6 +29,8 @@ const SURFACE_EVENT_TYPES = new Set<string>([
  * surface-eligible event that is MISSING its mandatory marker (e.g. validating
  * a seed/load log); use {@link isSurfaceEvent} to narrow to a fully-formed
  * {@link SurfaceEvent} with `surfaceOp` present.
+ * @param type - the event type string to test.
+ * @returns true when the type is one of the five message-producing types.
  */
 export function isSurfaceEligibleType(type: string): boolean {
   return SURFACE_EVENT_TYPES.has(type)
@@ -38,6 +40,8 @@ export function isSurfaceEligibleType(type: string): boolean {
  * Narrow a {@link SessionEvent} to {@link SurfaceEvent}: checks that the
  * event's `type` is surface-eligible AND that `surfaceOp` is present.
  * The narrowed type has mandatory {@link SurfaceOp}.
+ * @param event - the event to narrow.
+ * @returns true when the event is surface-eligible and carries its `surfaceOp` marker.
  */
 export function isSurfaceEvent(event: SessionEvent): event is SurfaceEvent {
   if (!SURFACE_EVENT_TYPES.has(event.type)) return false
@@ -72,6 +76,9 @@ export class SurfaceManager {
   /** The last processed seq. -1 forces a full rebuild on first access. */
   private _lastProcessedSeq = -1
 
+  /** Rewrite generation — see {@link replaceGeneration}. */
+  private _replaceGeneration = 0
+
   constructor(private log: readonly SessionEvent[]) {}
 
   /**
@@ -83,6 +90,23 @@ export class SurfaceManager {
     this._lastProcessedSeq = -1
     this._nodes = []
     this._nodeBySeq.clear()
+    // A wholesale rebuild is a rewrite: bump the generation so incremental
+    // consumers (the session's derived-message cache) discard their view.
+    this._replaceGeneration += 1
+  }
+
+  /**
+   * The surface's rewrite generation: bumped by every folded `replace` op and
+   * by {@link invalidate}. A replace is the ONE operation that rewrites the
+   * surface non-monotonically, so an incremental consumer of {@link nodes}
+   * (the session's derived-message cache) compares this between visits — an
+   * unchanged generation guarantees every node it has not seen is a pure tail
+   * append; a changed one means its view must rebuild. Monotonic: it never
+   * moves backwards, so comparisons cannot be fooled by a re-fold.
+   */
+  get replaceGeneration(): number {
+    if (this._lastProcessedSeq < this.log.length - 1) this._processDelta()
+    return this._replaceGeneration
   }
 
   /** The surface nodes in linked-list order (head to tail). */
@@ -155,5 +179,6 @@ export class SurfaceManager {
     if (nextNode) nextNode.prev = newSeq
     this._nodes.splice(startIdx, 0, newNode)
     this._nodeBySeq.set(newSeq, newNode)
+    this._replaceGeneration += 1
   }
 }

packages/core/session/src/tool-pairing.ts

@@ -74,6 +74,12 @@ function nodeDelta(event: SessionEvent): number {
  * surface successor (`SurfaceNode.next`), or `null` when `end` is the tail —
  * for the cut after `end`.
  *
+ * @param nodes - the surface linked list in head→tail order.
+ * @param events - the session log each node's `seq` indexes into.
+ * @param beforeSeq - names the cut (the node it sits immediately before);
+ *   `null` — or any seq not on the surface — means the after-tail cut.
+ * @returns true when every `tool-call` before the cut is answered before it
+ *   (the unanswered-call depth at the cut is zero).
  * @throws if the surface prefix drives the unanswered-call depth negative — a
  *   `tool/result` with no preceding open `tool-call` on the surface. That is a
  *   corrupt surface (a structural invariant violation), surfaced loudly here

packages/core/session/src/types.ts

@@ -1,10 +1,14 @@
 import type { Branded } from '@deepseek-ai/dsh-brand'
-import type { CallId, ContentBlock, MessageSource, StreamChunk, TokenUsage } from '@deepseek-ai/dsh-llm'
+import type { CallId, ContentBlock, LlmCallConfig, MessageSource, StreamChunk, TokenUsage, ToolSchema } from '@deepseek-ai/dsh-llm'
 
 /** Identifies one session in the store (and its persistence artifacts). */
 export type SessionId = Branded<'SessionId'>
 
-/** Brand a string as a {@link SessionId}. */
+/**
+ * Brand a string as a {@link SessionId}.
+ * @param id - the raw session id string.
+ * @returns the same string, branded (a compile-time cast — no runtime cost).
+ */
 export function SessionId(id: string): SessionId {
   return id as SessionId
 }
@@ -102,6 +106,7 @@ export interface TurnTriggerMap {
   injection: { kind: 'injection'; source: MessageSource }
 }
 
+/** The union over {@link TurnTriggerMap} — what started a turn; plugins extend it by merging variants into the map. */
 export type TurnTrigger = TurnTriggerMap[keyof TurnTriggerMap]
 
 /**
@@ -156,6 +161,7 @@ export interface TurnEndReasonMap {
   interrupted: { kind: 'interrupted' }
 }
 
+/** The union over {@link TurnEndReasonMap} — why a turn ended; plugins extend it by merging variants into the map. */
 export type TurnEndReason = TurnEndReasonMap[keyof TurnEndReasonMap]
 
 /**
@@ -176,6 +182,67 @@ export interface TodoItem {
   status: 'pending' | 'in_progress' | 'completed'
 }
 
+/**
+ * The request header: everything about an LLM request besides its message
+ * content — the call configuration plus the rendered system prompt and tool
+ * schemas. Logged session state (the reconstructability RFC): a
+ * {@link SessionEventMap} `request/header` snapshot installs one, a
+ * `request/header-delta` amends it, and folding those events over the log
+ * (`foldRequestHeader`) reconstructs the header any request was built under.
+ * Canonical form: an empty system prompt and an empty tool list are ABSENT
+ * fields, matching how requests are built.
+ */
+export interface EpochHeader {
+  /** The conversation's call configuration (model + sampling scalars). */
+  config: LlmCallConfig
+  /** Rendered system prompt text; absent for a system-less request. */
+  system?: string
+  /** Assembled tool schemas; absent for a tool-less request. */
+  tools?: ToolSchema[]
+}
+
+/**
+ * Why a `request/header` snapshot was appended: `'initial'` — the log's first
+ * header (a new conversation); `'resume'` — a loop instance's first request
+ * over a log that already has header events (process restart, fork seed);
+ * `'fallback'` — a mid-run change the delta encoding could not round-trip
+ * (e.g. a pure tool reordering), recorded whole instead.
+ */
+export type RequestHeaderReason = 'initial' | 'resume' | 'fallback'
+
+/**
+ * Line-level edit of the system prompt: keep the first `keepStart` and last
+ * `keepEnd` lines of the previous text, with `insert` replacing everything
+ * between. Computed as a common-prefix/common-suffix trim — deterministic,
+ * library-free, degenerating to a full replacement when nothing is shared.
+ * Absence is encoded as zero lines (the canonical form has no empty-string
+ * system), so a transition to or from "no system prompt" round-trips.
+ */
+export interface SystemDelta {
+  /** Lines kept from the start of the previous system prompt. */
+  keepStart: number
+  /** Lines kept from the end of the previous system prompt. */
+  keepEnd: number
+  /** Lines replacing everything between the kept edges. */
+  insert: string[]
+}
+
+/**
+ * Tool-set edit keyed by tool name (names are unique — the registry rejects
+ * duplicates): `removed` names drop, `changed` schemas replace their
+ * predecessor in place, `added` schemas append at the end. A change this
+ * encoding cannot express (a pure reordering) fails the writer's round-trip
+ * guard and is recorded as a `'fallback'` snapshot instead.
+ */
+export interface ToolsDelta {
+  /** Schemas appended to the end of the tool list. */
+  added: ToolSchema[]
+  /** Names of schemas dropped from the tool list. */
+  removed: string[]
+  /** Schemas replacing the same-named predecessor in place. */
+  changed: ToolSchema[]
+}
+
 /**
  * The session event vocabulary — the append-only source of truth for an
  * agent's whole interaction history. The LLM message history is *derived*
@@ -274,8 +341,33 @@ export interface SessionEventMap {
    * cordis-catalog row.
    */
   'todo/write': { todos: TodoItem[] }
+  /**
+   * Full snapshot of the {@link EpochHeader} the NEXT request is built under,
+   * with the {@link RequestHeaderReason} it was recorded whole. Appended by
+   * the loop inside the step, before dispatch, on a loop instance's first
+   * request-building step (`'initial'`/`'resume'`) or when a delta failed its
+   * round-trip guard (`'fallback'`); always records what the request actually
+   * used, post-`agent/request`. Anchors the header fold: reconstruction reads
+   * the latest snapshot and applies the deltas after it. NOT a
+   * {@link SurfaceEventType}: it produces no LLM message — it is the request
+   * envelope, logged so every request is a pure function of the session log
+   * (the reconstructability RFC).
+   */
+  'request/header': { header: EpochHeader; reason: RequestHeaderReason }
+  /**
+   * Amendment to the folded {@link EpochHeader}: at least one of a
+   * {@link SystemDelta}, a {@link ToolsDelta}, or a whole replacement
+   * {@link LlmCallConfig} (four scalars — not worth diffing). Appended by the
+   * loop inside the step, before dispatch, when the header for this request
+   * differs from the fold of the log so far; the writer verifies
+   * `applyHeaderDelta(previous, delta)` reproduces the new header exactly and
+   * falls back to a `'fallback'` `request/header` snapshot when it cannot, so
+   * a logged delta ALWAYS round-trips. NOT a {@link SurfaceEventType}.
+   */
+  'request/header-delta': { system?: SystemDelta; tools?: ToolsDelta; config?: LlmCallConfig }
 }
 
+/** The appendable event-type keys of {@link SessionEventMap}, plugin-merged extensions included. */
 export type SessionEventType = keyof SessionEventMap
 
 /**

packages/core/session/tests/derived-cache.spec.ts

@@ -0,0 +1,112 @@
+/**
+ * Derived-message cache tests: the session projects each surface node exactly
+ * once (O(new nodes) per call), rebuilds on a surface rewrite (replace /
+ * invalidate — the replaceGeneration signal), returns a fresh array snapshot
+ * per call over shared frozen messages, and stays deep-equal to a from-scratch
+ * replay derivation at every step — the incremental==scratch property the
+ * reconstructability RFC's invariant enforces in dev at request time.
+ */
+
+import { describe, expect, it } from 'vitest'
+import { Session, SessionId } from '@deepseek-ai/dsh-session'
+
+function userText(session: Session, text: string): void {
+  session.append('user/message', { content: [{ type: 'text', text }], source: { kind: 'user' } }, { surfaceOp: 'append' })
+}
+
+/** From-scratch oracle: replay the log into a fresh session and derive. */
+function scratch(session: Session): unknown {
+  return new Session(SessionId(`${session.id}-scratch-${session.seq}`), [...session.events]).deriveMessages()
+}
+
+describe('derived-message cache', () => {
+  it('stays deep-equal to a from-scratch replay derivation as the log grows', () => {
+    const session = new Session(SessionId('cache-grow'))
+    session.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
+    userText(session, 'one')
+    expect(session.deriveMessages()).toEqual(scratch(session))
+    userText(session, 'two')
+    session.append('assistant/message', { turn: 1, step: 1, content: [{ type: 'text', text: 'reply' }] }, { surfaceOp: 'append' })
+    expect(session.deriveMessages()).toEqual(scratch(session))
+    // An empty-content assistant/message (usage host) projects to nothing.
+    session.append('assistant/message', { turn: 1, step: 2, content: [], usage: { inputTokens: 1, outputTokens: 0 } }, { surfaceOp: 'append' })
+    expect(session.deriveMessages()).toEqual(scratch(session))
+  })
+
+  it('rebuilds on a surface replace and still matches scratch', () => {
+    const session = new Session(SessionId('cache-replace'))
+    session.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
+    userText(session, 'one')
+    userText(session, 'two')
+    const beforeReplace = session.deriveMessages()
+    expect(beforeReplace).toHaveLength(2)
+
+    const nodes = session.surface.nodes
+    session.append('context/message', {
+      content: [{ type: 'text', text: 'summary' }], source: { kind: 'plugin', plugin: 'compact' },
+    }, { surfaceOp: { op: 'replace', start: nodes[0]!.seq, end: nodes[1]!.seq }, sourceEventSeqs: [nodes[0]!.seq, nodes[1]!.seq] })
+
+    expect(session.deriveMessages()).toHaveLength(1)
+    expect(session.deriveMessages()).toEqual(scratch(session))
+    // The array a caller took before the replace is untouched.
+    expect(beforeReplace).toHaveLength(2)
+  })
+
+  it('returns a fresh array per call: later appends never grow a held snapshot', () => {
+    const session = new Session(SessionId('cache-snapshot'))
+    session.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
+    userText(session, 'one')
+    const first = session.deriveMessages()
+    userText(session, 'two')
+    const second = session.deriveMessages()
+    expect(first).toHaveLength(1)
+    expect(second).toHaveLength(2)
+    // Shared projection objects: the same frozen message instance, once ever.
+    expect(second[0]).toBe(first[0])
+    expect(Object.isFrozen(first[0])).toBe(true)
+  })
+
+  it('rebuilds after surface.invalidate() (the generation covers wholesale rebuilds too)', () => {
+    const session = new Session(SessionId('cache-invalidate'))
+    session.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
+    userText(session, 'one')
+    const before = session.deriveMessages()
+    session.surface.invalidate()
+    const after = session.deriveMessages()
+    expect(after).toEqual(before)
+    // A rebuild re-projects: fresh objects, same values.
+    expect(after[0]).not.toBe(before[0])
+  })
+})
+
+describe('Session.deriveEventMessage — the per-event projection', () => {
+  it('projects one appended event exactly as the full derivation projects its node', () => {
+    const session = new Session(SessionId('per-event'))
+    session.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
+    const event = session.append('user/message', { content: [{ type: 'text', text: 'hi' }], source: { kind: 'user' } }, { surfaceOp: 'append' })
+    // The fold path (deriveMessages) and the per-event path share the
+    // projection, so an external reconstructor cannot disagree with the cache.
+    expect(session.deriveEventMessage(event)).toEqual(session.deriveMessages().at(-1))
+  })
+
+  it('clones content off the log: the projection never aliases the logged event', () => {
+    const session = new Session(SessionId('per-event-clone'))
+    session.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
+    const event = session.append('user/message', { content: [{ type: 'text', text: 'orig' }], source: { kind: 'user' } }, { surfaceOp: 'append' })
+    const message = session.deriveEventMessage(event)!
+    expect(message.content).not.toBe(event.data.content)
+    // deriveEventMessage returns an unfrozen clone (the cache freezes ITS
+    // copies); mutating it must not reach the log.
+    ;(message.content[0] as { text: string }).text = 'mutated'
+    expect(session.deriveMessages().at(-1)!.content).toEqual([{ type: 'text', text: 'orig' }])
+  })
+
+  it('projects null for events that produce no message (boundaries, empty assistant)', () => {
+    const session = new Session(SessionId('per-event-null'))
+    session.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
+    const boundary = session.append('step/start', { turn: 1, step: 1 })
+    expect(session.deriveEventMessage(boundary)).toBeNull()
+    const empty = session.append('assistant/message', { turn: 1, step: 1, content: [] }, { surfaceOp: 'append' })
+    expect(session.deriveEventMessage(empty)).toBeNull()
+  })
+})

packages/core/session/tests/fork.spec.ts

@@ -0,0 +1,240 @@
+import { describe, expect, it } from 'vitest'
+import { Context } from 'cordis'
+import { CallId } from '@deepseek-ai/dsh-llm'
+import SessionStore, { Session, SessionForkError, SessionId } from '@deepseek-ai/dsh-session'
+import type { SessionEvent, TurnEndReason } from '@deepseek-ai/dsh-session'
+
+async function setup(): Promise<{ ctx: Context; sessions: SessionStore }> {
+  const ctx = new Context()
+  await ctx.plugin(SessionStore)
+  return { ctx, sessions: ctx.sessions }
+}
+
+function appendClosedTurn(
+  session: Session,
+  turn: number,
+  text = `hello ${turn}`,
+  reason: TurnEndReason = { kind: 'completed' },
+): void {
+  session.append('turn/start', { turn, trigger: { kind: 'message', source: { kind: 'user' } } })
+  session.append('user/message', {
+    content: [{ type: 'text', text }],
+    source: { kind: 'user' },
+  }, { surfaceOp: 'append' })
+  session.append('turn/end', { turn, reason })
+}
+
+function appendOpenTurn(session: Session, turn: number): void {
+  session.append('turn/start', { turn, trigger: { kind: 'message', source: { kind: 'user' } } })
+  session.append('user/message', {
+    content: [{ type: 'text', text: `open ${turn}` }],
+    source: { kind: 'user' },
+  }, { surfaceOp: 'append' })
+}
+
+function firstUserMessage(events: readonly SessionEvent[]): SessionEvent<'user/message'> {
+  const event = events.find((e): e is SessionEvent<'user/message'> => e.type === 'user/message')
+  if (event === undefined) throw new Error('missing user/message')
+  return event
+}
+
+function lastSeq(session: Session): number {
+  const event = session.events.at(-1)
+  if (event === undefined) throw new Error('missing last event')
+  return event.seq
+}
+
+describe('SessionStore.fork', () => {
+  it('forks an empty live session as an empty child with lineage metadata', async () => {
+    const { ctx, sessions } = await setup()
+    const source = ctx.sessions.create(SessionId('empty-parent'), { meta: { cwd: '/workspace' } })
+
+    const child = sessions.fork(source, undefined, SessionId('empty-child'))
+
+    expect(child.events).toEqual([])
+    expect(child.header).toMatchObject({
+      id: SessionId('empty-child'),
+      cwd: '/workspace',
+      parentSession: SessionId('empty-parent'),
+      seedLength: 0,
+    })
+  })
+
+  it('forks the latest completed boundary by default and deep-clones seed events', async () => {
+    const { ctx, sessions } = await setup()
+    const source = ctx.sessions.create(SessionId('parent'), { meta: { cwd: '/workspace' } })
+    appendClosedTurn(source, 1, 'hello')
+
+    const child = sessions.fork(SessionId('parent'), undefined, SessionId('child'))
+
+    expect(child.events).toEqual(source.events)
+    expect(child.events).not.toBe(source.events)
+    expect(child.events[1]).not.toBe(source.events[1])
+    firstUserMessage(child.events).data.content[0] = { type: 'text', text: 'child mutation' }
+    expect(firstUserMessage(source.events).data.content).toEqual([{ type: 'text', text: 'hello' }])
+    expect(child.header).toMatchObject({
+      id: SessionId('child'),
+      cwd: '/workspace',
+      parentSession: SessionId('parent'),
+      seedLength: source.events.length,
+    })
+  })
+
+  it('forks from an earlier turn boundary even when the source currently has an open tail', async () => {
+    const { ctx, sessions } = await setup()
+    const source = ctx.sessions.create(SessionId('parent'), { meta: { cwd: '/workspace' } })
+    appendClosedTurn(source, 1, 'first')
+    const firstBoundary = lastSeq(source)
+    appendClosedTurn(source, 2, 'second')
+    appendOpenTurn(source, 3)
+
+    const child = sessions.fork(source, firstBoundary, SessionId('child-from-first'))
+
+    expect(child.events).toEqual(source.events.slice(0, firstBoundary + 1))
+    expect(child.header.seedLength).toBe(firstBoundary + 1)
+    expect(child.deriveMessages()).toEqual([{ role: 'user', content: [{ type: 'text', text: 'first' }] }])
+  })
+
+  it('accepts every turn/end reason as an explicit fork boundary', async () => {
+    const { ctx, sessions } = await setup()
+    const reasons: TurnEndReason[] = [
+      { kind: 'completed' },
+      { kind: 'aborted', reason: 'cancelled by user' },
+      { kind: 'error', step: 1, message: 'model failed', code: 'MODEL' },
+      { kind: 'disposed' },
+      { kind: 'max-tokens' },
+      { kind: 'interrupted' },
+    ]
+
+    for (const reason of reasons) {
+      const source = ctx.sessions.create(SessionId(`parent-${reason.kind}`))
+      appendClosedTurn(source, 1, reason.kind, reason)
+
+      const child = sessions.fork(source, lastSeq(source), SessionId(`child-${reason.kind}`))
+
+      expect(child.events.at(-1)?.type).toBe('turn/end')
+      expect(child.header.seedLength).toBe(source.events.length)
+    }
+  })
+
+  it('rejects invalid boundaries before creating a child', async () => {
+    const { ctx, sessions } = await setup()
+    const empty = ctx.sessions.create(SessionId('empty'))
+    expect(() => sessions.fork(empty, 0, SessionId('empty-child')))
+      .toThrow(new SessionForkError('fork boundary 0 does not exist in session "empty" (last seq: none)', 'INVALID_BOUNDARY'))
+    expect(ctx.sessions.get(SessionId('empty-child'))).toBeUndefined()
+
+    const source = ctx.sessions.create(SessionId('parent'))
+    appendClosedTurn(source, 1)
+    expect(() => sessions.fork(source, -1, SessionId('negative')))
+      .toThrow(/non-negative safe integer/)
+    expect(() => sessions.fork(source, 0.5, SessionId('fraction')))
+      .toThrow(/non-negative safe integer/)
+    expect(() => sessions.fork(source, Number.MAX_SAFE_INTEGER + 1, SessionId('unsafe')))
+      .toThrow(/non-negative safe integer/)
+    expect(() => sessions.fork(source, source.seq, SessionId('past-end')))
+      .toThrow(new SessionForkError(`fork boundary ${source.seq} does not exist in session "parent" (last seq: ${source.seq - 1})`, 'INVALID_BOUNDARY'))
+  })
+
+  it('rejects a corrupted live source whose array index no longer matches event seq', async () => {
+    const { ctx, sessions } = await setup()
+    const source = ctx.sessions.create(SessionId('corrupt-parent'))
+    appendClosedTurn(source, 1)
+    const mutableLog = (source as unknown as { log: SessionEvent[] }).log
+    mutableLog[2] = { ...mutableLog[2]!, seq: 99 }
+
+    expect(() => sessions.fork(source, 2, SessionId('corrupt-child')))
+      .toThrow(new SessionForkError('fork boundary 2 does not match a contiguous event seq in session "corrupt-parent"', 'INVALID_BOUNDARY'))
+    expect(ctx.sessions.get(SessionId('corrupt-child'))).toBeUndefined()
+  })
+
+  it('rejects an unknown live session id', async () => {
+    const { sessions } = await setup()
+
+    expect(() => sessions.fork(SessionId('missing')))
+      .toThrow(new SessionForkError('session "missing" not found', 'SESSION_NOT_FOUND'))
+  })
+
+  it('rejects a detached Session object that is not live in ctx.sessions', async () => {
+    const { sessions } = await setup()
+    const detached = new Session(SessionId('detached'))
+
+    expect(() => sessions.fork(detached))
+      .toThrow(new SessionForkError('session "detached" not found', 'SESSION_NOT_FOUND'))
+  })
+
+  it('rejects a stale Session object whose id is live on a different instance', async () => {
+    const { ctx, sessions } = await setup()
+    ctx.sessions.create(SessionId('same-id'))
+    const stale = new Session(SessionId('same-id'))
+
+    expect(() => sessions.fork(stale))
+      .toThrow(new SessionForkError('session "same-id" is not the live store instance', 'SESSION_NOT_LIVE'))
+  })
+
+  it('rejects selected slices whose boundary is inside an open turn', async () => {
+    const { ctx, sessions } = await setup()
+    const cases: [string, (session: Session) => number][] = [
+      ['turn/start', (session) => {
+        session.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
+        return lastSeq(session)
+      }],
+      ['step/start', (session) => {
+        session.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
+        session.append('step/start', { turn: 1, step: 1 })
+        return lastSeq(session)
+      }],
+      ['user/message', (session) => {
+        session.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
+        session.append('user/message', { content: [{ type: 'text', text: 'open' }], source: { kind: 'user' } }, { surfaceOp: 'append' })
+        return lastSeq(session)
+      }],
+      ['assistant/message', (session) => {
+        session.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
+        session.append('step/start', { turn: 1, step: 1 })
+        session.append('assistant/message', { turn: 1, step: 1, content: [{ type: 'text', text: 'partial' }] }, { surfaceOp: 'append' })
+        return lastSeq(session)
+      }],
+      ['tool/call', (session) => {
+        const callId = CallId('call-open')
+        session.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
+        session.append('step/start', { turn: 1, step: 1 })
+        session.append('assistant/message', {
+          turn: 1,
+          step: 1,
+          content: [{ type: 'tool-call', id: callId, name: 'bash', arguments: '{}' }],
+        }, { surfaceOp: 'append' })
+        session.append('tool/call', { turn: 1, step: 1, callId, name: 'bash', arguments: '{}' })
+        return lastSeq(session)
+      }],
+    ]
+
+    for (const [lastType, build] of cases) {
+      const source = ctx.sessions.create(SessionId(`open-${lastType}`))
+      const boundary = build(source)
+
+      expect(() => sessions.fork(source, boundary))
+        .toThrow(new SessionForkError(`fork boundary ${boundary} in session "open-${lastType}" must be turn/end, got ${lastType}`, 'OPEN_TURN'))
+    }
+  })
+
+  it('rejects a child session id that is already live with a typed fork error', async () => {
+    const { ctx, sessions } = await setup()
+    const source = ctx.sessions.create(SessionId('parent'))
+    appendClosedTurn(source, 1)
+    ctx.sessions.create(SessionId('child'))
+
+    expect(() => sessions.fork(source, undefined, SessionId('child')))
+      .toThrow(new SessionForkError('session "child" already exists', 'SESSION_ALREADY_EXISTS'))
+  })
+
+  it('rejects a duplicate child session id before validating the boundary', async () => {
+    const { ctx, sessions } = await setup()
+    const source = ctx.sessions.create(SessionId('open-parent'))
+    source.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
+    ctx.sessions.create(SessionId('child'))
+
+    expect(() => sessions.fork(source, undefined, SessionId('child')))
+      .toThrow(new SessionForkError('session "child" already exists', 'SESSION_ALREADY_EXISTS'))
+  })
+})

packages/core/session/tests/properties.spec.ts

@@ -109,15 +109,17 @@ describe('Session properties', () => {
     ))
   })
 
-  it('every derived message has a known role and decoupled content', () => {
+  it('every derived message has a known role and is frozen (append-only contract)', () => {
     fc.assert(fc.property(logArb, (events) => {
       const session = build(events)
       const messages = session.deriveMessages()
       const before = structuredClone(session.events)
       for (const m of messages) {
         expect(['user', 'assistant', 'system']).toContain(m.role)
-        // Mutating derived content must not touch the log (append-only).
-        m.content.push({ type: 'text', text: 'mutation' })
+        // Derived messages are frozen shared projections: mutation THROWS
+        // (strict mode) instead of relying on per-call clones for isolation.
+        expect(Object.isFrozen(m)).toBe(true)
+        expect(() => { m.content.push({ type: 'text', text: 'mutation' }) }).toThrow(TypeError)
       }
       expect(session.events).toEqual(before)
     }))

packages/core/session/tests/request-header.spec.ts

@@ -0,0 +1,140 @@
+/**
+ * Request-header utility tests: canonical form, the system line-diff
+ * (prefix/suffix trim), the name-keyed tools delta, config replacement, the
+ * round-trip contract (including the reorder case the encoding cannot
+ * express), and the log fold. These pin the reconstruction algebra: for every
+ * logged delta, apply(prev, delta) === next, and folding a log prefix yields
+ * the header its next request was built under.
+ */
+
+import { describe, expect, it } from 'vitest'
+import { Session, SessionId, applyHeaderDelta, canonicalHeader, diffHeader, foldRequestHeader } from '@deepseek-ai/dsh-session'
+import type { EpochHeader, SessionEvent } from '@deepseek-ai/dsh-session'
+import type { ToolSchema } from '@deepseek-ai/dsh-llm'
+
+const CONFIG = { model: 'm' }
+
+function tool(name: string, description = 'd'): ToolSchema {
+  return { name, description, parameters: { type: 'object' } }
+}
+
+/** Round-trip helper: diff must reproduce `next` from `prev` exactly. */
+function roundTrip(prev: EpochHeader, next: EpochHeader): ReturnType<typeof diffHeader> {
+  const delta = diffHeader(prev, next)
+  if (delta !== undefined) {
+    expect(applyHeaderDelta(prev, delta)).toEqual(canonicalHeader(next))
+  }
+  return delta
+}
+
+describe('canonicalHeader', () => {
+  it('normalizes empty system and empty tools to absent fields', () => {
+    expect(canonicalHeader({ config: CONFIG, system: '', tools: [] })).toEqual({ config: CONFIG })
+    const full = canonicalHeader({ config: CONFIG, system: 's', tools: [tool('a')] })
+    expect(full.system).toBe('s')
+    expect(full.tools).toHaveLength(1)
+  })
+})
+
+describe('diffHeader / applyHeaderDelta', () => {
+  it('returns undefined for equal headers', () => {
+    const header = canonicalHeader({ config: CONFIG, system: 'a\nb', tools: [tool('t')] })
+    expect(diffHeader(header, header)).toBeUndefined()
+  })
+
+  it('encodes a mid-prompt line change as a prefix/suffix trim', () => {
+    const prev = canonicalHeader({ config: CONFIG, system: 'keep1\nold\nkeep2\nkeep3' })
+    const next = canonicalHeader({ config: CONFIG, system: 'keep1\nnew A\nnew B\nkeep2\nkeep3' })
+    const delta = roundTrip(prev, next)
+    expect(delta?.system).toEqual({ keepStart: 1, keepEnd: 2, insert: ['new A', 'new B'] })
+    expect(delta?.tools).toBeUndefined()
+    expect(delta?.config).toBeUndefined()
+  })
+
+  it('degenerates to a full replacement when nothing is shared, and round-trips absence transitions', () => {
+    const none = canonicalHeader({ config: CONFIG })
+    const some = canonicalHeader({ config: CONFIG, system: 'x\ny' })
+    const gained = roundTrip(none, some)
+    expect(gained?.system).toEqual({ keepStart: 0, keepEnd: 0, insert: ['x', 'y'] })
+    const lost = roundTrip(some, none)
+    expect(lost?.system).toEqual({ keepStart: 0, keepEnd: 0, insert: [] })
+  })
+
+  it('does not double-count overlapping prefix and suffix (repeated lines)', () => {
+    const prev = canonicalHeader({ config: CONFIG, system: 'a\na' })
+    const next = canonicalHeader({ config: CONFIG, system: 'a\na\na' })
+    roundTrip(prev, next)
+  })
+
+  it('encodes tool addition, removal, and in-place schema change by name', () => {
+    const prev = canonicalHeader({ config: CONFIG, tools: [tool('keep'), tool('drop'), tool('edit', 'before')] })
+    const next = canonicalHeader({ config: CONFIG, tools: [tool('keep'), tool('edit', 'after'), tool('new')] })
+    const delta = roundTrip(prev, next)
+    expect(delta?.tools?.added.map(t => t.name)).toEqual(['new'])
+    expect(delta?.tools?.removed).toEqual(['drop'])
+    expect(delta?.tools?.changed.map(t => t.name)).toEqual(['edit'])
+  })
+
+  it('round-trips a tool set gained from a tool-less header and lost back to one', () => {
+    const none = canonicalHeader({ config: CONFIG })
+    const some = canonicalHeader({ config: CONFIG, tools: [tool('t')] })
+    const gained = roundTrip(none, some)
+    expect(gained?.tools?.added.map(t => t.name)).toEqual(['t'])
+    const lost = roundTrip(some, none)
+    expect(lost?.tools?.removed).toEqual(['t'])
+  })
+
+  it('cannot express a pure reordering — the writer detects it via the round-trip check', () => {
+    const prev = canonicalHeader({ config: CONFIG, tools: [tool('a'), tool('b')] })
+    const next = canonicalHeader({ config: CONFIG, tools: [tool('b'), tool('a')] })
+    const delta = diffHeader(prev, next)
+    // A delta IS produced (the lists differ)…
+    expect(delta).toBeDefined()
+    // …but applying it cannot reproduce the new order — exactly the case the
+    // writer's guard turns into a 'fallback' snapshot.
+    expect(applyHeaderDelta(prev, delta!)).not.toEqual(next)
+  })
+
+  it('replaces the config whole and leaves untouched parts alone', () => {
+    const prev = canonicalHeader({ config: { model: 'm' }, system: 's', tools: [tool('t')] })
+    const next = canonicalHeader({ config: { model: 'm2', temperature: 0.1 }, system: 's', tools: [tool('t')] })
+    const delta = roundTrip(prev, next)
+    expect(delta).toEqual({ config: { model: 'm2', temperature: 0.1 } })
+  })
+})
+
+describe('foldRequestHeader', () => {
+  function headerEvents(session: Session): readonly SessionEvent[] {
+    return session.events
+  }
+
+  it('returns undefined on a log with no header events', () => {
+    const session = new Session(SessionId('fold-none'))
+    session.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
+    expect(foldRequestHeader(headerEvents(session))).toBeUndefined()
+  })
+
+  it('folds snapshot then deltas into the header in force, skipping unrelated events', () => {
+    const session = new Session(SessionId('fold'))
+    session.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
+    const first = canonicalHeader({ config: { model: 'm' }, system: 'a\nb', tools: [tool('t')] })
+    session.append('request/header', { header: first, reason: 'initial' })
+    session.append('user/message', { content: [{ type: 'text', text: 'hi' }], source: { kind: 'user' } }, { surfaceOp: 'append' })
+
+    const second = canonicalHeader({ config: { model: 'm' }, system: 'a\nc', tools: [tool('t')] })
+    session.append('request/header-delta', diffHeader(first, second)!)
+    expect(foldRequestHeader(headerEvents(session))).toEqual(second)
+
+    // A later snapshot replaces the state wholesale (the 'resume'/'fallback' anchor).
+    const third = canonicalHeader({ config: { model: 'other' } })
+    session.append('request/header', { header: third, reason: 'resume' })
+    expect(foldRequestHeader(headerEvents(session))).toEqual(third)
+  })
+
+  it('throws on a delta before any snapshot (corrupt log)', () => {
+    const session = new Session(SessionId('fold-corrupt'))
+    session.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
+    session.append('request/header-delta', { config: { model: 'x' } })
+    expect(() => foldRequestHeader(headerEvents(session))).toThrow(/before any request\/header snapshot/)
+  })
+})

packages/core/session/tests/session.spec.ts

@@ -61,8 +61,10 @@ describe('Session', () => {
 
   it('replays identically from a seeded event log', () => {
     const original = new Session(SessionId('s3'))
+    original.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
     original.append('user/message', { content: [{ type: 'text', text: 'q' }], source: { kind: 'user' } }, { surfaceOp: 'append' })
     original.append('assistant/message', { turn: 1, step: 1, content: [{ type: 'text', text: 'a' }] }, { surfaceOp: 'append' })
+    original.append('turn/end', { turn: 1, reason: { kind: 'completed' } })
 
     const replayed = new Session(SessionId('s3-replay'), [...original.events])
     expect(replayed.deriveMessages()).toEqual(original.deriveMessages())
@@ -78,19 +80,25 @@ describe('Session', () => {
     }, { surfaceOp: 'append' })
     const before = structuredClone(session.events)
 
-    // A request middleware / adapter mutates the messages it was handed.
+    // A misbehaving consumer tries to mutate the messages it was handed.
+    // Derived messages are frozen shared projections (cloned once off the
+    // log, then deep-frozen): every mutation attempt THROWS in strict mode —
+    // isolation by unrepresentability, not by per-call cloning.
     const messages = session.deriveMessages()
     const userBlock = messages[0]!.content[0]!
-    if (userBlock.type === 'text') userBlock.text = 'HACKED'
+    expect(() => { if (userBlock.type === 'text') userBlock.text = 'HACKED' }).toThrow(TypeError)
     const toolBlock = messages[1]!.content[0]!
-    if (toolBlock.type === 'tool-result') {
-      toolBlock.content.push({ type: 'text', text: 'injected' })
-    }
-    messages[0]!.content.push({ type: 'text', text: 'extra' })
+    expect(() => {
+      if (toolBlock.type === 'tool-result') toolBlock.content.push({ type: 'text', text: 'injected' })
+    }).toThrow(TypeError)
+    expect(() => { messages[0]!.content.push({ type: 'text', text: 'extra' }) }).toThrow(TypeError)
+    // The returned ARRAY is the caller's own snapshot, though — reordering it
+    // is the caller's business and never reaches the cache or the log.
+    messages.reverse()
 
     // The log is unchanged: deep-equal to the snapshot taken before mutation.
     expect(session.events).toEqual(before)
-    // And a fresh derivation still reflects the original content.
+    // And a fresh derivation still reflects the original content and order.
     expect(session.deriveMessages()[0]!.content).toEqual([{ type: 'text', text: 'original' }])
   })
 
@@ -417,7 +425,9 @@ describe('todo/write event', () => {
 
   it('round-trips through a seeded replay identically (durable, no surfaceOp needed)', () => {
     const original = new Session(SessionId('t4'))
+    original.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
     original.append('todo/write', { todos: [{ content: 'only', status: 'completed' }] })
+    original.append('turn/end', { turn: 1, reason: { kind: 'completed' } })
     // Seeding a non-surface event with no surfaceOp must not throw.
     const replayed = new Session(SessionId('t4-replay'), [...original.events])
     expect(replayed.events.findLast(e => e.type === 'todo/write')!.data.todos)

packages/core/session/tests/surface.spec.ts

@@ -334,3 +334,26 @@ describe('surface type guards', () => {
     expect(isSurfaceEvent(markerless)).toBe(false)
   })
 })
+
+describe('SurfaceManager.replaceGeneration', () => {
+  it('folds the pending log delta on access and counts replaces and invalidations', () => {
+    const s = new Session(SessionId('gen'))
+    s.append('turn/start', { turn: 1, trigger: { kind: 'message', source: { kind: 'user' } } })
+    s.append('user/message', { content: [{ type: 'text', text: 'one' }], source: { kind: 'user' } }, { surfaceOp: 'append' })
+    s.append('user/message', { content: [{ type: 'text', text: 'two' }], source: { kind: 'user' } }, { surfaceOp: 'append' })
+    // Read the generation FIRST — before nodes — so the getter itself folds
+    // the pending delta rather than piggybacking on a nodes read.
+    expect(s.surface.replaceGeneration).toBe(0)
+
+    const nodes = s.surface.nodes
+    s.append('context/message', {
+      content: [{ type: 'text', text: 'summary' }], source: { kind: 'plugin', plugin: 'compact' },
+    }, { surfaceOp: { op: 'replace', start: nodes[0]!.seq, end: nodes[1]!.seq }, sourceEventSeqs: [nodes[0]!.seq, nodes[1]!.seq] })
+    expect(s.surface.replaceGeneration).toBe(1)
+
+    // invalidate() is a rewrite too: the generation moves forward (and the
+    // refold re-counts the replace), never backwards.
+    s.surface.invalidate()
+    expect(s.surface.replaceGeneration).toBeGreaterThan(1)
+  })
+})

packages/core/system-prompt/README.md

@@ -7,15 +7,16 @@ System prompt assembly registry. Plugins contribute ordered text sections, tool-
 | Key | Default | Meaning |
 |---|---|---|
 | `persona` | `''` | The deployment persona: the ONE deployment-authored prompt fragment, rendered as the order-0 `deployment:persona` section and shared by every agent in the context (subagents included). A template — complete `{{…}}` groups are interpreted strictly against the registered variables (the shipped loop registers `{{model}}`/`{{cwd}}`), with no escape syntax for literal braces yet. Empty ⇒ the section is dropped at render. |
+| `toolOrder` | — | Explicit model-facing tool order, as a list of `ToolSchema.name`s with one `'<unlisted-tools>'` rest entry (`TOOL_ORDER_REST`): listed tools take their listed position, unlisted tools land at the rest entry in lexicographic name order. Absent ⇒ plain lexicographic name order. Applied to the collected tools BEFORE the `system-prompt/assemble` waterfall — like the sections' `order` sort, it canonicalizes what the registry contributed (registration order is a plugin-load artifact), and a waterfall listener that mutates the list owns the determinism of what it emits. Misconfiguration fails loud: a list without exactly one rest entry, or with duplicates, throws at load; a listed name with no registered tool rejects every `assemble()`; a tool provider returning the reserved rest-entry name also rejects. Under the shipped loop the turn fails before any model request. Why a central list and not per-plugin weights: [Explicit model-facing tool order](../../../docs/rfc/implemented/feature/2026-07-06-explicit-tool-order.md). |
 
 ## Service: `SystemPrompt` (ctx key: `systemPrompt`)
 
 ### Public API
 
 - `ctx.systemPrompt.section(section: PromptSection): () => void` Contribute a section. Duplicate names throw. Disposed with the calling fiber.
-- `ctx.systemPrompt.tools(provider: () => ToolSchema[]): () => void` Contribute tool schemas (evaluated at each assembly). Disposed with the calling fiber.
+- `ctx.systemPrompt.tools(provider: () => ToolSchema[]): () => void` Contribute tool schemas (evaluated at each assembly). A provider must not return a schema named `TOOL_ORDER_REST`; that name is reserved for `toolOrder`'s rest entry. Disposed with the calling fiber.
 - `ctx.systemPrompt.variable(name: string, provider: (context) => string | undefined): () => void` Contribute a prompt variable, referenced from section text as `{{name}}`. Duplicate or unreferenceable names throw; `undefined` means "no value for this assembly". Disposed with the calling fiber.
-- `ctx.systemPrompt.assemble(context?: AssembleContext): Promise<PromptAssembly>` Assemble the prompt for one caller. Runs through the `system-prompt/assemble` waterfall.
+- `ctx.systemPrompt.assemble(context?: AssembleContext): Promise<PromptAssembly>` Assemble the prompt for one caller. Runs through the `system-prompt/assemble` waterfall. Rejects when a configured `toolOrder` names a tool no provider contributed, or when a provider returns the reserved rest-entry name.
 
 ### Events
 

packages/core/system-prompt/src/index.ts

@@ -88,7 +88,8 @@ export interface AssembledSection {
  *
  * Tool schemas are part of the assembly by design: "what the model is told it
  * can do" is one coherent thing managed here, even though adapters transmit
- * `tools` as a separate wire field rather than prompt text.
+ * `tools` as a separate wire field rather than prompt text. They arrive in
+ * the canonical model-facing order (see {@link Config.toolOrder}).
  *
  * `variables` carries every registered prompt variable resolved against this
  * assembly's context — key present means registered, `undefined` value means
@@ -110,6 +111,71 @@ const VARIABLE_NAME = /^[a-z][a-z0-9_]*$/
 /** A complete `{{...}}` reference group at the scan position (validated after). */
 const GROUP_AT = /^\{\{([^{}]*)\}\}/
 
+/**
+ * The rest entry for {@link Config.toolOrder}: the position where registered
+ * tools not named in the list are inserted (in lexicographic name order).
+ * Reserved: collected tool schemas using this name are rejected before
+ * ordering, so the marker can never collide with a real model-facing tool.
+ */
+export const TOOL_ORDER_REST = '<unlisted-tools>'
+
+/**
+ * Validate a configured tool-order list's shape at service construction:
+ * the {@link TOOL_ORDER_REST} rest entry exactly once, no duplicate names.
+ * Returns the list (or undefined when unconfigured); throws otherwise,
+ * failing the service at load — a bad order config must never reach an
+ * assembly. Whether every listed name matches a registered tool is checked
+ * at each assembly instead ({@link orderTools}): tool plugins register after
+ * this service constructs, so the tool set does not exist yet here.
+ */
+function validateToolOrder(toolOrder: string[] | undefined): string[] | undefined {
+  if (toolOrder === undefined) return undefined
+  const seen = new Set<string>()
+  for (const name of toolOrder) {
+    if (seen.has(name)) throw new Error(`toolOrder lists "${name}" more than once`)
+    seen.add(name)
+  }
+  if (!seen.has(TOOL_ORDER_REST)) {
+    throw new Error(`toolOrder must contain the "${TOOL_ORDER_REST}" rest entry (where unlisted tools are inserted)`)
+  }
+  return toolOrder
+}
+
+/**
+ * Order collected tool schemas by the validated policy: with no configured
+ * list, plain lexicographic name order; with one, listed names take their
+ * listed position and every unlisted tool lands at the
+ * {@link TOOL_ORDER_REST} rest entry in lexicographic name order. A listed
+ * name with no collected tool throws — misconfiguration fails loud, and this
+ * is the earliest moment the registered tool set exists to check against
+ * (tool plugins register after the service constructs, so load time is too
+ * early): the assembly rejects, failing the caller's turn before any model
+ * request. Never drops a tool, and both sorts are stable, so tools sharing a
+ * name keep their collection order.
+ */
+function orderTools(tools: ToolSchema[], toolOrder: string[] | undefined): ToolSchema[] {
+  const reserved = tools.find(tool => tool.name === TOOL_ORDER_REST)
+  if (reserved !== undefined) {
+    throw new Error(`tool provider returned reserved tool name "${TOOL_ORDER_REST}" (reserved for toolOrder's rest entry)`)
+  }
+  if (toolOrder === undefined) return tools.sort(compareToolNames)
+  const registered = new Set(tools.map(tool => tool.name))
+  const unknown = toolOrder.filter(name => name !== TOOL_ORDER_REST && !registered.has(name))
+  if (unknown.length > 0) {
+    throw new Error(`toolOrder lists unregistered tool${unknown.length > 1 ? 's' : ''} ${unknown.map(name => `"${name}"`).join(', ')}; registered tools: ${[...registered].sort().join(', ') || '(none)'}`)
+  }
+  const listed = new Set(toolOrder)
+  const rest = tools.filter(tool => !listed.has(tool.name)).sort(compareToolNames)
+  return toolOrder.flatMap(name =>
+    name === TOOL_ORDER_REST ? rest : tools.filter(tool => tool.name === name))
+}
+
+/** Lexicographic (code-unit) name comparison — locale-independent, so the order is identical on every machine. */
+function compareToolNames(a: ToolSchema, b: ToolSchema): number {
+  return a.name < b.name ? -1 : a.name > b.name ? 1 : 0
+}
+
+/** Plugin config: the deployment-authored fragment of the system prompt (see {@link Config.persona} for its contract). */
 export interface Config {
   /**
    * The deployment's persona — the ONE deployment-authored fragment of the
@@ -124,6 +190,29 @@ export interface Config {
    * deployment opens with the harness identity alone.
    */
   persona?: string
+  /**
+   * Explicit model-facing tool order, as a list of `ToolSchema.name`s: listed
+   * tools take their listed position, and tools absent from the list are
+   * inserted at the {@link TOOL_ORDER_REST} (`'<unlisted-tools>'`) entry in
+   * lexicographic name order. A configured list must contain the rest entry
+   * exactly once, no duplicate names, and no name without a registered tool —
+   * a misconfigured order blocks work instead of silently reaching a model
+   * request: shape violations throw at load, and an unregistered name rejects
+   * every assembly. `TOOL_ORDER_REST` is reserved for the list marker and may
+   * not be a collected tool name; such a provider output also rejects the
+   * assembly. The single assembly-time validation rejects either failure
+   * before any model request — the earliest moment the registered tool set
+   * exists to check against, since tool plugins register after this service
+   * constructs. When omitted, tools are ordered lexicographically by name.
+   * Applied to the tools
+   * {@link SystemPrompt.assemble} collects, BEFORE the
+   * `system-prompt/assemble` waterfall — like the sections' `order` sort, it
+   * canonicalizes what the registry contributed (registration order is a
+   * plugin-load artifact); a waterfall listener that mutates the tool list
+   * owns the determinism of what it emits. Rationale (and why not per-plugin
+   * weights): docs/rfc/implemented/feature/2026-07-06-explicit-tool-order.md.
+   */
+  toolOrder?: string[]
 }
 
 /**
@@ -138,6 +227,10 @@ export interface Config {
  * while a `}}` still follows (e.g. `{{{model}}}`, `{{a{b}}`) all throw. A
  * lone `{{` with no `}}` anywhere after it is ordinary prose and passes
  * through verbatim. Substituted values are never re-scanned.
+ * @param assembly - the assembly to render (typically the awaited result of
+ *   {@link SystemPrompt.assemble}); only `sections` and `variables` are read.
+ * @returns the full system prompt text; `''` when every section renders empty
+ *   (the caller then sends no system prompt at all).
  */
 export function renderPrompt(assembly: PromptAssembly): string {
   return assembly.sections
@@ -198,14 +291,23 @@ function interpolate(section: AssembledSection, variables: Record<string, string
 export class SystemPrompt extends Service {
   static Config: z<Config> = z.object({
     persona: z.string().default(''),
+    // A schemastery array defaults to [] when omitted, but an omitted
+    // toolOrder must stay absent ("lexicographic order"), not become an
+    // explicitly-configured empty list (which is invalid — it lacks the
+    // rest entry). Forcing the default to undefined keeps the key out of the
+    // validated config; the cast is needed because .default() expects the
+    // array type.
+    toolOrder: z.array(z.string()).default(undefined as unknown as string[]),
   })
 
   private sections: PromptSection[] = []
   private toolProviders: (() => ToolSchema[])[] = []
   private variableProviders = new Map<string, (context: AssembleContext) => string | undefined>()
+  private readonly toolOrder: string[] | undefined
 
   constructor(ctx: Context, public config: Config) {
     super(ctx, 'systemPrompt')
+    this.toolOrder = validateToolOrder(config.toolOrder)
     // The harness-owned openers. They live HERE (not on the loop plugin) so a
     // deployment that swaps in a different loop keeps them: the identity is a
     // harness fact stated ahead of everything, and the persona is the
@@ -261,7 +363,10 @@ export class SystemPrompt extends Service {
   /**
    * Contribute a tool-schema provider that is evaluated at each assembly
    * call (so it can reflect the live registry state). The provider is
-   * removed when the calling fiber is disposed. Emits `system-prompt/change`.
+   * removed when the calling fiber is disposed. A provider must not return a
+   * schema named {@link TOOL_ORDER_REST}; that name is reserved for
+   * {@link Config.toolOrder}'s rest entry and rejects the assembly. Emits
+   * `system-prompt/change`.
    * @param provider - evaluated at every {@link assemble} for fresh schemas.
    * @returns the disposer that removes the provider.
    */
@@ -318,19 +423,28 @@ export class SystemPrompt extends Service {
 
   /**
    * Assemble the current prompt for one caller: section texts are resolved
-   * against `context` and sorted by order, tools collected from all
-   * providers, and every registered variable resolved against `context` into
-   * `assembly.variables`. Tool schemas are deep-cloned because adapters and
-   * request waterfalls may mutate schema objects. Runs through the
-   * `system-prompt/assemble` waterfall, giving listeners the opportunity to
-   * mutate or replace the assembly before it reaches the model. Await the
-   * result before reading the assembly values — waterfall listeners may be
-   * async. Interpolation happens later, in {@link renderPrompt}.
+   * against `context` and sorted by order, tools collected from all providers
+   * and put in the canonical model-facing order ({@link Config.toolOrder}, or
+   * lexicographic name order when unconfigured — provider registration order
+   * is a plugin-load artifact and never reaches the assembly; a configured
+   * order naming a tool no provider contributed rejects the assembly), and every
+   * registered variable resolved against `context` into `assembly.variables`.
+   * Tool schemas are deep-cloned because adapters and request waterfalls may
+   * mutate schema objects. Runs through the `system-prompt/assemble`
+   * waterfall, giving listeners the opportunity to mutate or replace the
+   * assembly before it reaches the model — like the sections' `order` sort,
+   * tool canonicalization happens on the initial assembly, and a listener
+   * owns the determinism of whatever it emits. Await the result before
+   * reading the assembly values — waterfall listeners may be async.
+   * Interpolation happens later, in {@link renderPrompt}.
    * @param context - what this assembly is for (defaults to an empty context;
    *   see {@link AssembleContext}).
    * @returns the assembly after the waterfall has run.
    */
-  assemble(context: AssembleContext = {}): Promise<PromptAssembly> {
+  // async so the misconfigured-toolOrder throw in orderTools surfaces as a
+  // rejection: a Promise-returning method must not throw synchronously
+  // (`assemble().catch(...)` would miss it).
+  async assemble(context: AssembleContext = {}): Promise<PromptAssembly> {
     const variables: Record<string, string | undefined> = {}
     for (const [name, provider] of this.variableProviders) {
       variables[name] = provider(context)
@@ -343,8 +457,10 @@ export class SystemPrompt extends Service {
           text: typeof section.text === 'function' ? section.text(context) : section.text,
         }))
         .sort((a, b) => a.order - b.order),
-      tools: this.toolProviders.flatMap(provider =>
-        provider().map(tool => ({ ...tool, parameters: structuredClone(tool.parameters) }))),
+      tools: orderTools(
+        this.toolProviders.flatMap(provider =>
+          provider().map(tool => ({ ...tool, parameters: structuredClone(tool.parameters) }))),
+        this.toolOrder),
       variables,
     }
     return this.ctx.waterfall(this, 'system-prompt/assemble', assembly, context, () => Promise.resolve(assembly))

packages/core/system-prompt/tests/tool-order.spec.ts

@@ -0,0 +1,115 @@
+import { describe, expect, it } from 'vitest'
+import { Context } from 'cordis'
+import SystemPrompt, { PromptAssembly, TOOL_ORDER_REST } from '@deepseek-ai/dsh-system-prompt'
+import type { ToolSchema } from '@deepseek-ai/dsh-llm'
+
+function tool(name: string, description = name): ToolSchema {
+  return { name, description, parameters: { type: 'object', properties: {} } }
+}
+
+async function mount(config: { persona?: string; toolOrder?: string[] } = {}): Promise<Context> {
+  const ctx = new Context()
+  await ctx.plugin(SystemPrompt, config)
+  return ctx
+}
+
+function names(assembly: PromptAssembly): string[] {
+  return assembly.tools.map(t => t.name)
+}
+
+describe('SystemPrompt tool order', () => {
+  // The ONE place the public constant's value is pinned; everything else
+  // (tests and deployment configs alike) references TOOL_ORDER_REST.
+  it('exports the rest entry as "<unlisted-tools>"', () => {
+    expect(TOOL_ORDER_REST).toBe('<unlisted-tools>')
+  })
+
+  it('assembles tools in lexicographic name order when no toolOrder is configured', async () => {
+    const ctx = await mount()
+    ctx.systemPrompt.tools(() => [tool('charlie'), tool('alpha')])
+    ctx.systemPrompt.tools(() => [tool('bravo')])
+    expect(names(await ctx.systemPrompt.assemble())).toEqual(['alpha', 'bravo', 'charlie'])
+  })
+
+  it('assembles the same order regardless of provider registration order', async () => {
+    const forward = await mount()
+    forward.systemPrompt.tools(() => [tool('alpha')])
+    forward.systemPrompt.tools(() => [tool('zulu')])
+    const backward = await mount()
+    backward.systemPrompt.tools(() => [tool('zulu')])
+    backward.systemPrompt.tools(() => [tool('alpha')])
+    expect(names(await forward.systemPrompt.assemble())).toEqual(['alpha', 'zulu'])
+    expect(names(await backward.systemPrompt.assemble())).toEqual(['alpha', 'zulu'])
+  })
+
+  it('applies a configured toolOrder: listed positions, rest at the rest entry lexicographically', async () => {
+    const ctx = await mount({ toolOrder: ['todo_write', TOOL_ORDER_REST, 'bash'] })
+    ctx.systemPrompt.tools(() => [tool('bash'), tool('echo_b'), tool('todo_write'), tool('echo_a')])
+    expect(names(await ctx.systemPrompt.assemble())).toEqual(['todo_write', 'echo_a', 'echo_b', 'bash'])
+  })
+
+  it('rejects the assembly when toolOrder names a tool that is not registered (misconfiguration blocks work)', async () => {
+    const ctx = await mount({ toolOrder: ['todo_write', 'ghost', TOOL_ORDER_REST, 'wraith'] })
+    ctx.systemPrompt.tools(() => [tool('bash'), tool('todo_write')])
+    await expect(ctx.systemPrompt.assemble()).rejects.toThrow(
+      'toolOrder lists unregistered tools "ghost", "wraith"; registered tools: bash, todo_write')
+  })
+
+  it('names the single unregistered tool when no tools are registered at all', async () => {
+    const ctx = await mount({ toolOrder: ['ghost', TOOL_ORDER_REST] })
+    await expect(ctx.systemPrompt.assemble()).rejects.toThrow(
+      'toolOrder lists unregistered tool "ghost"; registered tools: (none)')
+  })
+
+  it.each([
+    ['without an explicit toolOrder', undefined],
+    ['with only the rest entry configured', [TOOL_ORDER_REST]],
+  ])('rejects a provider tool named like the reserved rest entry %s', async (_case, toolOrder) => {
+    const ctx = await mount(toolOrder === undefined ? {} : { toolOrder })
+    ctx.systemPrompt.tools(() => [tool(TOOL_ORDER_REST)])
+    await expect(ctx.systemPrompt.assemble()).rejects.toThrow(
+      `tool provider returned reserved tool name "${TOOL_ORDER_REST}"`)
+  })
+
+  it('keeps collection order between tools that share a name (stable sort)', async () => {
+    const ctx = await mount()
+    ctx.systemPrompt.tools(() => [tool('dup', 'first'), tool('anchor'), tool('dup', 'second')])
+    const assembly = await ctx.systemPrompt.assemble()
+    expect(assembly.tools.map(t => t.description)).toEqual(['anchor', 'first', 'second'])
+  })
+
+  it('canonicalizes BEFORE the assemble waterfall: listeners see the ordered list and own their own edits', async () => {
+    const ctx = await mount()
+    ctx.systemPrompt.tools(() => [tool('zulu'), tool('alpha')])
+    let seen: string[] | undefined
+    ctx.on('system-prompt/assemble', function (assembly, _context, next) {
+      seen = assembly.tools.map(t => t.name)
+      // A listener-appended tool is NOT re-sorted — same contract as sections:
+      // canonicalization applies to what the registry contributed, and a
+      // listener owns the determinism of what it emits.
+      assembly.tools.push(tool('aardvark'))
+      return next()
+    })
+    const assembly = await ctx.systemPrompt.assemble()
+    expect(seen).toEqual(['alpha', 'zulu'])
+    expect(names(assembly)).toEqual(['alpha', 'zulu', 'aardvark'])
+  })
+
+  it.each([
+    ['an empty list', []],
+    ['a list without the rest entry', ['bash', 'todo_write']],
+  ])('rejects %s at load (the rest entry is required)', async (_case, toolOrder) => {
+    await expect(new Context().plugin(SystemPrompt, { toolOrder })).rejects.toThrow(`must contain the "${TOOL_ORDER_REST}" rest entry`)
+  })
+
+  it.each([
+    ['a duplicate tool name', ['bash', 'bash', TOOL_ORDER_REST]],
+    ['a duplicate rest entry', [TOOL_ORDER_REST, 'bash', TOOL_ORDER_REST]],
+  ])('rejects %s at load', async (_case, toolOrder) => {
+    await expect(new Context().plugin(SystemPrompt, { toolOrder })).rejects.toThrow('more than once')
+  })
+
+  it('throws from direct construction too', () => {
+    expect(() => new SystemPrompt(new Context(), { toolOrder: ['bash'] })).toThrow('rest entry')
+  })
+})

packages/core/tools/README.md

@@ -1,6 +1,6 @@
 # dsh-tools
 
-Tool registry and execution pipeline. Tool plugins register their schemas and executors; the agent loop executes each call through `tools/pre-execute` (the allow/deny gate) → core dispatch → `tools/post-execute` (inspect/replace the result, attach context).
+Tool registry and execution pipeline. Tool plugins register their schemas and executors; the agent loop executes each call through `tools/pre-execute` (the allow/deny gate) → `tools/execute` (an around-dispatch wrapper for timeout/retry/metrics plugins) → `tools/post-execute` (inspect/replace the result, attach context).
 
 ## Service: `ToolRegistry` (ctx key: `tools`)
 
@@ -8,8 +8,8 @@ Tool registry and execution pipeline. Tool plugins register their schemas and ex
 
 - `ctx.tools.register(definition: ToolDefinition): () => void` Register a tool. Disposed with the calling fiber.
 - `ctx.tools.get(name: string): ToolDefinition | undefined`
-- `ctx.tools.schemas(): ToolSchema[]` Schemas of all registered tools (without the `execute` functions). The shipped tools' schemas are catalogued in [docs/tool-catalog/tools.md](../../../docs/tool-catalog/tools.md), generated by booting each tool plugin and harvesting this method (see [the tool-schema-catalog RFC](../../../docs/rfc/implemented/process/2026-07-02-tool-schema-catalog.md)).
-- `ctx.tools.execute(exec: ToolExecution): Promise<ToolExecutionResult>` Execute one tool call through the `tools/pre-execute` → dispatch → `tools/post-execute` pipeline.
+- `ctx.tools.schemas(): ToolSchema[]` Schemas of all registered tools (without the `execute` functions). The shipped tools' schemas are catalogued in [docs/tool-catalog.md](../../../docs/tool-catalog.md), generated by booting each tool plugin and harvesting this method (see [the tool-schema-catalog RFC](../../../docs/rfc/implemented/process/2026-07-02-tool-schema-catalog.md)).
+- `ctx.tools.execute(exec: ToolExecution): Promise<ToolExecutionResult>` Execute one tool call through the `tools/pre-execute` → `tools/execute` → `tools/post-execute` pipeline.
 
 ### Injected services
 
@@ -20,12 +20,13 @@ Tool registry and execution pipeline. Tool plugins register their schemas and ex
 | Event | Mode | Purpose |
 |---|---|---|
 | `tools/pre-execute` | waterfall | Allow/deny gate BEFORE a tool runs (sandbox, permission, hooks); returns `PreToolDecision` |
+| `tools/execute` | waterfall | Around-dispatch wrapper (timeout, retry, metrics): `(exec, next)` → the dispatched `ToolExecutionResult`; `next()` is dispatch-with-normalization |
 | `tools/post-execute` | waterfall | Inspect/replace the result AFTER a tool runs, attach context; returns `PostToolDecision` |
 | `tools/change` | emit | A tool was registered or unregistered |
 
 ### Key types
 
-- `ToolDefinition` — `ToolSchema` + `execute(args, exec): Promise<ContentBlock[] | { content: ContentBlock[]; meta? }>` (the bare array is the model-facing content; the object form additionally attaches an opaque, JSON-serializable `meta` presentation payload persisted on the `tool/result` event and handed back to `presentResult`), plus optional `presentCall(args)` / `presentResult(args, result)` for tool-owned UI presentation (see below).
+- `ToolDefinition` — `ToolSchema` + `execute(args, exec): Promise<ContentBlock[] | { content: ContentBlock[]; meta? }>` (the bare array is the model-facing content; the object form additionally attaches an opaque, JSON-serializable `meta` presentation payload persisted on the `tool/result` event and handed back to `presentResult`), plus optional `presentCall(args)` / `presentResult(args, result)` for tool-owned UI presentation (see below). It also carries an optional cooperative timeout budget `timeoutMs?: number` (ms) enforced by `@deepseek-ai/dsh-timeout-policy`, never sent to the model.
 - `ToolExecution` — one pending tool call: `{ callId, name, arguments, agent?, signal? }`.
 - `ToolExecutionResult` — outcome: `{ callId, content, isError, error?, additionalContext?, meta? }`. On failure with a `HarnessError`, `error: { name, code }` carries the structured failure class alongside the model-facing text (the loop forwards it onto the `tool/result` session event for retry/sandbox plugins and replay). `additionalContext` (a `HookContext`) ferries any `tools/post-execute` context up to the loop, which buffers it and appends it as a `context/message` after all `tool/result`s in the step. `meta` is the tool's opaque presentation payload from a successful `execute` (the object return form); the loop forwards it onto the `tool/result` session event for result-card rendering.
 - `PreToolDecision` — `{kind:'allow'}` | `{kind:'deny', reason}` | `{kind:'ask', reason?}`. Input rewrite (changing `arguments`) is deliberately NOT offered (it would desync the pre-execution audit/history/UI from what ran — its own proposed RFC); `ask` degrades to `deny` until the permission system lands.
@@ -35,7 +36,7 @@ Tool registry and execution pipeline. Tool plugins register their schemas and ex
 ### Extension points
 
 - Tool plugins call `ctx.tools.register()` — schemas flow into the assembly automatically.
-- `tools/pre-execute` is the allow/deny gate (sandbox, permission, hooks): listeners receive `(exec, next)` and call `next()` to delegate to the default (allow) or return a `PreToolDecision` to short-circuit; a `deny`/`ask` skips dispatch and yields an `isError` result. `tools/post-execute` is the inspect/transform seam: `(exec, result, next)` → a `PostToolDecision` that can replace content, block with feedback, or attach `additionalContext`. Core dispatch sits between them as plain code; the tool body keeps its own try/catch so a thrown tool still reaches `post-execute` as an `isError`. Both follow the typed-Decision idiom shared with the `agent/*` interception seams (see [`dsh-agent`](../agent/README.md)).
+- `tools/pre-execute` is the allow/deny gate (sandbox, permission, hooks): listeners receive `(exec, next)` and call `next()` to delegate to the default (allow) or return a `PreToolDecision` to short-circuit; a `deny`/`ask` skips dispatch and yields an `isError` result. `tools/execute` is the around-dispatch seam (timeout, retry, metrics): listeners receive `(exec, next)` and call `next()` to delegate to core dispatch (returning its `ToolExecutionResult`, optionally wrapped), or return a replacement result to short-circuit dispatch; the base `next()` IS dispatch-with-normalization, so `await next()` already yields an `isError` result for a thrown/unknown tool (never a raw throw). A wrapper mutates `exec` in place before `next()` — e.g. replacing `exec.signal` with a per-call deadline — because cordis `next()` ignores passed arguments. `tools/post-execute` is the inspect/transform seam: `(exec, result, next)` → a `PostToolDecision` that can replace content, block with feedback, or attach `additionalContext`. Core dispatch is the base of the `tools/execute` waterfall; the tool body keeps its own try/catch so a thrown tool still reaches `post-execute` as an `isError`. All follow the typed-Decision idiom shared with the `agent/*` interception seams (see [`dsh-agent`](../agent/README.md)); `@deepseek-ai/dsh-timeout-policy` is the reference `tools/execute` wrapper.
 - MCP servers: one plugin per server, discover tools, call `ctx.tools.register()` with the server's schemas.
 
 ### Typed tool parameter schemas
@@ -71,6 +72,14 @@ A `defineTool` tool also **validates the model-generated arguments against its `
 
 See `defineTool`, `validateArgs`, `ToolArgsError`, `SchemaSpec`, `InferArgs`, and `schemaSpecToJsonSchema` in the public API for details.
 
+`defineTool` also validates an optional `timeoutMs` at definition time when present: it must be a positive finite number, or the helper throws — the budget is attached to the produced `ToolDefinition` (for `@deepseek-ai/dsh-timeout-policy`) and never reaches the model.
+
+### Structured-output schema subset
+
+A separate vocabulary for callers that DEMAND a machine-readable value from an agent — the subagent seam's `SubagentStartRequest.outputSchema` (and, by extension, a workflow's `agent({ schema })`). Unlike `SchemaSpec` (the author-facing DSL for tool parameters), a `StructuredOutputSchema` is an object-rooted **raw JSON Schema subset** as data: it travels verbatim to the model as a forced tool's `parameters`, and the produced value is validated against it.
+
+The subset is deliberately narrow and REJECTS LOUD outside it — accepting a keyword the validator doesn't enforce would validate less than the schema promises (accepted-then-ignored). Supported: single-string `type` (`object`/`array`/`string`/`number`/`integer`/`boolean`/`null`; type arrays rejected), `properties`/`required`/`additionalProperties` (boolean; every `required` key must be declared), `items`, scalar-only `enum`/`const`; annotations (`description`/`title`/`default`/`examples`) are ignored but must still be JSON data. `assertSupportedOutputSchema(schema)` throws `OutputSchemaError` (`code: 'UNSUPPORTED_SCHEMA'`, listing every violation) for anything else; `validateStructuredValue(schema, value)` returns path-qualified violations (empty = valid, total — never throws).
+
 ### Tool-owned UI presentation
 
 A tool owns how ITS calls render in a UI (an editor's tool-call card, a CLI log line) — a UI plugin must NOT special-case tool names. A `ToolDefinition` may declare two optional, pure, display-only methods that return a **`card`-tagged render intent** (a discriminated union — a tool declares its card kind once and a UI bridge switches on `card`):

packages/core/tools/src/index.ts

@@ -1,9 +1,10 @@
 /**
  * Tool registry and execution pipeline. Plugins register tools; the registry
  * feeds schemas into the system prompt, and `execute()` dispatches each call
- * through `tools/pre-execute` (the allow/deny gate) → core dispatch →
- * `tools/post-execute` (inspect/replace the result, attach context) for
- * sandbox, permission, and hook plugins to gate or transform a call.
+ * through `tools/pre-execute` (the allow/deny gate) → `tools/execute` (an
+ * around-dispatch wrapper for timeout/retry/metrics plugins) → `tools/post-execute`
+ * (inspect/replace the result, attach context) for sandbox, permission, and hook
+ * plugins to gate or transform a call.
  *
  * @module @deepseek-ai/dsh-tools
  */
@@ -28,6 +29,16 @@ export {
   type JsonSchemaObject,
 } from './schema.ts'
 
+export {
+  assertSupportedOutputSchema,
+  validateStructuredValue,
+  OutputSchemaError,
+  type StructuredOutputSchema,
+  type StructuredSchemaNode,
+  type StructuredSchemaType,
+  type StructuredScalar,
+} from './json-schema.ts'
+
 // The render-intent vocabulary a tool declares via `presentCall`/`presentResult`
 // lives in its own UI-facing module; re-export it so `@deepseek-ai/dsh-tools`
 // stays the single public surface for consumers (producers + the ACP bridge).
@@ -64,17 +75,37 @@ declare module 'cordis' {
      * @mode waterfall
      */
     'tools/pre-execute'(this: ToolRegistry, exec: ToolExecution, next: () => Promise<PreToolDecision>): Promise<PreToolDecision>
+    /**
+     * Around-dispatch waterfall wrapping the registry's core tool dispatch,
+     * between the `tools/pre-execute` gate and the `tools/post-execute` seam. A
+     * listener receives `(exec, next)`: call `next()` to delegate to dispatch
+     * (returning its {@link ToolExecutionResult}, optionally wrapped), or return a
+     * replacement result without calling `next()` to short-circuit dispatch. The
+     * base `next()` IS the dispatch-with-normalization thunk — a thrown tool (or
+     * unknown tool) is already normalized to an `isError` result by the time a
+     * listener's `await next()` returns, so a wrapper never sees a raw throw from
+     * the tool body. This is the seam a timeout/retry/metrics plugin wraps: it can
+     * mutate `exec` (e.g. replace `exec.signal` with a per-call deadline) BEFORE
+     * `next()` and inspect the result AFTER. (Cordis `next()` ignores any passed
+     * arguments and re-invokes downstream with the shared payload, so a wrapper
+     * mutates `exec` in place rather than passing a new object to `next()`.)
+     * Multiple listeners compose by registration order — an outer one wraps the
+     * inner ones plus dispatch.
+     * @param exec - the allowed call about to dispatch (name, parsed arguments, caller agent, signal).
+     * @mode waterfall
+     */
+    'tools/execute'(this: ToolRegistry, exec: ToolExecution, next: () => Promise<ToolExecutionResult>): Promise<ToolExecutionResult>
     /**
      * Waterfall AFTER a tool runs — where hook plugins inspect the result and
      * accept it (optionally REPLACING the model-facing content, and/or attaching
      * `additionalContext` for the next request) or block it with corrective
      * `feedback` (Claude Code's `PostToolUse`). Listeners receive
      * `(exec, result, next)`: call `next()` to delegate to the default (accept
-     * unchanged), or return a {@link PostToolDecision} to override. The core tool
-     * dispatch sits between the two waterfalls as plain code, all inside
-     * `execute`'s outer try/catch (and the tool body keeps its own inner
-     * try/catch, so a thrown tool still reaches `post-execute` as an `isError`
-     * result).
+     * unchanged), or return a {@link PostToolDecision} to override. Core tool
+     * dispatch runs earlier as the base `next()` of the `tools/execute`
+     * waterfall, all inside `execute`'s outer try/catch (and the tool body keeps
+     * its own inner try/catch, so a thrown tool still reaches `post-execute` as an
+     * `isError` result).
      * @param exec - the call that just ran (name, parsed arguments, caller agent).
      * @param result - the dispatch outcome a listener may accept, replace, or block.
      * @mode waterfall
@@ -107,6 +138,14 @@ export type ToolExecuteReturn = ContentBlock[] | { content: ContentBlock[]; meta
 /** A registered tool: its schema plus the execution function. */
 export interface ToolDefinition extends ToolSchema {
   execute(args: unknown, exec: ToolExecution): Promise<ToolExecuteReturn>
+  /**
+   * Cooperative tool-call timeout budget in milliseconds. Omit for no deadline.
+   * Enforced by `@deepseek-ai/dsh-timeout-policy` (a `tools/execute` wrapper); it
+   * is NEVER sent to the model — `schemas()` whitelists only name/description/
+   * parameters. Declaring it asserts this tool forwards `exec.signal` to a
+   * cooperative implementation that can reach quiescence when the signal aborts.
+   */
+  timeoutMs?: number
   /**
    * Optional: how to present the PENDING state of one call in a UI, derived from
    * the call's `args` (parsed arguments, `unknown` — the tool validates/narrows
@@ -261,7 +300,7 @@ function errorInfo(error: unknown): ToolErrorInfo | undefined {
 
 /**
  * Tool registry (`ctx.tools`): tool plugins register definitions; the agent
- * loop executes calls through the `tools/pre-execute` → dispatch →
+ * loop executes calls through the `tools/pre-execute` → `tools/execute` →
  * `tools/post-execute` pipeline. The registry contributes its schemas into the
  * system-prompt assembly.
  */
@@ -335,18 +374,20 @@ export class ToolRegistry extends Service {
   }
 
   /**
-   * Execute one tool call through the `tools/pre-execute` → dispatch →
-   * `tools/post-execute` pipeline. The two waterfalls are the gate (allow/deny)
-   * and the inspect/transform seam; core dispatch sits between them as plain
-   * code. The whole thing is wrapped in one outer try/catch so a throwing
-   * listener (in either waterfall) becomes an `isError` result instead of
-   * failing the turn; the tool body ALSO keeps its own inner try/catch, so a
-   * thrown tool becomes an `isError` result that `post-execute` listeners can
-   * still inspect. If the tool is not registered, the result is an `isError`
-   * carrying a `UNKNOWN_TOOL` structured error. A thrown {@link HarnessError}
-   * surfaces its `{ name, code }` on the result.
+   * Execute one tool call through the `tools/pre-execute` → `tools/execute`
+   * (around dispatch) → `tools/post-execute` pipeline. `pre-execute` is the gate
+   * (allow/deny), `tools/execute` wraps core dispatch (a timeout/retry/metrics
+   * seam), and `post-execute` is the inspect/transform seam; core dispatch sits
+   * as the base `next()` of the `tools/execute` waterfall. The whole thing is
+   * wrapped in one outer try/catch so a throwing listener (in any waterfall)
+   * becomes an `isError` result instead of failing the turn; the tool body ALSO
+   * keeps its own inner try/catch, so a thrown tool becomes an `isError` result
+   * that `tools/execute` and `post-execute` listeners can still inspect. If the
+   * tool is not registered, the result is an `isError` carrying a `UNKNOWN_TOOL`
+   * structured error. A thrown {@link HarnessError} surfaces its `{ name, code }`
+   * on the result.
    * @param exec - the call to run (name, parsed arguments, caller agent, signal).
-   * @returns the final result after both waterfalls; failures resolve as
+   * @returns the final result after every waterfall; failures resolve as
    *   `isError` results, never rejections.
    */
   async execute(exec: ToolExecution): Promise<ToolExecutionResult> {
@@ -372,23 +413,30 @@ export class ToolRegistry extends Service {
         return await this.postExecute(exec, denied)
       }
 
-      // --- Core dispatch (plain code between the waterfalls). The tool body's
-      // own try/catch turns a throw into an isError result so post-execute can
-      // inspect it; an unknown tool routes through the same catch. ---
-      let result: ToolExecutionResult
-      try {
-        const tool = this.store.get(exec.name)
-        if (!tool) throw new ToolNotFoundError(exec.name)
-        // Normalize the two `execute` return shapes: a bare ContentBlock[] (no
-        // meta) or a { content, meta } object (a tool attaching a private
-        // presentation payload). An array IS the content; the object carries it.
-        const returned = await tool.execute(exec.arguments, exec)
-        const content = Array.isArray(returned) ? returned : returned.content
-        const meta = Array.isArray(returned) ? undefined : returned.meta
-        result = { callId: exec.callId, content, isError: false, ...meta !== undefined ? { meta } : {} }
-      } catch (error: unknown) {
-        result = toolErrorResult(exec.callId, error)
-      }
+      // --- Around-dispatch: tools/execute. The base `next` is the dispatch-
+      // with-normalization thunk — the tool body's own try/catch turns a throw
+      // into an isError result so a wrapper (and post-execute) can inspect it;
+      // an unknown tool routes through the same catch. A `tools/execute` listener
+      // (e.g. a timeout plugin) wraps this thunk: it may mutate `exec` before
+      // delegating and inspect the normalized result after. ---
+      const result = await this.ctx.waterfall(
+        this, 'tools/execute', exec,
+        async (): Promise<ToolExecutionResult> => {
+          try {
+            const tool = this.store.get(exec.name)
+            if (!tool) throw new ToolNotFoundError(exec.name)
+            // Normalize the two `execute` return shapes: a bare ContentBlock[] (no
+            // meta) or a { content, meta } object (a tool attaching a private
+            // presentation payload). An array IS the content; the object carries it.
+            const returned = await tool.execute(exec.arguments, exec)
+            const content = Array.isArray(returned) ? returned : returned.content
+            const meta = Array.isArray(returned) ? undefined : returned.meta
+            return { callId: exec.callId, content, isError: false, ...meta !== undefined ? { meta } : {} }
+          } catch (error: unknown) {
+            return toolErrorResult(exec.callId, error)
+          }
+        },
+      )
 
       return await this.postExecute(exec, result)
     } catch (error: unknown) {

packages/core/tools/src/json-schema.ts

@@ -0,0 +1,345 @@
+/**
+ * Structured-output JSON Schema subset: the vocabulary a caller uses to demand
+ * a machine-readable result from a subagent (`SubagentStartRequest.outputSchema`)
+ * or a workflow `agent()` call.
+ *
+ * This is deliberately NOT full JSON Schema. The schema travels verbatim to the
+ * model as a forced tool's `parameters`, and the value the model produces is
+ * validated here — so every accepted keyword must be one this module actually
+ * enforces. Accepting a keyword we don't enforce would validate less than the
+ * schema promises (accepted-then-ignored), so anything outside the subset is
+ * REJECTED LOUD by {@link assertSupportedOutputSchema} instead. The subset:
+ *
+ * - `type` — a single string (`object`/`array`/`string`/`number`/`integer`/
+ *   `boolean`/`null`); type ARRAYS (`["string","null"]`) are rejected.
+ * - `properties`/`required`/`additionalProperties` (boolean) on objects; every
+ *   `required` key must be declared in `properties`. `additionalProperties`
+ *   absent keeps standard JSON Schema semantics (extra keys allowed).
+ * - `items` on arrays (absent ⇒ any JSON items).
+ * - `enum` (non-empty, scalars only) and `const` (scalar) on scalar types.
+ * - Annotations `description`/`title`/`default`/`examples` are allowed and
+ *   ignored (they constrain nothing), except that they must still be JSON data
+ *   — the schema is serialized onto the wire, so a non-JSON annotation would be
+ *   silently mangled.
+ *
+ * Values checked by {@link validateStructuredValue} are expected to be plain
+ * host-realm JSON data (model tool-call arguments are parsed wire JSON; a
+ * caller holding foreign-realm data materializes it first).
+ *
+ * @module dsh-tools/json-schema
+ */
+
+import { assertNever, HarnessError } from '@deepseek-ai/dsh-llm'
+
+/** The scalar values `enum`/`const` may carry (finite numbers only). */
+export type StructuredScalar = string | number | boolean | null
+
+/** The `type` keywords the subset accepts. */
+export type StructuredSchemaType = 'object' | 'array' | 'string' | 'number' | 'integer' | 'boolean' | 'null'
+
+/**
+ * One node of the structured-output schema subset. Recursive via `properties`
+ * and `items`; see the module doc for the exact keyword semantics.
+ */
+export interface StructuredSchemaNode {
+  type: StructuredSchemaType
+  /** Nested property schemas (`type: 'object'` only). */
+  properties?: Record<string, StructuredSchemaNode>
+  /** Required property names; each must appear in `properties`. */
+  required?: string[]
+  /** `false` rejects undeclared keys; absent/`true` allows them (JSON Schema default). */
+  additionalProperties?: boolean
+  /** Item schema (`type: 'array'` only); absent ⇒ any JSON items. */
+  items?: StructuredSchemaNode
+  /** Allowed values (scalar types only). */
+  enum?: StructuredScalar[]
+  /** The single allowed value (scalar types only). */
+  const?: StructuredScalar
+  /** Annotation, ignored for validation. */
+  description?: string
+  /** Annotation, ignored for validation. */
+  title?: string
+  /** Annotation, ignored for validation (must still be JSON data). */
+  default?: unknown
+  /** Annotation, ignored for validation (must still be JSON data). */
+  examples?: unknown
+}
+
+/** A structured-output schema: an OBJECT-rooted {@link StructuredSchemaNode}. */
+export type StructuredOutputSchema = StructuredSchemaNode & { type: 'object' }
+
+/**
+ * Thrown by {@link assertSupportedOutputSchema} when a schema falls outside the
+ * supported subset. Extends {@link HarnessError} (`code: 'UNSUPPORTED_SCHEMA'`)
+ * so seam code and tool results can route on it; `violations` lists every
+ * offending path, not just the first.
+ */
+export class OutputSchemaError extends HarnessError {
+  /** The individual violation messages, in walk order. */
+  readonly violations: string[]
+
+  constructor(violations: string[]) {
+    super(`unsupported output schema: ${violations.join('; ')}`, 'UNSUPPORTED_SCHEMA')
+    this.name = 'OutputSchemaError'
+    this.violations = violations
+  }
+}
+
+/** The keywords the subset accepts, checked (`constraint`) or ignored (`annotation`). */
+const CONSTRAINT_KEYWORDS = new Set(['type', 'properties', 'required', 'additionalProperties', 'items', 'enum', 'const'])
+const ANNOTATION_KEYWORDS = new Set(['description', 'title', 'default', 'examples'])
+
+const SCHEMA_TYPES: readonly StructuredSchemaType[] = ['object', 'array', 'string', 'number', 'integer', 'boolean', 'null']
+
+/**
+ * Whether a value is a PLAIN JSON object — non-null, non-array, and with a
+ * prototype chain of at most one link (`null`-proto, or any realm's
+ * `Object.prototype`, whose own prototype is `null`). Realm-agnostic on
+ * purpose: a schema materialized in another realm carries THAT realm's
+ * `Object.prototype`, which an identity check would wrongly reject. Exotic
+ * hosts (`Date`, `Map`, class instances) have longer chains and are rejected —
+ * they would serialize lossily (`Date` → string, `Map` → `{}`) instead of
+ * failing loud.
+ */
+function isObjectLike(value: unknown): value is Record<string, unknown> {
+  if (typeof value !== 'object' || value === null || Array.isArray(value)) return false
+  const proto: unknown = Object.getPrototypeOf(value)
+  return proto === null || Object.getPrototypeOf(proto) === null
+}
+
+/** Whether a value is a supported scalar (`enum`/`const` member): string, finite number, boolean, or null. */
+function isStructuredScalar(value: unknown): value is StructuredScalar {
+  return value === null || typeof value === 'string' || typeof value === 'boolean'
+    || (typeof value === 'number' && Number.isFinite(value))
+}
+
+/**
+ * Whether a value is JSON data (annotation payloads only): scalars, arrays, and
+ * object-likes of such values. Realm-agnostic on purpose (no prototype check) —
+ * the schema may have been materialized from another realm; structural JSON-ness
+ * is what the wire needs. Cycles are rejected via `seen`.
+ */
+function isJsonData(value: unknown, seen: Set<object>): boolean {
+  if (isStructuredScalar(value)) return true
+  // The scalar check above already returned for null, so `object` here is a real object.
+  if (typeof value !== 'object') return false
+  if (seen.has(value)) return false
+  seen.add(value)
+  try {
+    if (Array.isArray(value)) return value.every(entry => isJsonData(entry, seen))
+    // A non-plain object (Date, Map, class instance) is NOT JSON data even when
+    // it has no enumerable values — it would serialize lossily, not loudly.
+    if (!isObjectLike(value)) return false
+    return Object.values(value).every(entry => isJsonData(entry, seen))
+  } finally {
+    seen.delete(value)
+  }
+}
+
+/** Collect subset violations for one schema node (recursive walk). */
+function checkSchemaNode(node: unknown, path: string, violations: string[], seen: Set<object>): void {
+  if (!isObjectLike(node)) {
+    violations.push(`${path} must be a schema object`)
+    return
+  }
+  if (seen.has(node)) {
+    violations.push(`${path} is circular`)
+    return
+  }
+  seen.add(node)
+
+  for (const key of Object.keys(node)) {
+    if (CONSTRAINT_KEYWORDS.has(key)) continue
+    if (ANNOTATION_KEYWORDS.has(key)) {
+      if (!isJsonData(node[key], new Set())) violations.push(`${path}.${key} annotation must be JSON data`)
+      continue
+    }
+    violations.push(`${path}.${key} is not a supported keyword (subset: type/properties/required/additionalProperties/items/enum/const + annotations)`)
+  }
+  if (typeof node.description !== 'undefined' && typeof node.description !== 'string') {
+    violations.push(`${path}.description must be a string`)
+  }
+  if (typeof node.title !== 'undefined' && typeof node.title !== 'string') {
+    violations.push(`${path}.title must be a string`)
+  }
+
+  const type = node.type
+  if (typeof type !== 'string' || !(SCHEMA_TYPES as readonly unknown[]).includes(type)) {
+    violations.push(Array.isArray(type)
+      ? `${path}.type must be a single type string (type arrays are not supported)`
+      : `${path}.type must be one of ${SCHEMA_TYPES.join('/')}`)
+    seen.delete(node)
+    return
+  }
+  const schemaType = type as StructuredSchemaType
+
+  // Keywords that only make sense on one type are rejected elsewhere — an
+  // `items` on an object (or `properties` on a string) is a schema-author bug
+  // the subset surfaces rather than ignores.
+  const allowedFor: Record<string, StructuredSchemaType[]> = {
+    properties: ['object'],
+    required: ['object'],
+    additionalProperties: ['object'],
+    items: ['array'],
+    enum: ['string', 'number', 'integer', 'boolean', 'null'],
+    const: ['string', 'number', 'integer', 'boolean', 'null'],
+  }
+  for (const [key, types] of Object.entries(allowedFor)) {
+    if (key in node && !types.includes(schemaType)) {
+      violations.push(`${path}.${key} is not supported on type "${schemaType}"`)
+    }
+  }
+
+  switch (schemaType) {
+    case 'object': {
+      const properties = node.properties
+      if (properties !== undefined) {
+        if (!isObjectLike(properties)) {
+          violations.push(`${path}.properties must be an object of schemas`)
+        } else {
+          for (const [key, child] of Object.entries(properties)) {
+            checkSchemaNode(child, `${path}.properties.${key}`, violations, seen)
+          }
+        }
+      }
+      const required = node.required
+      if (required !== undefined) {
+        if (!Array.isArray(required) || required.some(entry => typeof entry !== 'string')) {
+          violations.push(`${path}.required must be an array of strings`)
+        } else {
+          const declared = isObjectLike(properties) ? properties : {}
+          // The guard above proved every entry is a string.
+          for (const key of required as string[]) {
+            // Own-property check: `in` would let inherited names (`toString`)
+            // satisfy the declared-in-properties contract via the prototype.
+            if (!Object.hasOwn(declared, key)) violations.push(`${path}.required names "${key}" which is not in properties`)
+          }
+        }
+      }
+      if (node.additionalProperties !== undefined && typeof node.additionalProperties !== 'boolean') {
+        violations.push(`${path}.additionalProperties must be a boolean`)
+      }
+      break
+    }
+    case 'array': {
+      if (node.items !== undefined) checkSchemaNode(node.items, `${path}.items`, violations, seen)
+      break
+    }
+    case 'string':
+    case 'number':
+    case 'integer':
+    case 'boolean':
+    case 'null': {
+      const allowed = node.enum
+      if (allowed !== undefined) {
+        if (!Array.isArray(allowed) || allowed.length === 0 || !allowed.every(entry => isStructuredScalar(entry))) {
+          violations.push(`${path}.enum must be a non-empty array of scalars`)
+        }
+      }
+      if ('const' in node && !isStructuredScalar(node.const)) {
+        violations.push(`${path}.const must be a scalar`)
+      }
+      break
+    }
+    /* v8 ignore start -- defensive: schemaType was membership-checked against SCHEMA_TYPES above, so no runtime value reaches here */
+    default:
+      assertNever(schemaType, 'assertSupportedOutputSchema')
+    /* v8 ignore stop */
+  }
+
+  seen.delete(node)
+}
+
+/**
+ * Assert `schema` is a supported {@link StructuredOutputSchema} — object-rooted
+ * and entirely within the enforced subset. Throws {@link OutputSchemaError}
+ * (`UNSUPPORTED_SCHEMA`) listing EVERY violation; returns (and narrows) on
+ * success. Call this at the seam boundary, before any child is created.
+ * @param schema - the caller-supplied schema (unknown until asserted).
+ * @returns nothing — the assertion signature narrows `schema` to
+ * {@link StructuredOutputSchema} in the caller's scope on normal return.
+ */
+export function assertSupportedOutputSchema(schema: unknown): asserts schema is StructuredOutputSchema {
+  const violations: string[] = []
+  checkSchemaNode(schema, 'schema', violations, new Set())
+  if (violations.length === 0 && (schema as StructuredSchemaNode).type !== 'object') {
+    violations.push('schema.type must be "object" (structured output is object-rooted)')
+  }
+  if (violations.length > 0) throw new OutputSchemaError(violations)
+}
+
+/** Collect violations for one value against an (already asserted) schema node. */
+function checkValue(node: StructuredSchemaNode, value: unknown, path: string): string[] {
+  switch (node.type) {
+    case 'object': {
+      if (!isObjectLike(value)) return [`"${path}" must be an object`]
+      const violations: string[] = []
+      const properties = node.properties ?? {}
+      // Own-property discipline throughout: JSON carries own enumerable
+      // properties only, so an inherited `toString` must not satisfy
+      // `required`, dodge `additionalProperties: false`, or be validated as if
+      // the value carried it.
+      for (const key of node.required ?? []) {
+        if (!Object.hasOwn(value, key) || value[key] === undefined) violations.push(`missing required property "${path}.${key}"`)
+      }
+      for (const [key, child] of Object.entries(properties)) {
+        if (!Object.hasOwn(value, key) || value[key] === undefined) continue
+        violations.push(...checkValue(child, value[key], `${path}.${key}`))
+      }
+      if (node.additionalProperties === false) {
+        for (const key of Object.keys(value)) {
+          if (!Object.hasOwn(properties, key)) violations.push(`"${path}.${key}" is not a declared property (additionalProperties: false)`)
+        }
+      }
+      return violations
+    }
+    case 'array': {
+      if (!Array.isArray(value)) return [`"${path}" must be an array`]
+      if (!node.items) return []
+      const items = node.items
+      return value.flatMap((entry, index) => checkValue(items, entry, `${path}[${index}]`))
+    }
+    case 'string': {
+      if (typeof value !== 'string') return [`"${path}" must be a string`]
+      break
+    }
+    case 'number': {
+      if (typeof value !== 'number' || !Number.isFinite(value)) return [`"${path}" must be a finite number`]
+      break
+    }
+    case 'integer': {
+      if (typeof value !== 'number' || !Number.isInteger(value)) return [`"${path}" must be an integer`]
+      break
+    }
+    case 'boolean': {
+      if (typeof value !== 'boolean') return [`"${path}" must be a boolean`]
+      break
+    }
+    case 'null': {
+      if (value !== null) return [`"${path}" must be null`]
+      break
+    }
+    default:
+      return assertNever(node.type, 'validateStructuredValue')
+  }
+  // Scalar constraint checks, shared by every scalar branch above.
+  if (node.enum && !node.enum.includes(value)) {
+    return [`"${path}" must be one of ${JSON.stringify(node.enum)}`]
+  }
+  if ('const' in node && value !== node.const) {
+    return [`"${path}" must be ${JSON.stringify(node.const)}`]
+  }
+  return []
+}
+
+/**
+ * Validate a value against an (already {@link assertSupportedOutputSchema}-
+ * asserted) schema. Returns human-readable, path-qualified violation messages
+ * — empty means valid. Total: never throws, however malformed the value.
+ * @param schema - the asserted schema to check against.
+ * @param value - the candidate value (e.g. parsed tool-call arguments).
+ * @returns every violation found, in walk order (empty = valid).
+ */
+export function validateStructuredValue(schema: StructuredOutputSchema, value: unknown): string[] {
+  return checkValue(schema, value, 'value')
+}

packages/core/tools/src/schema.ts

@@ -155,6 +155,9 @@ export interface JsonSchemaObject {
  * `properties`, `required` array).
  *
  * This is a plain function — no schemastery or other framework dependency.
+ * @param spec - the author-facing per-property schema to convert.
+ * @returns the wire-format JSON Schema; the top-level `required` array is
+ *   omitted entirely when no property is marked required.
  */
 export function schemaSpecToJsonSchema(spec: SchemaSpec): JsonSchemaObject {
   const properties: Record<string, unknown> = {}
@@ -269,6 +272,9 @@ function checkSpec(spec: SchemaSpec, value: unknown, path: string): string[] {
  * keys are allowed (no `additionalProperties: false`); `default` is not
  * applied; an `object`/`array` prop without `properties`/`items` only
  * type-checks; `enum` is membership (strings only).
+ * @param spec - the declared parameter schema to validate against.
+ * @param args - the model-generated arguments, however malformed.
+ * @returns the violation messages in declaration order; empty means valid.
  */
 export function validateArgs(spec: SchemaSpec, args: unknown): string[] {
   return checkSpec(spec, args, '')
@@ -289,6 +295,13 @@ export interface DefineToolOptions<S extends SchemaSpec> {
    * standard JSON Schema at runtime.
    */
   parameters: S
+  /**
+   * Optional cooperative tool-call timeout budget in milliseconds. When given it
+   * must be a positive finite number; it is attached to the produced
+   * {@link ToolDefinition} for `@deepseek-ai/dsh-timeout-policy` to enforce and
+   * is never sent to the model.
+   */
+  timeoutMs?: number
   /**
    * Tool execution function. `args` is typed as {@link InferArgs<S>} — zero
    * casts needed. Returns either a bare {@link ContentBlock}`[]` (model-facing
@@ -340,6 +353,13 @@ export interface DefineToolOptions<S extends SchemaSpec> {
  * Raw JSON-Schema tool definitions (from MCP servers) are still accepted
  * by `ToolRegistry.register()` directly — `defineTool` is sugar for
  * first-party plugin authors.
+ * @param options - the tool's name, description, typed parameter schema,
+ *   execute body, and optional presenters.
+ * @returns a registry-ready {@link ToolDefinition}: its `execute` validates the
+ *   raw args first (throwing {@link ToolArgsError} on mismatch, which the
+ *   registry turns into an isError result), and its presenters validate softly
+ *   (returning undefined on mismatch, since replay may feed them older-schema
+ *   args).
  */
 export function defineTool<S extends SchemaSpec>(options: DefineToolOptions<S>): ToolDefinition {
   // Object-literal execute methods don't use `this`; the reference is safe.
@@ -349,10 +369,14 @@ export function defineTool<S extends SchemaSpec>(options: DefineToolOptions<S>):
   const userPresentCall = options.presentCall
   // eslint-disable-next-line @typescript-eslint/unbound-method
   const userPresentResult = options.presentResult
+  if (options.timeoutMs !== undefined && (!Number.isFinite(options.timeoutMs) || options.timeoutMs <= 0)) {
+    throw new Error(`defineTool(${options.name}): timeoutMs must be a positive finite number`)
+  }
   const tool: ToolDefinition = {
     name: options.name,
     description: options.description,
     parameters: schemaSpecToJsonSchema(options.parameters) as unknown as Record<string, unknown>,
+    ...(options.timeoutMs !== undefined ? { timeoutMs: options.timeoutMs } : {}),
     async execute(args: unknown, exec: ToolExecution): Promise<ToolExecuteReturn> {
       // Validate the model-generated args before the typed body runs. On
       // mismatch we throw ToolArgsError; the registry turns it into an

packages/core/tools/tests/json-schema.spec.ts

@@ -0,0 +1,304 @@
+import { describe, expect, it } from 'vitest'
+import {
+  assertSupportedOutputSchema,
+  OutputSchemaError,
+  validateStructuredValue,
+  type StructuredOutputSchema,
+} from '../src/json-schema.ts'
+
+/** Assert-and-narrow helper: the asserted schema, typed. */
+function asserted(schema: unknown): StructuredOutputSchema {
+  assertSupportedOutputSchema(schema)
+  return schema
+}
+
+/** The violations OutputSchemaError carries for a bad schema (throws if it passes). */
+function violationsOf(schema: unknown): string[] {
+  try {
+    assertSupportedOutputSchema(schema)
+  } catch (error: unknown) {
+    if (error instanceof OutputSchemaError) return error.violations
+    throw error
+  }
+  throw new Error('expected the schema to be rejected')
+}
+
+describe('assertSupportedOutputSchema', () => {
+  it('accepts a representative subset schema (all supported keywords)', () => {
+    const schema = asserted({
+      type: 'object',
+      description: 'a finding',
+      title: 'Finding',
+      properties: {
+        file: { type: 'string', description: 'path' },
+        line: { type: 'integer' },
+        severity: { type: 'string', enum: ['low', 'high'] },
+        kind: { type: 'string', const: 'bug' },
+        score: { type: 'number' },
+        confirmed: { type: 'boolean' },
+        parent: { type: 'null' },
+        tags: { type: 'array', items: { type: 'string' } },
+        nested: {
+          type: 'object',
+          properties: { x: { type: 'number', default: 3, examples: [1, 2] } },
+          additionalProperties: false,
+        },
+        anything: { type: 'array' },
+      },
+      required: ['file', 'line'],
+      additionalProperties: true,
+    })
+    expect(schema.type).toBe('object')
+  })
+
+  it('rejects a non-object root (scalar/array-rooted schemas)', () => {
+    expect(violationsOf({ type: 'string' })).toEqual(['schema.type must be "object" (structured output is object-rooted)'])
+    expect(violationsOf({ type: 'array', items: { type: 'string' } }))
+      .toContain('schema.type must be "object" (structured output is object-rooted)')
+  })
+
+  it('rejects non-object schema nodes and missing/unknown type', () => {
+    expect(violationsOf('nope')).toEqual(['schema must be a schema object'])
+    expect(violationsOf(null)).toEqual(['schema must be a schema object'])
+    expect(violationsOf([])).toEqual(['schema must be a schema object'])
+    expect(violationsOf({})).toEqual(['schema.type must be one of object/array/string/number/integer/boolean/null'])
+    expect(violationsOf({ type: 'tuple' })[0]).toMatch(/type must be one of/)
+    expect(violationsOf({ type: 'object', properties: { a: 'str' } })).toEqual(['schema.properties.a must be a schema object'])
+  })
+
+  it('rejects type ARRAYS with a dedicated message', () => {
+    expect(violationsOf({ type: ['string', 'null'] }))
+      .toEqual(['schema.type must be a single type string (type arrays are not supported)'])
+  })
+
+  it('rejects unsupported constraint keywords loudly (never accepted-then-ignored)', () => {
+    for (const keyword of ['oneOf', 'anyOf', 'allOf', 'not', 'pattern', 'minimum', 'maxLength', '$ref']) {
+      const bad = violationsOf({ type: 'object', [keyword]: [] })
+      expect(bad.some(v => v.includes(`schema.${keyword} is not a supported keyword`))).toBe(true)
+    }
+  })
+
+  it('reports EVERY violation, not just the first', () => {
+    const bad = violationsOf({
+      type: 'object',
+      pattern: 'x',
+      properties: { a: { type: 'weird' }, b: { type: 'string', minimum: 1 } },
+    })
+    expect(bad.length).toBe(3)
+  })
+
+  it('rejects keywords on the wrong type (items on object, properties on string, enum on object)', () => {
+    expect(violationsOf({ type: 'object', items: { type: 'string' } }))
+      .toEqual(['schema.items is not supported on type "object"'])
+    expect(violationsOf({ type: 'object', properties: { a: { type: 'string', properties: {} } } }))
+      .toEqual(['schema.properties.a.properties is not supported on type "string"'])
+    expect(violationsOf({ type: 'object', enum: [1] }))
+      .toEqual(['schema.enum is not supported on type "object"'])
+    expect(violationsOf({ type: 'object', properties: { a: { type: 'array', const: 1 } } }))
+      .toEqual(['schema.properties.a.const is not supported on type "array"'])
+  })
+
+  it('validates required: must be string[] naming declared properties', () => {
+    expect(violationsOf({ type: 'object', required: 'file' }))
+      .toEqual(['schema.required must be an array of strings'])
+    expect(violationsOf({ type: 'object', required: [1] }))
+      .toEqual(['schema.required must be an array of strings'])
+    expect(violationsOf({ type: 'object', properties: { a: { type: 'string' } }, required: ['b'] }))
+      .toEqual(['schema.required names "b" which is not in properties'])
+    expect(violationsOf({ type: 'object', required: ['a'] }))
+      .toEqual(['schema.required names "a" which is not in properties'])
+  })
+
+  it('validates additionalProperties must be boolean and enum/const must be scalars', () => {
+    expect(violationsOf({ type: 'object', additionalProperties: {} }))
+      .toEqual(['schema.additionalProperties must be a boolean'])
+    expect(violationsOf({ type: 'object', properties: { a: { type: 'string', enum: [] } } }))
+      .toEqual(['schema.properties.a.enum must be a non-empty array of scalars'])
+    expect(violationsOf({ type: 'object', properties: { a: { type: 'string', enum: [{}] } } }))
+      .toEqual(['schema.properties.a.enum must be a non-empty array of scalars'])
+    expect(violationsOf({ type: 'object', properties: { a: { type: 'string', enum: 'x' } } }))
+      .toEqual(['schema.properties.a.enum must be a non-empty array of scalars'])
+    expect(violationsOf({ type: 'object', properties: { a: { type: 'number', enum: [Number.NaN] } } }))
+      .toEqual(['schema.properties.a.enum must be a non-empty array of scalars'])
+    expect(violationsOf({ type: 'object', properties: { a: { type: 'string', const: {} } } }))
+      .toEqual(['schema.properties.a.const must be a scalar'])
+  })
+
+  it('rejects non-string description/title and non-JSON annotation payloads', () => {
+    expect(violationsOf({ type: 'object', description: 7 }))
+      .toEqual(['schema.description must be a string'])
+    expect(violationsOf({ type: 'object', title: 7 }))
+      .toEqual(['schema.title must be a string'])
+    expect(violationsOf({ type: 'object', default: () => 1 }))
+      .toEqual(['schema.default annotation must be JSON data'])
+    expect(violationsOf({ type: 'object', examples: [undefined] }))
+      .toEqual(['schema.examples annotation must be JSON data'])
+    expect(violationsOf({ type: 'object', examples: [Number.POSITIVE_INFINITY] }))
+      .toEqual(['schema.examples annotation must be JSON data'])
+    // A cyclic annotation payload is caught by the JSON-data walk.
+    const cyclicAnnotation: Record<string, unknown> = {}
+    cyclicAnnotation.self = cyclicAnnotation
+    expect(violationsOf({ type: 'object', default: cyclicAnnotation }))
+      .toEqual(['schema.default annotation must be JSON data'])
+    // Object/array annotations that ARE JSON data pass.
+    asserted({ type: 'object', default: { a: [1, 'x', null, true] } })
+  })
+
+  it('rejects a circular schema instead of recursing forever', () => {
+    const node: Record<string, unknown> = { type: 'object' }
+    node.properties = { self: node }
+    expect(violationsOf(node)).toEqual(['schema.properties.self is circular'])
+  })
+
+  it('accepts the same subschema object reused in two SIBLING positions (a DAG, not a cycle)', () => {
+    const leaf = { type: 'string' }
+    asserted({ type: 'object', properties: { a: leaf, b: leaf } })
+  })
+
+  it('required cannot be satisfied by INHERITED names — `toString` is not a declared property', () => {
+    // `'toString' in {}` is true via Object.prototype; the declared-property
+    // contract must be an own-property check.
+    expect(violationsOf({ type: 'object', properties: {}, required: ['toString'] }))
+      .toEqual(['schema.required names "toString" which is not in properties'])
+  })
+
+  it('rejects exotic host objects where the subset expects plain JSON structure', () => {
+    // A Map as `properties` has no own enumerable entries: structurally it
+    // would read as "no properties" and serialize to {} — lossy, not loud.
+    expect(violationsOf({ type: 'object', properties: new Map() }))
+      .toEqual(['schema.properties must be an object of schemas'])
+    // A Date node is not a schema object even though Object.values(date) is [].
+    expect(violationsOf({ type: 'object', properties: { at: new Date(0) } }))
+      .toEqual(['schema.properties.at must be a schema object'])
+  })
+
+  it('rejects exotic annotation payloads that would serialize lossily', () => {
+    expect(violationsOf({ type: 'object', default: new Date(0) }))
+      .toEqual(['schema.default annotation must be JSON data'])
+    expect(violationsOf({ type: 'object', examples: [new Map()] }))
+      .toEqual(['schema.examples annotation must be JSON data'])
+  })
+})
+
+describe('validateStructuredValue', () => {
+  const schema = asserted({
+    type: 'object',
+    properties: {
+      file: { type: 'string' },
+      line: { type: 'integer' },
+      score: { type: 'number' },
+      confirmed: { type: 'boolean' },
+      parent: { type: 'null' },
+      severity: { type: 'string', enum: ['low', 'high'] },
+      kind: { type: 'string', const: 'bug' },
+      tags: { type: 'array', items: { type: 'string' } },
+      free: { type: 'array' },
+      nested: { type: 'object', properties: { x: { type: 'number' } }, required: ['x'], additionalProperties: false },
+    },
+    required: ['file'],
+  })
+
+  it('accepts a fully valid value (empty violations)', () => {
+    expect(validateStructuredValue(schema, {
+      file: 'a.ts', line: 3, score: 0.5, confirmed: true, parent: null,
+      severity: 'high', kind: 'bug', tags: ['x'], free: [1, { any: true }], nested: { x: 1 },
+    })).toEqual([])
+  })
+
+  it('reports missing required and wrong root type', () => {
+    expect(validateStructuredValue(schema, {})).toEqual(['missing required property "value.file"'])
+    expect(validateStructuredValue(schema, 'nope')).toEqual(['"value" must be an object'])
+    expect(validateStructuredValue(schema, [])).toEqual(['"value" must be an object'])
+  })
+
+  it('type-checks every scalar branch with path-qualified messages', () => {
+    expect(validateStructuredValue(schema, { file: 1 })).toEqual(['"value.file" must be a string'])
+    expect(validateStructuredValue(schema, { file: 'a', line: 1.5 })).toEqual(['"value.line" must be an integer'])
+    expect(validateStructuredValue(schema, { file: 'a', line: 'x' })).toEqual(['"value.line" must be an integer'])
+    expect(validateStructuredValue(schema, { file: 'a', score: 'x' })).toEqual(['"value.score" must be a finite number'])
+    expect(validateStructuredValue(schema, { file: 'a', score: Number.NaN })).toEqual(['"value.score" must be a finite number'])
+    expect(validateStructuredValue(schema, { file: 'a', confirmed: 'yes' })).toEqual(['"value.confirmed" must be a boolean'])
+    expect(validateStructuredValue(schema, { file: 'a', parent: 0 })).toEqual(['"value.parent" must be null'])
+  })
+
+  it('enforces enum membership and const equality', () => {
+    expect(validateStructuredValue(schema, { file: 'a', severity: 'mid' }))
+      .toEqual(['"value.severity" must be one of ["low","high"]'])
+    expect(validateStructuredValue(schema, { file: 'a', kind: 'feature' }))
+      .toEqual(['"value.kind" must be "bug"'])
+  })
+
+  it('checks arrays per index; an items-less array accepts anything', () => {
+    expect(validateStructuredValue(schema, { file: 'a', tags: 'x' })).toEqual(['"value.tags" must be an array'])
+    expect(validateStructuredValue(schema, { file: 'a', tags: ['ok', 2] })).toEqual(['"value.tags[1]" must be a string'])
+    expect(validateStructuredValue(schema, { file: 'a', free: [{ deep: [1] }, null] })).toEqual([])
+  })
+
+  it('recurses into nested objects: required + additionalProperties: false', () => {
+    expect(validateStructuredValue(schema, { file: 'a', nested: {} }))
+      .toEqual(['missing required property "value.nested.x"'])
+    expect(validateStructuredValue(schema, { file: 'a', nested: { x: 1, y: 2 } }))
+      .toEqual(['"value.nested.y" is not a declared property (additionalProperties: false)'])
+    expect(validateStructuredValue(schema, { file: 'a', nested: 3 }))
+      .toEqual(['"value.nested" must be an object'])
+  })
+
+  it('a required key present-but-undefined counts as missing', () => {
+    expect(validateStructuredValue(schema, { file: undefined })).toEqual(['missing required property "value.file"'])
+  })
+
+  it('inherited properties satisfy nothing: required, additionalProperties, and recursion are own-property only', () => {
+    // required: ['toString'] must NOT be satisfied by Object.prototype.toString.
+    expect(validateStructuredValue(
+      asserted({ type: 'object', properties: { toString: { type: 'string' } }, required: ['toString'] }),
+      {},
+    )).toEqual(['missing required property "value.toString"'])
+    // additionalProperties: false must flag an OWN `toString` key even though
+    // `'toString' in properties` is true via the prototype.
+    expect(validateStructuredValue(
+      asserted({ type: 'object', additionalProperties: false }),
+      { toString: 1 },
+    )).toEqual(['"value.toString" is not a declared property (additionalProperties: false)'])
+    // A declared property the value does NOT carry must not be validated
+    // against the value's INHERITED member (constructor is a function on
+    // every plain object's prototype, not a carried property).
+    expect(validateStructuredValue(
+      asserted({ type: 'object', properties: { constructor: { type: 'string' } } }),
+      {},
+    )).toEqual([])
+  })
+
+  it('a non-plain object value is not an object in the JSON sense', () => {
+    expect(validateStructuredValue(asserted({ type: 'object' }), new Date(0)))
+      .toEqual(['"value" must be an object'])
+  })
+
+  it('collects multiple violations across branches in one pass', () => {
+    expect(validateStructuredValue(schema, { line: 'x', severity: 'mid' })).toEqual([
+      'missing required property "value.file"',
+      '"value.line" must be an integer',
+      '"value.severity" must be one of ["low","high"]',
+    ])
+  })
+
+  it('null-typed const/enum work through the scalar path', () => {
+    const nullish = asserted({ type: 'object', properties: { a: { type: 'null', const: null } } })
+    expect(validateStructuredValue(nullish, { a: null })).toEqual([])
+  })
+
+  it('rejects a non-object properties value in the schema walk', () => {
+    expect(violationsOf({ type: 'object', properties: [] }))
+      .toEqual(['schema.properties must be an object of schemas'])
+  })
+
+  it('an object schema without properties/required only type-checks its value', () => {
+    const bare = asserted({ type: 'object' })
+    expect(validateStructuredValue(bare, { any: ['thing'] })).toEqual([])
+    expect(validateStructuredValue(bare, 7)).toEqual(['"value" must be an object'])
+  })
+
+  it('validateStructuredValue throws on a type the assert would never let through (assertNever backstop)', () => {
+    const forged = { type: 'tuple' } as unknown as StructuredOutputSchema
+    expect(() => validateStructuredValue(forged, 1)).toThrow(/tuple/)
+  })
+})

packages/core/tools/tests/tools.spec.ts

@@ -62,6 +62,17 @@ describe('ToolRegistry', () => {
     expect(schema.execute).toBeUndefined()
   })
 
+  it('schemas() excludes timeoutMs — the budget must never reach the model', async () => {
+    const ctx = await setup()
+    ctx.tools.register(defineTool({
+      name: 'budgeted', description: 'has a budget', parameters: {}, timeoutMs: 5_000,
+      async execute() { return [{ type: 'text' as const, text: 'ok' }] },
+    }))
+    const schema = ctx.tools.schemas().find(s => s.name === 'budgeted')
+    expect(schema).toBeDefined()
+    expect('timeoutMs' in (schema as object)).toBe(false)
+  })
+
   it('executes a tool and returns its content', async () => {
     const ctx = await setup()
     ctx.tools.register(echoTool)
@@ -272,6 +283,148 @@ describe('ToolRegistry', () => {
     expect(order).toEqual(['pre:before', 'pre:after', 'post:before', 'post:after'])
   })
 
+  it('runs tools/execute after an allowed pre-execute, around dispatch, and before post-execute', async () => {
+    const ctx = await setup()
+    const order: string[] = []
+    ctx.tools.register(defineTool({
+      name: 'traced',
+      description: 'echo',
+      parameters: { text: { type: 'string' } },
+      async execute(args) {
+        order.push('dispatch')
+        return [{ type: 'text' as const, text: args.text ?? '' }]
+      },
+    }))
+
+    ctx.on('tools/pre-execute', async (_exec, next) => { order.push('pre'); return next() })
+    ctx.on('tools/execute', async (_exec, next) => {
+      order.push('execute:before')
+      const result = await next()
+      order.push('execute:after')
+      return result
+    })
+    ctx.on('tools/post-execute', async (_exec, _result, next) => { order.push('post'); return next() })
+
+    const result = await ctx.tools.execute({ callId: CallId('c1'), name: 'traced', arguments: { text: 'hi' } })
+    expect(result).toEqual({ callId: CallId('c1'), content: [{ type: 'text', text: 'hi' }], isError: false })
+    // The around seam wraps dispatch; pre gates before it, post runs over its result.
+    expect(order).toEqual(['pre', 'execute:before', 'dispatch', 'execute:after', 'post'])
+  })
+
+  it('a pre-execute deny short-circuits before tools/execute (the seam never runs)', async () => {
+    const ctx = await setup()
+    ctx.tools.register(echoTool)
+
+    let entered = false
+    ctx.on('tools/pre-execute', async (_exec, _next): Promise<PreToolDecision> => ({ kind: 'deny', reason: 'nope' }))
+    ctx.on('tools/execute', async (_exec, next) => { entered = true; return next() })
+
+    const result = await ctx.tools.execute({ callId: CallId('c1'), name: 'echo', arguments: { text: 'hi' } })
+    expect(result.isError).toBe(true)
+    expect(result.content[0]).toMatchObject({ text: 'Error: nope' })
+    expect(entered).toBe(false) // a denied call never enters the around-dispatch seam
+  })
+
+  it('a thrown tool is normalized to an isError result BEFORE a tools/execute listener sees next()', async () => {
+    const ctx = await setup()
+    ctx.tools.register({
+      ...echoTool,
+      name: 'boom',
+      async execute() { throw new HarnessError('kaboom', 'BOOM') },
+    })
+
+    let seen: { isError: boolean; error?: unknown } | undefined
+    ctx.on('tools/execute', async (_exec, next) => {
+      const result = await next()
+      // The base next() IS dispatch-with-normalization: the wrapper sees the
+      // normalized isError result, never a raw throw from the tool body.
+      seen = { isError: result.isError, error: result.error }
+      return result
+    })
+
+    const result = await ctx.tools.execute({ callId: CallId('c1'), name: 'boom', arguments: {} })
+    expect(seen).toEqual({ isError: true, error: { name: 'HarnessError', code: 'BOOM' } })
+    expect(result.isError).toBe(true)
+    expect(result.content[0]).toMatchObject({ text: 'Error: kaboom' })
+  })
+
+  it('a thrown tool normalized inside tools/execute still reaches post-execute', async () => {
+    const ctx = await setup()
+    ctx.tools.register({
+      ...echoTool,
+      name: 'boom',
+      async execute() { throw new Error('exploded') },
+    })
+
+    let postSaw: boolean | undefined
+    ctx.on('tools/execute', async (_exec, next) => next())
+    ctx.on('tools/post-execute', async (_exec, result, next) => {
+      postSaw = result.isError
+      return next()
+    })
+
+    const result = await ctx.tools.execute({ callId: CallId('c1'), name: 'boom', arguments: {} })
+    expect(postSaw).toBe(true) // the normalized isError still flows through post-execute
+    expect(result.isError).toBe(true)
+    expect(result.content[0]).toMatchObject({ text: 'Error: exploded' })
+  })
+
+  it('a tools/execute listener can replace exec.signal for the dispatched tool (deadline pattern)', async () => {
+    const ctx = await setup()
+    let seenSignal: AbortSignal | undefined
+    ctx.tools.register({
+      ...echoTool,
+      name: 'signal-probe',
+      async execute(_args, exec) {
+        seenSignal = exec.signal
+        return [{ type: 'text' as const, text: 'ok' }]
+      },
+    })
+
+    const upstream = new AbortController().signal
+    const replacement = new AbortController().signal
+    ctx.on('tools/execute', async (exec, next) => {
+      expect(exec.signal).toBe(upstream)
+      // Cordis next() ignores passed arguments, so a wrapper mutates exec in
+      // place (the documented "mutate the shared object, then delegate" idiom).
+      exec.signal = replacement
+      return next()
+    })
+
+    await ctx.tools.execute({ callId: CallId('c1'), name: 'signal-probe', arguments: {}, signal: upstream })
+    expect(seenSignal).toBe(replacement) // dispatch saw the wrapper's replacement, not the upstream
+  })
+
+  it('a tools/execute listener can short-circuit dispatch by returning a result without next()', async () => {
+    const ctx = await setup()
+    let dispatched = false
+    ctx.tools.register({
+      ...echoTool,
+      name: 'never-runs',
+      async execute() { dispatched = true; return [] },
+    })
+
+    ctx.on('tools/execute', async (exec, _next): Promise<import('@deepseek-ai/dsh-tools').ToolExecutionResult> =>
+      ({ callId: exec.callId, content: [{ type: 'text', text: 'short-circuited' }], isError: false }))
+
+    const result = await ctx.tools.execute({ callId: CallId('c1'), name: 'never-runs', arguments: {} })
+    expect(dispatched).toBe(false) // returning without next() skips core dispatch
+    expect(result.content[0]).toMatchObject({ text: 'short-circuited' })
+  })
+
+  it('returns an isError result when a tools/execute listener throws', async () => {
+    const ctx = await setup()
+    ctx.tools.register(echoTool)
+    ctx.on('tools/execute', async () => { throw new Error('wrapper broke') })
+
+    const result = await ctx.tools.execute({ callId: CallId('c1'), name: 'echo', arguments: { text: 'hi' } })
+    expect(result).toEqual({
+      callId: CallId('c1'),
+      content: [{ type: 'text', text: 'Error: wrapper broke' }],
+      isError: true,
+    })
+  })
+
   it('returns an isError result when a tools/pre-execute listener throws', async () => {
     const ctx = await setup()
     ctx.tools.register(echoTool)
@@ -989,6 +1142,38 @@ describe('defineTool validation (the runtime-validation RFC, part 1)', () => {
     const result = await ctx.tools.execute({ callId: CallId('c1'), name: 'raw', arguments: {} })
     expect(result.isError).toBe(false)
   })
+
+  it('attaches a positive-finite timeoutMs to the definition', () => {
+    const tool = defineTool({
+      name: 'x', description: 'd', parameters: {}, timeoutMs: 30_000,
+      async execute() { return [{ type: 'text' as const, text: 'ok' }] },
+    })
+    expect(tool.timeoutMs).toBe(30_000)
+  })
+
+  it('omits timeoutMs when not declared', () => {
+    const tool = defineTool({
+      name: 'x', description: 'd', parameters: {},
+      async execute() { return [{ type: 'text' as const, text: 'ok' }] },
+    })
+    expect(tool.timeoutMs).toBeUndefined()
+  })
+
+  it('throws when timeoutMs is zero or negative', () => {
+    const make = (ms: number) => defineTool({
+      name: 'x', description: 'd', parameters: {}, timeoutMs: ms,
+      async execute() { return [{ type: 'text' as const, text: 'ok' }] },
+    })
+    expect(() => make(0)).toThrow('timeoutMs must be a positive finite number')
+    expect(() => make(-5)).toThrow('positive finite number')
+  })
+
+  it('throws when timeoutMs is non-finite', () => {
+    expect(() => defineTool({
+      name: 'x', description: 'd', parameters: {}, timeoutMs: Infinity,
+      async execute() { return [{ type: 'text' as const, text: 'ok' }] },
+    })).toThrow('positive finite number')
+  })
 })
 
 describe('defineTool presentation (presentCall / presentResult)', () => {

packages/fs/fs-local/src/fsio.ts

@@ -129,6 +129,9 @@ export interface LocalDirEntry {
  * and intermediate directories are created by the write. Two input paths
  * reaching the same file via symlinks share one key. Falls back to the absolute
  * path only when no ancestor (not even the filesystem root) can be resolved.
+ * @param cwd - base directory a relative `path` resolves against.
+ * @param path - absolute or relative path; empty/whitespace-only throws `FS_NOT_FOUND`.
+ * @returns the absolute display path plus the realpath-derived stable target key.
  */
 export async function resolveLocalTarget(cwd: string, path: string): Promise<LocalTarget> {
   if (path.trim().length === 0) throw new FsError('file_path must be a non-empty string', 'FS_NOT_FOUND')
@@ -165,7 +168,11 @@ export async function resolveLocalTarget(cwd: string, path: string): Promise<Loc
   }
 }
 
-/** Probe a path for its version, mode, type, and size. Null if absent. */
+/**
+ * Probe a path for its version, mode, type, and size. Null if absent.
+ * @param absolutePath - the path to stat (typically a target key; symlinks are followed).
+ * @returns the metadata, or null when the path — or a parent segment — does not exist.
+ */
 export async function probe(absolutePath: string): Promise<PathInfo | null> {
   try {
     const info = await stat(absolutePath)
@@ -200,6 +207,9 @@ async function resolveListedChildTarget(parent: LocalTarget, name: string): Prom
  * List direct children of a directory in stable name order. Each child includes
  * a resolved target plus stat metadata when still available; file contents are
  * never read.
+ * @param target - the resolved directory to list; a missing or non-directory target throws.
+ * @param signal - aborts the listing, checked between children (`FS_ABORTED`).
+ * @returns one entry per direct child, sorted by name.
  */
 export async function listDirectory(target: LocalTarget, signal?: AbortSignal): Promise<LocalDirEntry[]> {
   throwIfAborted(signal, 'list')
@@ -290,6 +300,9 @@ async function statRegularFile(target: LocalTarget, verb: 'read', signal?: Abort
 /**
  * Read a whole regular UTF-8 text file into a single decoded string. Rejects
  * non-regular files, invalid UTF-8, and NUL-byte binary samples.
+ * @param target - the resolved file to read.
+ * @param signal - aborts the read (`FS_ABORTED`).
+ * @returns the full decoded text, byte-for-byte (no normalization).
  */
 export async function readWholeText(target: LocalTarget, signal?: AbortSignal): Promise<string> {
   await statRegularFile(target, 'read', signal)
@@ -305,6 +318,9 @@ export async function readWholeText(target: LocalTarget, signal?: AbortSignal):
  * Stream a whole regular UTF-8 text file as decoded text chunks. Same text
  * semantics as {@link readWholeText} (regular-file check, binary/NUL rejection,
  * cross-chunk UTF-8 decoding), but never holds the whole file in memory.
+ * @param target - the resolved file to stream.
+ * @param signal - aborts the stream, including between chunks (`FS_ABORTED`).
+ * @returns decoded text chunks in file order; chunk boundaries carry no meaning.
  */
 export async function* streamWholeText(target: LocalTarget, signal?: AbortSignal): AsyncIterable<string> {
   await statRegularFile(target, 'read', signal)
@@ -352,6 +368,11 @@ async function removeStagingDirOrThrow(stagingDir: string, originalError: unknow
  * (`0o700`) staging directory, fsync, optionally chmod to the final mode while
  * still private, then rename over the target. `mode` (when given) preserves an
  * existing file's permissions across the replace.
+ * @param absolutePath - the final destination (typically a target key); missing parent dirs are created.
+ * @param content - the full UTF-8 text to write.
+ * @param mode - final file mode applied before the rename (an existing file's, to preserve permissions); undefined leaves `0o600`.
+ * @param signal - aborts the write (`FS_ABORTED`); checked before the rename, so the target is never left torn.
+ * @param internals - test seam for pinning temp names and observing the staged file.
  */
 export async function writeFileAtomic(
   absolutePath: string,
@@ -409,6 +430,12 @@ export async function writeFileAtomic(
 /** Line ending style detected before LF normalization. */
 export type LineEndings = 'LF' | 'CRLF'
 
+/**
+ * Collapse CRLF to LF — the canonical in-memory form every edit/diff basis
+ * uses. Lone `\r` bytes (not followed by `\n`) are left untouched.
+ * @param content - decoded text in whatever line-ending style the file had.
+ * @returns the text with every `\r\n` pair replaced by `\n`.
+ */
 function normalizeLineEndings(content: string): string {
   return content.replaceAll('\r\n', '\n')
 }
@@ -420,6 +447,14 @@ function detectLineEndings(raw: string): LineEndings {
   return crlfCount > lfCount ? 'CRLF' : 'LF'
 }
 
+/**
+ * Convert LF-normalized content back to the line-ending style detected at read
+ * time, for write-back. `LF` returns the content unchanged; `CRLF` re-normalizes
+ * first so an already-CRLF sequence is never doubled to `\r\r\n`.
+ * @param content - the LF-normalized (edited) text.
+ * @param lineEndings - the original file's style, as detected by {@link readForEdit}.
+ * @returns the text in the original file's line-ending style.
+ */
 function restoreLineEndings(content: string, lineEndings: LineEndings): string {
   return lineEndings === 'LF' ? content : normalizeLineEndings(content).split('\n').join('\r\n')
 }
@@ -438,6 +473,10 @@ function countOccurrences(content: string, needle: string): number {
 /**
  * Read and decode a file for editing: rejects binaries, returns LF-normalized
  * content plus the original line-ending style for write-back.
+ * @param absolutePath - the file to read (typically a target key).
+ * @param displayPath - the caller-facing path used in error messages.
+ * @param signal - aborts the read (`FS_ABORTED`).
+ * @returns the LF-normalized content and the detected style to restore on write-back.
  */
 export async function readForEdit(
   absolutePath: string,
@@ -459,6 +498,9 @@ export async function readForEdit(
  * prior bytes, so an undiffable prior file simply yields no contextual-hunk basis
  * (the caller treats `null` the same as an absent file: the result renders a
  * whole-file diff rather than an applied hunk).
+ * @param absolutePath - the file to read (typically a target key); it must exist.
+ * @param signal - aborts the read (`FS_ABORTED`).
+ * @returns the LF-normalized text, or null for a binary or non-UTF-8 file.
  */
 export async function readTextForDiff(absolutePath: string, signal?: AbortSignal): Promise<string | null> {
   const buffer = await readFileAbortable(absolutePath, 'read', signal)
@@ -477,6 +519,12 @@ export async function readTextForDiff(absolutePath: string, signal?: AbortSignal
  * `FS_EDIT_NOT_FOUND` on empty `oldString` or zero matches and
  * `FS_AMBIGUOUS_EDIT` on multiple matches when `replaceAll` is false. Returns
  * the edited content (still LF-normalized) and the replacement count.
+ * @param content - the current file content, already LF-normalized.
+ * @param oldString - literal text to find; CRLF inside it is normalized to LF before matching.
+ * @param newString - literal replacement text, normalized the same way.
+ * @param replaceAll - replace every match instead of requiring exactly one.
+ * @param displayPath - the caller-facing path used in error messages.
+ * @returns the edited LF-normalized content plus how many occurrences were replaced.
  */
 export function applyLiteralEdit(
   content: string,

packages/fs/fs-local/src/index.ts

@@ -73,6 +73,7 @@ export class LocalFileSystem extends FileSystem {
     cwd: z.string().default(process.cwd()),
   })
 
+  /** Validated config (schemastery applied the defaults before construction). */
   readonly config: ResolvedConfig
   /** Test seam forwarded to fsio (force streaming path, pin temp names). */
   internals: FsIoInternals = {}

packages/fs/fs/src/types.ts

@@ -29,7 +29,12 @@ import type { Branded } from '@deepseek-ai/dsh-brand'
  */
 export type FsTargetKey = Branded<'FsTargetKey'>
 
-/** Brand a string as an {@link FsTargetKey}. */
+/**
+ * Brand a string as an {@link FsTargetKey}. For backend use only — a consumer
+ * never manufactures a key, it receives one from `resolve()`.
+ * @param key - the backend's raw key string (the local backend passes a realpath).
+ * @returns the same string, branded; no validation is performed.
+ */
 export function FsTargetKey(key: string): FsTargetKey {
   return key as FsTargetKey
 }
@@ -42,7 +47,12 @@ export function FsTargetKey(key: string): FsTargetKey {
  */
 export type FsVersion = Branded<'FsVersion'>
 
-/** Brand a string as an {@link FsVersion}. */
+/**
+ * Brand a string as an {@link FsVersion}. For backend use only — a consumer
+ * never manufactures a version, it receives one from `stat`/write/edit outcomes.
+ * @param v - the backend's raw version string (the local backend derives it from mtime+size).
+ * @returns the same string, branded; no validation is performed.
+ */
 export function FsVersion(v: string): FsVersion {
   return v as FsVersion
 }

packages/fs/tool-fs/src/diff.ts

@@ -40,6 +40,10 @@ export type FsDiffMeta = { diffs: FileDiff[] }
  * (a pure insertion) reports `oldText: null` (nothing to diff against), mirroring
  * the call-time card's new-file convention. The unified-diff "\ No newline at end
  * of file" markers are dropped — they annotate the patch, not file content.
+ * @param path - the path stamped on every produced diff (the model-facing `file_path`; the bridge relativizes it).
+ * @param before - the file text before the change (the backend's LF-normalized diff basis).
+ * @param after - the file text after the change, on the same basis.
+ * @returns one diff per applied hunk, in file order; empty when the texts are identical.
  */
 export function computeHunkDiffs(path: string, before: string, after: string): FileDiff[] {
   const patch = structuredPatch('', '', before, after, undefined, undefined, { context: DIFF_CONTEXT })
@@ -83,6 +87,8 @@ function isFileDiff(value: unknown): value is FileDiff {
  * it validates defensively rather than trusting the payload — a bad `meta` yields
  * `undefined`, and the caller decides the fallback (edit → the generic result
  * rendering; write → an args-derived whole-file diff), never a thrown presenter.
+ * @param meta - the opaque `tool/result` meta payload (live or replayed from the session log).
+ * @returns the validated non-empty hunk list, or undefined for an absent/empty/malformed payload.
  */
 export function diffsFromMeta(meta: unknown): FileDiff[] | undefined {
   if (typeof meta !== 'object' || meta === null || Array.isArray(meta)) return undefined

packages/fs/tool-fs/src/edit.ts

@@ -29,7 +29,13 @@ interface EditInput {
   replaceAll: boolean
 }
 
-/** Validate value constraints the schema DSL can't express. */
+/**
+ * Validate value constraints the schema DSL can't express: a non-blank
+ * `file_path`, a non-empty `old_string`, and `old_string !== new_string`
+ * (an equal pair would be a guaranteed no-op edit).
+ * @param args - the schema-validated raw tool arguments.
+ * @returns the camelCased input with `replace_all` defaulted to false.
+ */
 export function parseEditArgs(args: { file_path: string; old_string: string; new_string: string; replace_all?: boolean }): EditInput {
   if (args.file_path.trim().length === 0) throw new Error('file_path must be a non-empty string')
   if (args.old_string.length === 0) throw new Error('old_string must be a non-empty string')
@@ -42,14 +48,22 @@ export function parseEditArgs(args: { file_path: string; old_string: string; new
   }
 }
 
-/** Format an edit success (single-match or replace-all) as a Claude-style model-facing message. */
+/**
+ * Format an edit success (single-match or replace-all) as a Claude-style model-facing message.
+ * @param displayPath - the backend-resolved path shown to the model.
+ * @param replaceAll - selects the all-occurrences wording over the single-replacement one.
+ * @returns the confirmation sentence the model sees as the tool result.
+ */
 export function formatEditOutput(displayPath: string, replaceAll: boolean): string {
   return replaceAll
     ? `The file ${displayPath} has been updated. All occurrences were successfully replaced.`
     : `The file ${displayPath} has been updated successfully.`
 }
 
-/** Register the `edit` tool and its system-prompt guidance. */
+/**
+ * Register the `edit` tool and its system-prompt guidance.
+ * @param ctx - the plugin context; registrations are effects scoped to it, and execution uses its `fs` service.
+ */
 export function applyEditTool(ctx: Context): void {
   ctx.systemPrompt.section({
     name: 'tool:edit',

packages/fs/tool-fs/src/read-render.ts

@@ -119,6 +119,10 @@ function finish(acc: WindowAccumulator, request: ReadWindow, displayPath: string
  * path serves both. Scans for newlines with a capped line buffer (a newline-free
  * giant line is truncated, never buffered past `request.maxLineLength`),
  * enforces the byte cap, and throws `FS_NOT_FOUND` for an offset past EOF.
+ * @param chunks - decoded text chunks in file order; chunk boundaries carry no meaning.
+ * @param request - the resolved window; the caller has already applied its defaults and caps.
+ * @param displayPath - the caller-facing path used in the offset-out-of-range error.
+ * @returns the numbered window lines, the total line count seen, and the byte-cap truncation flag.
  */
 export async function buildWindow(
   chunks: AsyncIterable<string> | Iterable<string>,
@@ -156,7 +160,12 @@ export async function buildWindow(
   return finish(acc, request, displayPath)
 }
 
-/** Format a read outcome as one OpenCode-style line-numbered text block body. */
+/**
+ * Format a read outcome as one OpenCode-style line-numbered text block body.
+ * @param displayPath - the backend-resolved path rendered in the envelope's `<path>` element.
+ * @param outcome - the windowed read to render.
+ * @returns the model-facing envelope: numbered lines plus a continuation or end-of-file footer.
+ */
 export function formatReadOutput(displayPath: string, outcome: FileReadOutcome): string {
   const endLine = outcome.lines.at(-1)?.number ?? Math.max(0, outcome.offset - 1)
   let footer: string

packages/fs/tool-fs/src/read.ts

@@ -58,7 +58,12 @@ function parsePositiveInteger(value: number, name: string): number {
   return value
 }
 
-/** Validate value constraints the schema DSL can't express. `maxLimit` is the deployment's line cap. */
+/**
+ * Validate value constraints the schema DSL can't express. `maxLimit` is the deployment's line cap.
+ * @param args - the schema-validated raw tool arguments; `offset`/`limit` must be positive integers when given.
+ * @param maxLimit - the configured line cap: both the default `limit` and the largest one accepted.
+ * @returns the validated input with `offset` defaulted to 1 and `limit` to `maxLimit`.
+ */
 export function parseReadArgs(args: { file_path: string; offset?: number; limit?: number }, maxLimit: number): ReadInput {
   if (args.file_path.trim().length === 0) throw new Error('file_path must be a non-empty string')
   const offset = args.offset === undefined ? 1 : parsePositiveInteger(args.offset, 'offset')
@@ -67,7 +72,11 @@ export function parseReadArgs(args: { file_path: string; offset?: number; limit?
   return { filePath: args.file_path, offset, limit }
 }
 
-/** Register the `read` tool and its system-prompt guidance. */
+/**
+ * Register the `read` tool and its system-prompt guidance.
+ * @param ctx - the plugin context; registrations are effects scoped to it, and execution uses its `fs` service.
+ * @param caps - the deployment's resolved read caps (plugin config after defaulting).
+ */
 export function applyReadTool(ctx: Context, caps: ReadToolCaps): void {
   ctx.systemPrompt.section({
     name: 'tool:read',

packages/fs/tool-fs/src/session-cwd.ts

@@ -18,7 +18,11 @@
 
 import type { ToolExecution } from '@deepseek-ai/dsh-tools'
 
-/** The session workspace cwd for this call, or `undefined` when none applies. */
+/**
+ * The session workspace cwd for this call, or `undefined` when none applies.
+ * @param exec - the tool-execution context; only its optional `agent` is read.
+ * @returns the calling agent's session cwd, or undefined for a non-agent caller (the backend then applies its own default).
+ */
 export function sessionCwd(exec: ToolExecution): string | undefined {
   return exec.agent?.session.header.cwd
 }

packages/fs/tool-fs/src/write.ts

@@ -21,13 +21,23 @@ import type {} from '@deepseek-ai/dsh-system-prompt'
 import { computeHunkDiffs, diffsFromMeta, type FsDiffMeta } from './diff.ts'
 import { sessionCwd } from './session-cwd.ts'
 
-/** Validate value constraints the schema DSL can't express. */
+/**
+ * Validate value constraints the schema DSL can't express: only a non-blank
+ * `file_path` — an empty `content` is legitimate (it writes an empty file).
+ * @param args - the schema-validated raw tool arguments.
+ * @returns the camelCased input; `content` passes through untouched.
+ */
 export function parseWriteArgs(args: { file_path: string; content: string }): { filePath: string; content: string } {
   if (args.file_path.trim().length === 0) throw new Error('file_path must be a non-empty string')
   return { filePath: args.file_path, content: args.content }
 }
 
-/** Format a write outcome as one model-facing text block body. */
+/**
+ * Format a write outcome as one model-facing text block body.
+ * @param displayPath - the backend-resolved path rendered in the envelope's `<path>` element.
+ * @param outcome - the write outcome; its `operation` selects the Created/Updated wording.
+ * @returns the model-facing confirmation envelope (no file content is echoed back).
+ */
 export function formatWriteOutput(displayPath: string, outcome: FsWriteOutcome): string {
   const verb = outcome.operation === 'create' ? 'Created' : 'Updated'
   return `<path>${displayPath}</path>
@@ -37,7 +47,10 @@ ${verb} file
 </content>`
 }
 
-/** Register the `write` tool and its system-prompt guidance. */
+/**
+ * Register the `write` tool and its system-prompt guidance.
+ * @param ctx - the plugin context; registrations are effects scoped to it, and execution uses its `fs` service.
+ */
 export function applyWriteTool(ctx: Context): void {
   ctx.systemPrompt.section({
     name: 'tool:write',

packages/guard/README.md

@@ -0,0 +1,9 @@
+# guard/ — loop-hygiene guard family
+
+Behavioral guard plugins that watch the agent loop for unproductive patterns and nudge the model back on course. A single **product** package — there is no interface/implementation seam here, because a guard is a self-contained consumer of existing core seams (`tools/post-execute`, `agent/prompt-submit`, `agent/status`), not a swappable capability.
+
+| Package | Role | ctx key |
+|---|---|---|
+| `repeat-tool-guard/` | Advisory reminders when an agent loops on identical tool calls | (listens on `ctx.tools`' waterfalls) |
+
+Reminders travel as `additionalContext` on the `tools/post-execute` decision; the agent loop appends them as logged `context/message` events after the step's tool results (see [the tools package](../core/tools)), so everything a guard says to the model is reconstructable from the session log.

packages/guard/repeat-tool-guard/README.md

@@ -0,0 +1,37 @@
+# @deepseek-ai/dsh-repeat-tool-guard
+
+An advisory loop-breaker, not a model-facing tool: it never appears in the tool list, never vetoes or rewrites a call, and adds exactly one behavior — it watches each agent's stream of tool calls, counts runs of consecutive calls to the same tool with identical canonicalized arguments, and at configured run lengths injects an escalating advisory reminder telling the model to stop repeating itself, re-read the last result, and either change approach or conclude. The decision (retry differently, gather more evidence, or finish) stays entirely with the model: a legitimately repeated call is delayed by nothing and blocked by nothing. Decision record: [the repeat-tool-guard RFC](../../../docs/rfc/implemented/feature/2026-07-08-repeat-tool-guard.md).
+
+## Config
+
+```yaml
+- id: repeat-tool-guard
+  name: '@deepseek-ai/dsh-repeat-tool-guard'
+  config:
+    thresholds: [3, 5, 8]        # default; consecutive counts that trigger a reminder
+    include: []                  # tool-name patterns to track; empty ⇒ all tools
+    exclude: [todo_write]        # tool-name patterns transparent to the chain
+    argumentsPreviewChars: 500   # default; cap on arguments quoted in the detailed reminder
+```
+
+`thresholds` fails loud at plugin load: an empty list, a non-integer, a value below 2, or a duplicate throws, never a silent fall-back to defaults; `argumentsPreviewChars` equally rejects anything but an integer >= 1. The list is normalized to ascending order; the FIRST threshold delivers a short generic nudge, every later threshold delivers the detailed form naming the tool, the run length, and the canonical arguments — head-truncated at `argumentsPreviewChars` with an omitted-count marker, so a looping `write`/`edit` payload cannot ride into the next request unbounded (the chain key always compares the FULL canonical string; the cap bounds the reminder, never the detection).
+
+`include`/`exclude` entries support `*` wildcards and are predicates over whatever tools exist at call time, not references to registry entries — a pattern matching no currently registered tool is NOT an error (`exclude: [mcp_*]` stays valid in a deployment that loads no MCP tools), unlike `toolOrder`'s referent check.
+
+## Chain semantics
+
+The chain key is `(tool name, canonical arguments)` — canonicalization is a deep key-sort plus `JSON.stringify`, so argument objects differing only in property order count as identical. A call identical to the previous tracked call increments the agent's consecutive counter; a different tracked call resets it to 1.
+
+- **Untracked calls are transparent to the chain.** A call excluded by `include`/`exclude` neither increments nor resets the counter, so `grep X → todo_write → grep X` still counts as two consecutive `grep X` when `todo_write` is excluded. This is what makes exclusion useful: bookkeeping tools interleaved into a loop must not launder it.
+- **Denied calls count.** Detection sits on `tools/post-execute`, which also runs for calls a `tools/pre-execute` listener denied — a model hammering a denied call is exactly the loop worth breaking.
+- **Calls without an agent are ignored.** A direct `ctx.tools.execute()` caller has no model to remind and no `AgentId` to key on.
+- **Per-agent keying.** The tool registry is context-level and subagents interleave through the same waterfall, so chains are keyed by `AgentId`; one agent's repetition never trips another's reminder. A user prompt (`agent/prompt-submit`) resets the submitting agent's chain; agent disposal drops its state.
+- **In-memory only.** A session resumed from persistence starts with a fresh chain — the guard is a heuristic nudge, not a logged invariant, later reminders are the accepted cost.
+
+## Reminder delivery
+
+Reminders ride the post-execute decision's `additionalContext` (source `{kind: 'plugin', plugin: 'repeat-tool-guard'}`), never a `content` replacement: the `tool/result` event stays the tool's own output for audit. The loop buffers the context and appends it as a `context/message` after the step's tool results, which the session renders as the tagged synthetic-user envelope — so the reminder is model-visible, source-attributed, and reconstructable from the session log with no new session event. The guard always delegates via `next()` and folds its reminder onto the downstream decision (both variants — a blocked call still gets the nudge); when a downstream listener attached its own `additionalContext`, the fold concatenates content and carries the guard's `source` (a `HookContext` holds one `MessageSource`; `source.kind` is what framing depends on).
+
+## Testing
+
+Unit suites drive a real agent loop against a mock adapter (no network) and cover the chain semantics above to per-file 100%. The snapshot tier owns the transcript surface: a scripted-replay scenario repeats a call five times and pins both reminder tiers (gentle at 3, detailed at 5) as `context/message`s in the ACP transcript.

packages/guard/repeat-tool-guard/package.json

@@ -0,0 +1,41 @@
+{
+  "name": "@deepseek-ai/dsh-repeat-tool-guard",
+  "description": "Repeat-tool-call guard plugin: advisory reminders when an agent loops on identical tool calls",
+  "version": "0.0.1",
+  "private": true,
+  "type": "module",
+  "main": "lib/index.js",
+  "types": "lib/types/index.d.ts",
+  "exports": {
+    ".": {
+      "types": "./lib/types/index.d.ts",
+      "default": "./lib/index.js"
+    },
+    "./src/*": "./src/*",
+    "./package.json": "./package.json"
+  },
+  "files": [
+    "lib/index.js",
+    "lib/types/**/*.d.ts",
+    "lib/types/**/*.d.ts.map",
+    "src"
+  ],
+  "license": "BSD-3-Clause",
+  "dependencies": {
+    "schemastery": "^3.18.0"
+  },
+  "peerDependencies": {
+    "@deepseek-ai/dsh-agent": "^0.0.1",
+    "@deepseek-ai/dsh-tools": "^0.0.1",
+    "cordis": "^4.0.0-rc.6"
+  },
+  "devDependencies": {
+    "@deepseek-ai/dsh-agent": "workspace:^",
+    "@deepseek-ai/dsh-agent-loop": "workspace:^",
+    "@deepseek-ai/dsh-llm": "workspace:^",
+    "@deepseek-ai/dsh-session": "workspace:^",
+    "@deepseek-ai/dsh-system-prompt": "workspace:^",
+    "@deepseek-ai/dsh-tools": "workspace:^",
+    "cordis": "^4.0.0-rc.6"
+  }
+}

packages/guard/repeat-tool-guard/src/index.ts

@@ -0,0 +1,268 @@
+/**
+ * Repeat-tool-call guard: advisory loop-breaker for agents stuck re-issuing
+ * the same tool call with identical arguments.
+ *
+ * Not a model-facing tool — it registers no tool, never vetoes or rewrites a
+ * call, and adds exactly one behavior: watch each agent's stream of tool calls
+ * through the `tools/post-execute` waterfall, count runs of consecutive calls
+ * to the same tool with identical canonicalized arguments, and at configured
+ * run lengths fold an escalating advisory reminder onto the decision's
+ * `additionalContext`. The loop appends that context as a logged
+ * `context/message` after the step's tool results, so the reminder is
+ * model-visible, source-attributed, and reconstructable from the session log
+ * with no new session event. Decision record:
+ * docs/rfc/implemented/feature/2026-07-08-repeat-tool-guard.md.
+ *
+ * ```yaml
+ * - id: repeat-tool-guard
+ *   name: '@deepseek-ai/dsh-repeat-tool-guard'
+ *   config:
+ *     thresholds: [3, 5, 8]   # consecutive counts that trigger a reminder
+ *     include: []             # tool-name patterns to track; empty = all tools
+ *     exclude: [todo_write]   # tool-name patterns transparent to the chain
+ * ```
+ *
+ * Chain state is keyed per {@link AgentId} — the tool registry is a
+ * context-level singleton whose waterfalls interleave every agent's calls, so
+ * a shared counter would let one agent's repetition trip another's reminder.
+ * State is in-memory only: a session resumed from persistence starts with a
+ * fresh chain (the guard is a heuristic nudge, not a logged invariant).
+ *
+ * Plugin export shape: named exports, NO default. The cordis Loader's
+ * `unwrapExports` does `exports.default ?? exports`, so a stray default would
+ * collapse the module to the bare `apply` (see docs/postmortem/0001).
+ *
+ * @module @deepseek-ai/dsh-repeat-tool-guard
+ */
+
+import type { Context } from 'cordis'
+import z from 'schemastery'
+import type { AgentId, HookContext, PromptDecision } from '@deepseek-ai/dsh-agent'
+import type { MessageSource } from '@deepseek-ai/dsh-llm'
+import type { PostToolDecision, ToolExecution } from '@deepseek-ai/dsh-tools'
+
+export const name = 'repeat-tool-guard'
+
+/**
+ * Plugin config, validated by the same-named schemastery schema plus the
+ * load-time checks in `apply` (misconfiguration fails loud: an empty
+ * `thresholds` list, a non-integer, a value below 2, or a duplicate throws at
+ * plugin load, never a silent fall-back). `include`/`exclude` entries are
+ * `*`-wildcard predicates over tool names at call time, not references to
+ * registry entries — a pattern matching no currently registered tool is valid
+ * (`exclude: [mcp_*]` must stay legal in a deployment that loads no MCP tools).
+ */
+export interface Config {
+  /** Consecutive-repeat counts that trigger a reminder (default `[3, 5, 8]`). */
+  thresholds?: number[]
+  /** Tool-name patterns to track; empty means every tool is tracked. */
+  include?: string[]
+  /** Tool-name patterns transparent to the chain (neither count nor reset). */
+  exclude?: string[]
+  /**
+   * Maximum characters of canonical arguments quoted in the DETAILED reminder
+   * (default 500). Large payloads (a `write` body, a long command) would
+   * otherwise ride into the next request unbounded — precisely in a loop
+   * scenario; the cap bounds the reminder, never the detection (the chain key
+   * always compares the FULL canonical string).
+   */
+  argumentsPreviewChars?: number
+}
+
+export const Config: z<Config> = z.object({
+  thresholds: z.array(z.number()).default([3, 5, 8]),
+  include: z.array(z.string()).default([]),
+  exclude: z.array(z.string()).default([]),
+  argumentsPreviewChars: z.number().default(500),
+})
+
+/**
+ * The `{kind:'plugin'}` source stamped on every reminder this guard injects —
+ * the label is load-bearing (an unlabeled context would render as a user
+ * prompt in derived history).
+ */
+const PLUGIN_SOURCE: MessageSource = { kind: 'plugin', plugin: 'repeat-tool-guard' }
+
+/**
+ * The gentle first-threshold reminder. Keyed to `thresholds[0]`, not a literal
+ * count, so a custom first threshold keeps the gentle-then-detailed escalation.
+ */
+const GENTLE_REMINDER =
+  'You are repeating the exact same tool call with identical arguments. '
+  + 'Carefully analyze the previous result before calling again: if the task is '
+  + 'not complete, try a different approach or different arguments instead of '
+  + 'repeating the call.'
+
+/** The detailed later-threshold reminder naming the tool, the run length, and the canonical arguments. */
+function detailedReminder(toolName: string, count: number, canonicalArguments: string): string {
+  return 'Repeated tool call detected:\n'
+    + `- tool: ${toolName}\n`
+    + `- consecutive_calls: ${count}\n`
+    + `- arguments: ${canonicalArguments}\n`
+    + 'The repeated calls are not making progress. Do not call this tool with '
+    + 'these exact arguments again. Inspect the latest result and choose a '
+    + 'different action, different arguments, or finish the task if enough '
+    + 'evidence has been gathered.'
+}
+
+/**
+ * Deep key-sort of a parsed-JSON value so two argument objects that differ
+ * only in property order canonicalize identically. Arguments reach the guard
+ * as the loop's `JSON.parse` output (or its raw-string fallback for malformed
+ * argument JSON), so JSON's value domain is the whole input domain — no
+ * bigint, cycle, or `undefined` handling exists because no input path can
+ * produce them.
+ */
+function sortJsonValue(value: unknown): unknown {
+  if (Array.isArray(value)) return value.map(sortJsonValue)
+  if (value !== null && typeof value === 'object') {
+    const record = value as Record<string, unknown>
+    const sorted: Record<string, unknown> = {}
+    for (const key of Object.keys(record).sort()) {
+      sorted[key] = sortJsonValue(record[key])
+    }
+    return sorted
+  }
+  return value
+}
+
+/** Canonical string form of a call's arguments: deep key-sort, then stringify. */
+function canonicalize(argumentsValue: unknown): string {
+  return JSON.stringify(sortJsonValue(argumentsValue))
+}
+
+/** Compile one `*`-wildcard pattern to an anchored RegExp (every other regex metacharacter is matched literally). */
+function wildcardToRegExp(pattern: string): RegExp {
+  const escaped = pattern.replace(/[|\\{}()[\]^$+?.]/g, String.raw`\$&`)
+  return new RegExp(`^${escaped.replaceAll('*', '.*')}$`)
+}
+
+/**
+ * Head-truncate the canonical arguments for quoting in the detailed reminder,
+ * marking how much was omitted. Bounds only the model-visible text — the
+ * chain key always uses the full canonical string.
+ */
+function previewArguments(canonical: string, cap: number): string {
+  if (canonical.length <= cap) return canonical
+  return `${canonical.slice(0, cap)}… (+${canonical.length - cap} more chars)`
+}
+
+/**
+ * Validate `thresholds` per the fail-loud contract and return them sorted
+ * ascending (the escalation rule reads `thresholds[0]` as the gentle tier, so
+ * order is normalized here, once).
+ */
+function validateThresholds(values: number[]): number[] {
+  if (values.length === 0) {
+    throw new Error('repeat-tool-guard: `thresholds` must not be empty')
+  }
+  for (const value of values) {
+    if (!Number.isInteger(value) || value < 2) {
+      throw new Error(`repeat-tool-guard: invalid threshold ${value} — every threshold must be an integer >= 2`)
+    }
+  }
+  if (new Set(values).size !== values.length) {
+    throw new Error('repeat-tool-guard: `thresholds` must not contain duplicates')
+  }
+  return [...values].sort((a, b) => a - b)
+}
+
+/**
+ * Concatenate the guard's reminder context with a downstream listener's
+ * optional one so folding drops neither. The merged block carries the guard's
+ * `source` — a `HookContext` holds one `MessageSource` and the seam cannot
+ * represent mixed provenance; the rendered `context/message` only
+ * distinguishes by `source.kind`, so a downstream plugin's text is still
+ * correctly framed as plugin context.
+ */
+function concatContext(ours: HookContext, theirs: HookContext | undefined): HookContext {
+  if (!theirs) return ours
+  return { content: [...ours.content, ...theirs.content], source: ours.source }
+}
+
+/** One agent's consecutive-repeat chain: the last tracked call's identity key and its run length. */
+interface Chain {
+  key: string
+  count: number
+}
+
+/**
+ * Install the guard's listeners.
+ * @param ctx - plugin context; listeners are scoped to it and disposed with it.
+ * @param config - validated {@link Config}; `thresholds` is re-checked fail-loud here.
+ */
+export function apply(ctx: Context, config: Config): void {
+  // schemastery's .default() guarantees the fields are set after validation.
+  const thresholds = validateThresholds(config.thresholds as number[])
+  const thresholdSet = new Set(thresholds)
+  const includePatterns = (config.include as string[]).map(wildcardToRegExp)
+  const excludePatterns = (config.exclude as string[]).map(wildcardToRegExp)
+  const argumentsPreviewChars = config.argumentsPreviewChars as number
+  if (!Number.isInteger(argumentsPreviewChars) || argumentsPreviewChars < 1) {
+    throw new Error(`repeat-tool-guard: invalid argumentsPreviewChars ${argumentsPreviewChars} — must be an integer >= 1`)
+  }
+
+  const chains = new Map<AgentId, Chain>()
+
+  /** Whether a tool participates in the chain (untracked calls are transparent: they neither count nor reset). */
+  function tracked(toolName: string): boolean {
+    if (includePatterns.length > 0 && !includePatterns.some(pattern => pattern.test(toolName))) return false
+    return !excludePatterns.some(pattern => pattern.test(toolName))
+  }
+
+  /**
+   * Advance the calling agent's chain for one attempt and return the reminder
+   * to deliver, if this attempt's run length hits a configured threshold.
+   * Counting happens here — in post-execute — because denied calls also flow
+   * through this waterfall (`ToolRegistry.execute` routes a deny through the
+   * same pipeline), and a model hammering a denied call is exactly the loop
+   * worth breaking.
+   */
+  function observe(exec: ToolExecution): HookContext | undefined {
+    // A direct `ctx.tools.execute()` caller has no model to remind and no id
+    // to key on; only agent-loop calls participate.
+    if (!exec.agent) return undefined
+    if (!tracked(exec.name)) return undefined
+    const canonical = canonicalize(exec.arguments)
+    const key = JSON.stringify([exec.name, canonical])
+    const chain = chains.get(exec.agent.id)
+    const count = chain !== undefined && chain.key === key ? chain.count + 1 : 1
+    chains.set(exec.agent.id, { key, count })
+    if (!thresholdSet.has(count)) return undefined
+    const text = count === thresholds[0]
+      ? GENTLE_REMINDER
+      : detailedReminder(exec.name, count, previewArguments(canonical, argumentsPreviewChars))
+    return { content: [{ type: 'text', text }], source: PLUGIN_SOURCE }
+  }
+
+  // Observe-and-enrich, never veto: count first (state advances regardless of
+  // the downstream outcome), DELEGATE so a later listener can still block or
+  // replace, then fold the reminder onto whatever came back — additionalContext
+  // rides both decision variants, so a blocked call still gets the nudge.
+  ctx.on('tools/post-execute', async (exec, _result, next): Promise<PostToolDecision> => {
+    const reminder = observe(exec)
+    const downstream = await next()
+    if (!reminder) return downstream
+    if (downstream.kind === 'block') {
+      return { kind: 'block', feedback: downstream.feedback, additionalContext: concatContext(reminder, downstream.additionalContext) }
+    }
+    return {
+      kind: 'accept',
+      ...downstream.content !== undefined ? { content: downstream.content } : {},
+      additionalContext: concatContext(reminder, downstream.additionalContext),
+    }
+  })
+
+  // A user interjection changes the context; repetition across it is not a
+  // loop. Pure reset hook: always delegates (attaching nothing, vetoing
+  // nothing).
+  ctx.on('agent/prompt-submit', (agent, _content, _source, next): Promise<PromptDecision> => {
+    chains.delete(agent.id)
+    return next()
+  })
+
+  // Drop state when an agent goes away, bounding the map over harness lifetime.
+  ctx.on('agent/status', (agent, status) => {
+    if (status === 'disposed') chains.delete(agent.id)
+  })
+}

packages/guard/repeat-tool-guard/tests/repeat-tool-guard.spec.ts

@@ -0,0 +1,401 @@
+import { describe, expect, it } from 'vitest'
+import { Context } from 'cordis'
+import LlmService, { CallId } from '@deepseek-ai/dsh-llm'
+import SessionStore, { type SessionEvent } from '@deepseek-ai/dsh-session'
+import SystemPrompt from '@deepseek-ai/dsh-system-prompt'
+import ToolRegistry, { defineTool } from '@deepseek-ai/dsh-tools'
+import AgentRegistry, { AgentId } from '@deepseek-ai/dsh-agent'
+import AgentLoop, { type ReactLoopAgent } from '@deepseek-ai/dsh-agent-loop'
+import * as RepeatToolGuard from '@deepseek-ai/dsh-repeat-tool-guard'
+import type { Config } from '@deepseek-ai/dsh-repeat-tool-guard'
+import { MockAdapter, textResponse, toolCallResponse } from '../../../core/agent-loop/tests/mock-adapter.ts'
+
+/**
+ * Behavior suite for the repeat-tool-call guard: chain semantics (identical /
+ * different-tracked / untracked-transparent / per-agent / resets), threshold
+ * escalation incl. the `thresholds[0]` gentle-text rule, canonicalization,
+ * fold-onto-downstream-decision, and fail-loud config validation — all driven
+ * through a real agent loop against a scripted mock adapter (no network).
+ */
+
+/** Boot the core spine + the guard; the caller registers adapters and extra listeners. */
+async function harness(config: Config = {}): Promise<Context> {
+  const ctx = new Context()
+  await ctx.plugin(LlmService)
+  await ctx.plugin(SessionStore)
+  await ctx.plugin(SystemPrompt)
+  await ctx.plugin(ToolRegistry)
+  await ctx.plugin(AgentRegistry)
+  await ctx.plugin(AgentLoop, { agents: [] })
+  await ctx.plugin(RepeatToolGuard, config)
+  ctx.tools.register(defineTool({ name: 'probe', description: 'p', parameters: {}, async execute() { return [{ type: 'text', text: 'ok' }] } }))
+  ctx.tools.register(defineTool({ name: 'other', description: 'o', parameters: {}, async execute() { return [{ type: 'text', text: 'ok' }] } }))
+  return ctx
+}
+
+function waitForIdle(ctx: Context, agent: ReactLoopAgent): Promise<void> {
+  return new Promise((resolve) => { const d = ctx.on('agent/status', (s, st) => { if (s === agent && st === 'idle') { d(); resolve() } }) })
+}
+
+/** Every `context/message` in the agent's log, flattened to joined text + source for terse assertions. */
+function reminders(agent: ReactLoopAgent): { text: string; source: unknown }[] {
+  return [...agent.session.events]
+    .filter((e): e is SessionEvent<'context/message'> => e.type === 'context/message')
+    .map(e => ({
+      text: e.data.content.map(block => block.type === 'text' ? block.text : '').join('|'),
+      source: e.data.source,
+    }))
+}
+
+const GUARD_SOURCE = { kind: 'plugin', plugin: 'repeat-tool-guard' }
+
+describe('threshold escalation', () => {
+  it('reminds gently at the first default threshold (3) and in detail at the second (5)', async () => {
+    const ctx = await harness()
+    const adapter = new MockAdapter([
+      ...Array.from({ length: 5 }, (_, i) => toolCallResponse(`c${i}`, 'probe', { q: 'same' })),
+      textResponse('done'),
+    ])
+    ctx.llm.registerAdapter(['mock'], adapter)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+    agent.send([{ type: 'text', text: 'go' }])
+    await waitForIdle(ctx, agent)
+
+    const found = reminders(agent)
+    expect(found).toHaveLength(2)
+    expect(found[0]!.text).toContain('repeating the exact same tool call')
+    expect(found[0]!.source).toEqual(GUARD_SOURCE)
+    expect(found[1]!.text).toContain('consecutive_calls: 5')
+    expect(found[1]!.text).toContain('- tool: probe')
+    expect(found[1]!.text).toContain('{"q":"same"}')
+    expect(found[1]!.source).toEqual(GUARD_SOURCE)
+  })
+
+  it('keys the gentle text to thresholds[0], not the literal 3', async () => {
+    const ctx = await harness({ thresholds: [4, 2] }) // unsorted on purpose: normalized ascending
+    const adapter = new MockAdapter([
+      ...Array.from({ length: 4 }, (_, i) => toolCallResponse(`c${i}`, 'probe', {})),
+      textResponse('done'),
+    ])
+    ctx.llm.registerAdapter(['mock'], adapter)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+    agent.send([{ type: 'text', text: 'go' }])
+    await waitForIdle(ctx, agent)
+
+    const found = reminders(agent)
+    expect(found).toHaveLength(2)
+    expect(found[0]!.text).toContain('repeating the exact same tool call') // gentle at 2
+    expect(found[1]!.text).toContain('consecutive_calls: 4') // detailed at 4
+  })
+})
+
+describe('chain semantics', () => {
+  it('caps the detailed reminder arguments at argumentsPreviewChars (detection still keys on the full string)', async () => {
+    const ctx = await harness({ thresholds: [2, 3], argumentsPreviewChars: 24 })
+    const bigPayload = 'x'.repeat(400)
+    const adapter = new MockAdapter([
+      toolCallResponse('c1', 'probe', { body: bigPayload }),
+      toolCallResponse('c2', 'probe', { body: bigPayload }),
+      toolCallResponse('c3', 'probe', { body: bigPayload }),
+      textResponse('done'),
+    ])
+    ctx.llm.registerAdapter(['mock'], adapter)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+    agent.send([{ type: 'text', text: 'go' }])
+    await waitForIdle(ctx, agent)
+
+    const found = reminders(agent)
+    expect(found).toHaveLength(2) // gentle at 2, detailed at 3 — full-key matching survived the cap
+    const detailed = found[1]!.text
+    expect(detailed).toContain('- arguments: {"body":"xxxxxxxxxxxxxx') // 24-char head
+    expect(detailed).toContain('… (+387 more chars)')
+    expect(detailed).not.toContain(bigPayload)
+  })
+
+  it('a different tracked call resets the chain', async () => {
+    const ctx = await harness()
+    const adapter = new MockAdapter([
+      toolCallResponse('c1', 'probe', { q: 1 }),
+      toolCallResponse('c2', 'probe', { q: 1 }),
+      toolCallResponse('c3', 'other', {}), // tracked, different → reset
+      toolCallResponse('c4', 'probe', { q: 1 }),
+      toolCallResponse('c5', 'probe', { q: 1 }),
+      toolCallResponse('c6', 'probe', { q: 1 }), // 3rd consecutive AFTER the reset
+      textResponse('done'),
+    ])
+    ctx.llm.registerAdapter(['mock'], adapter)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+    agent.send([{ type: 'text', text: 'go' }])
+    await waitForIdle(ctx, agent)
+
+    expect(reminders(agent)).toHaveLength(1)
+  })
+
+  it('excluded calls are transparent: they neither count nor reset', async () => {
+    const ctx = await harness({ exclude: ['other'] })
+    const adapter = new MockAdapter([
+      toolCallResponse('c1', 'probe', { q: 1 }),
+      toolCallResponse('c2', 'other', {}), // excluded → invisible to the chain
+      toolCallResponse('c3', 'probe', { q: 1 }),
+      toolCallResponse('c4', 'other', {}),
+      toolCallResponse('c5', 'probe', { q: 1 }), // 3rd consecutive probe
+      textResponse('done'),
+    ])
+    ctx.llm.registerAdapter(['mock'], adapter)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+    agent.send([{ type: 'text', text: 'go' }])
+    await waitForIdle(ctx, agent)
+
+    const found = reminders(agent)
+    expect(found).toHaveLength(1)
+    expect(found[0]!.text).toContain('repeating the exact same tool call')
+  })
+
+  it('include patterns track only matching tools (wildcard star)', async () => {
+    const ctx = await harness({ include: ['pro*'] })
+    const adapter = new MockAdapter([
+      toolCallResponse('c1', 'other', {}),
+      toolCallResponse('c2', 'other', {}),
+      toolCallResponse('c3', 'other', {}), // 3 identical, but untracked
+      toolCallResponse('c4', 'probe', {}),
+      toolCallResponse('c5', 'probe', {}),
+      toolCallResponse('c6', 'probe', {}), // 3 identical, tracked
+      textResponse('done'),
+    ])
+    ctx.llm.registerAdapter(['mock'], adapter)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+    agent.send([{ type: 'text', text: 'go' }])
+    await waitForIdle(ctx, agent)
+
+    const found = reminders(agent)
+    expect(found).toHaveLength(1)
+    expect(found[0]!.text).toContain('repeating the exact same tool call')
+  })
+
+  it('escapes regex metacharacters in patterns (a dot matches only a literal dot)', async () => {
+    const ctx = await harness({ exclude: ['pr.be'] }) // would match 'probe' as a regex; must not as a wildcard
+    const adapter = new MockAdapter([
+      ...Array.from({ length: 3 }, (_, i) => toolCallResponse(`c${i}`, 'probe', {})),
+      textResponse('done'),
+    ])
+    ctx.llm.registerAdapter(['mock'], adapter)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+    agent.send([{ type: 'text', text: 'go' }])
+    await waitForIdle(ctx, agent)
+
+    expect(reminders(agent)).toHaveLength(1) // probe was NOT excluded
+  })
+
+  it('canonicalization ignores property order, deeply', async () => {
+    const ctx = await harness()
+    const adapter = new MockAdapter([
+      toolCallResponse('c1', 'probe', { a: 1, nested: { x: [1, 2], y: null } }),
+      toolCallResponse('c2', 'probe', { nested: { y: null, x: [1, 2] }, a: 1 }),
+      toolCallResponse('c3', 'probe', { a: 1, nested: { x: [1, 2], y: null } }),
+      textResponse('done'),
+    ])
+    ctx.llm.registerAdapter(['mock'], adapter)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+    agent.send([{ type: 'text', text: 'go' }])
+    await waitForIdle(ctx, agent)
+
+    expect(reminders(agent)).toHaveLength(1) // all three canonicalize identically
+  })
+
+  it('keys chains per agent: one agent repeating never trips another', async () => {
+    const ctx = await harness()
+    ctx.llm.registerAdapter(['mock-a'], new MockAdapter([
+      toolCallResponse('a1', 'probe', { q: 1 }),
+      toolCallResponse('a2', 'probe', { q: 1 }),
+      textResponse('done'),
+    ]))
+    ctx.llm.registerAdapter(['mock-b'], new MockAdapter([
+      toolCallResponse('b1', 'probe', { q: 1 }),
+      toolCallResponse('b2', 'probe', { q: 1 }),
+      toolCallResponse('b3', 'probe', { q: 1 }),
+      textResponse('done'),
+    ]))
+    const agentA = ctx.agentLoop.create(AgentId('a'), { model: 'mock-a' })
+    const agentB = ctx.agentLoop.create(AgentId('b'), { model: 'mock-b' })
+    agentA.send([{ type: 'text', text: 'go' }])
+    agentB.send([{ type: 'text', text: 'go' }])
+    await Promise.all([waitForIdle(ctx, agentA), waitForIdle(ctx, agentB)])
+
+    expect(reminders(agentA)).toHaveLength(0) // 2 repeats < 3, despite B's 3 in the same registry
+    expect(reminders(agentB)).toHaveLength(1)
+  })
+
+  it('a new user prompt resets the chain', async () => {
+    const ctx = await harness()
+    const adapter = new MockAdapter([
+      toolCallResponse('c1', 'probe', { q: 1 }),
+      toolCallResponse('c2', 'probe', { q: 1 }),
+      textResponse('turn one done'),
+      toolCallResponse('c3', 'probe', { q: 1 }), // without the reset this would be the 3rd
+      textResponse('turn two done'),
+    ])
+    ctx.llm.registerAdapter(['mock'], adapter)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+    agent.send([{ type: 'text', text: 'go' }])
+    await waitForIdle(ctx, agent)
+    agent.send([{ type: 'text', text: 'again' }])
+    await waitForIdle(ctx, agent)
+
+    expect(reminders(agent)).toHaveLength(0)
+  })
+
+  it('drops an agent chain on disposal', async () => {
+    const ctx = await harness({ thresholds: [2] })
+    ctx.llm.registerAdapter(['mock'], new MockAdapter([
+      toolCallResponse('c1', 'probe', { q: 1 }),
+      textResponse('done'),
+      toolCallResponse('c2', 'probe', { q: 1 }), // same id, fresh agent: count 1, not 2
+      textResponse('done'),
+    ]))
+    // Loop agents are torn down by disposing the scope that created them
+    // (the loop.spec pattern): a child plugin fiber owns `first`.
+    let first!: ReactLoopAgent
+    const fiber = await ctx.plugin(Object.assign((inner: Context) => {
+      first = inner.agentLoop.create(AgentId('reused'), { model: 'mock' })
+    }, { inject: ['agentLoop'] }))
+    first.send([{ type: 'text', text: 'go' }])
+    await waitForIdle(ctx, first)
+    await fiber.dispose()
+    await first.done
+
+    const second = ctx.agentLoop.create(AgentId('reused'), { model: 'mock' })
+    second.send([{ type: 'text', text: 'go' }])
+    await waitForIdle(ctx, second)
+
+    expect(reminders(second)).toHaveLength(0)
+  })
+
+  it('counts denied calls: hammering a denied tool still draws the reminder', async () => {
+    const ctx = await harness({ thresholds: [2] })
+    ctx.on('tools/pre-execute', async () => ({ kind: 'deny' as const, reason: 'sealed' }))
+    const adapter = new MockAdapter([
+      toolCallResponse('c1', 'probe', { q: 1 }),
+      toolCallResponse('c2', 'probe', { q: 1 }),
+      textResponse('done'),
+    ])
+    ctx.llm.registerAdapter(['mock'], adapter)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+    agent.send([{ type: 'text', text: 'go' }])
+    await waitForIdle(ctx, agent)
+
+    expect(reminders(agent)).toHaveLength(1)
+  })
+
+  it('ignores direct executes with no agent (they neither crash nor advance any chain)', async () => {
+    const ctx = await harness({ thresholds: [2] })
+    const direct = await ctx.tools.execute({ callId: CallId('d1'), name: 'probe', arguments: { q: 1 } })
+    expect(direct.isError).toBe(false)
+
+    ctx.llm.registerAdapter(['mock'], new MockAdapter([
+      toolCallResponse('c1', 'probe', { q: 1 }), // if the direct call had counted, this would be #2
+      textResponse('done'),
+    ]))
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+    agent.send([{ type: 'text', text: 'go' }])
+    await waitForIdle(ctx, agent)
+
+    expect(reminders(agent)).toHaveLength(0)
+  })
+})
+
+describe('fold onto the downstream decision', () => {
+  it('folds the reminder onto a downstream block and keeps its feedback', async () => {
+    const ctx = await harness({ thresholds: [2] })
+    ctx.on('tools/post-execute', async () => ({
+      kind: 'block' as const,
+      feedback: [{ type: 'text' as const, text: 'nope' }],
+      additionalContext: { content: [{ type: 'text' as const, text: 'downstream-ctx' }], source: { kind: 'plugin' as const, plugin: 'test' } },
+    }))
+    const adapter = new MockAdapter([
+      toolCallResponse('c1', 'probe', { q: 1 }),
+      toolCallResponse('c2', 'probe', { q: 1 }),
+      textResponse('done'),
+    ])
+    ctx.llm.registerAdapter(['mock'], adapter)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+    agent.send([{ type: 'text', text: 'go' }])
+    await waitForIdle(ctx, agent)
+
+    const found = reminders(agent)
+    expect(found).toHaveLength(2)
+    // Call 1: below threshold — the downstream context passes through untouched.
+    expect(found[0]!.text).toBe('downstream-ctx')
+    expect(found[0]!.source).toEqual({ kind: 'plugin', plugin: 'test' })
+    // Call 2: reminder folded in front, single merged context, the guard's source.
+    expect(found[1]!.text).toContain('repeating the exact same tool call')
+    expect(found[1]!.text).toContain('|downstream-ctx')
+    expect(found[1]!.source).toEqual(GUARD_SOURCE)
+    // The block's feedback reached the tool result unchanged.
+    const results = [...agent.session.events].filter((e): e is SessionEvent<'tool/result'> => e.type === 'tool/result')
+    expect(results.every(r => r.data.isError)).toBe(true)
+    expect(results[1]!.data.content).toEqual([{ type: 'text', text: 'nope' }])
+  })
+
+  it('preserves a downstream accept content replacement while folding', async () => {
+    const ctx = await harness({ thresholds: [2] })
+    ctx.on('tools/post-execute', async () => ({
+      kind: 'accept' as const,
+      content: [{ type: 'text' as const, text: 'replaced' }],
+    }))
+    const adapter = new MockAdapter([
+      toolCallResponse('c1', 'probe', { q: 1 }),
+      toolCallResponse('c2', 'probe', { q: 1 }),
+      textResponse('done'),
+    ])
+    ctx.llm.registerAdapter(['mock'], adapter)
+    const agent = ctx.agentLoop.create(AgentId('a1'), { model: 'mock' })
+    agent.send([{ type: 'text', text: 'go' }])
+    await waitForIdle(ctx, agent)
+
+    const found = reminders(agent)
+    expect(found).toHaveLength(1)
+    expect(found[0]!.text).toContain('repeating the exact same tool call')
+    const results = [...agent.session.events].filter((e): e is SessionEvent<'tool/result'> => e.type === 'tool/result')
+    expect(results[1]!.data.content).toEqual([{ type: 'text', text: 'replaced' }])
+  })
+})
+
+describe('config validation fails loud', () => {
+  async function spine(): Promise<Context> {
+    const ctx = new Context()
+    await ctx.plugin(LlmService)
+    await ctx.plugin(SessionStore)
+    await ctx.plugin(SystemPrompt)
+    await ctx.plugin(ToolRegistry)
+    await ctx.plugin(AgentRegistry)
+    await ctx.plugin(AgentLoop, { agents: [] })
+    return ctx
+  }
+
+  it('rejects an empty thresholds list', async () => {
+    const ctx = await spine()
+    await expect(ctx.plugin(RepeatToolGuard, { thresholds: [] })).rejects.toThrow(/must not be empty/)
+  })
+
+  it('rejects a threshold below 2', async () => {
+    const ctx = await spine()
+    await expect(ctx.plugin(RepeatToolGuard, { thresholds: [1, 3] })).rejects.toThrow(/integer >= 2/)
+  })
+
+  it('rejects a non-integer threshold', async () => {
+    const ctx = await spine()
+    await expect(ctx.plugin(RepeatToolGuard, { thresholds: [2.5] })).rejects.toThrow(/integer >= 2/)
+  })
+
+  it('rejects duplicate thresholds', async () => {
+    const ctx = await spine()
+    await expect(ctx.plugin(RepeatToolGuard, { thresholds: [3, 3] })).rejects.toThrow(/duplicates/)
+  })
+
+  it('rejects a non-positive or fractional argumentsPreviewChars', async () => {
+    const ctx = await spine()
+    await expect(ctx.plugin(RepeatToolGuard, { argumentsPreviewChars: 0 })).rejects.toThrow(/argumentsPreviewChars/)
+    const ctx2 = await spine()
+    await expect(ctx2.plugin(RepeatToolGuard, { argumentsPreviewChars: 12.5 })).rejects.toThrow(/argumentsPreviewChars/)
+  })
+})

packages/guard/repeat-tool-guard/tsconfig.json

@@ -0,0 +1,30 @@
+{
+  "extends": "../../../tsconfig.base.json",
+  "compilerOptions": {
+    "rootDir": "src",
+    "outDir": "lib/types"
+  },
+  "include": [
+    "src"
+  ],
+  "references": [
+    {
+      "path": "../../../vendor/cosmokit"
+    },
+    {
+      "path": "../../../vendor/cordis"
+    },
+    {
+      "path": "../../../vendor/schemastery"
+    },
+    {
+      "path": "../../core/tools"
+    },
+    {
+      "path": "../../core/agent"
+    },
+    {
+      "path": "../../llm/llm"
+    }
+  ]
+}

packages/hooks/README.md

@@ -4,7 +4,7 @@ The hooks subsystem lets users extend the agent at lifecycle points the way Clau
 
 | Package | Role | Shape |
 |---|---|---|
-| `hook-protocol/` | Shared wire-protocol core: matcher primitive, exit-code/stdout codec, `runHook` (via `ctx.bash`), most-restrictive merge, `hook/*` session events | library (no plugin) |
+| `hook-protocol/` | Shared wire-protocol core: matcher primitive, exit-code/stdout codec, `runHook` (via `ctx.bash`), most-restrictive merge, `hook/*` session events, detached-run quiescence | library (no plugin) |
 | `hooks-claude/` | Bridge for a Claude Code `hooks.json` / settings | plugin |
 | `hooks-codex/` | Bridge for a Codex `hooks.json` | plugin |
 

packages/hooks/hook-protocol/README.md

@@ -13,6 +13,7 @@ Why a shared lib at all: Codex deliberately reimplements a *subset* of the Claud
 | Decode output | `parseHookOutput(exit, stdout, stderr)` → neutral `HookOutput` | maps the neutral `HookOutput` onto a seam-specific typed Decision |
 | Merge N hooks | `mergeHookOutputs(outputs)` → most-restrictive `MergedHookOutcome` | — |
 | Durable record | `appendHookInvoked` / `appendHookResult` (`hook/*` session events; the result's `decision`/`stderrSummary` derive from the `HookOutput` here) | calls them around each invocation |
+| Detached-run quiescence | `createDetachedRuns()` — track fire-and-forget run chains; `drain()` aborts, then awaits them | passes `signal` to each detached `runHook`, registers `drain` as its effect disposer |
 
 ## Primitives
 
@@ -20,10 +21,11 @@ Why a shared lib at all: Codex deliberately reimplements a *subset* of the Claud
 - **`runHook(bash, hook, options, now)`** — serialize `options.payload` to the hook's stdin (with a trailing newline iff `options.trailingNewline`), merge `options.env` after the executor's credential scrub (the `dsh-bash` trusted-plugin surface), honor the hook's `timeoutSec` (else `options.defaultTimeoutMs` — the bridge owns the default, its config defaulting to the lib's `DEFAULT_HOOK_TIMEOUT_MS` 10-minute reference), and decode the result (threading `options.expectedEventName` to the codec). Never throws: an executor rejection (infra fault) becomes a `HookOutput` with `exitCode: undefined` (a non-blocking error). `now` is injected for testable durations.
 - **`parseHookOutput(exitCode, stdout, stderr, expectedEventName?)`** — the exit-code + structured-stdout codec. Exit `0` → parse JSON stdout (lenient: non-JSON is left for the bridge); exit `2` → blocking error, `stderr` is the block reason (surfaced as `decision: 'block'`); other → non-blocking error. `hookSpecificOutput.permissionDecision` (allow/deny/ask) overrides a legacy top-level `decision`; `additionalContext`/`updatedInput`/`systemMessage`/`continue`/`stopReason` are parsed too. The schemas key the `hookSpecificOutput` block by `hookEventName`, so passing `expectedEventName` (the firing event) DISCARDS a block whose `hookEventName` names a different event — or omits it entirely — its event-scoped fields don't take effect (a `PreToolUse` block on a `Stop` hook is malformed, and so is a discriminator-less block that would otherwise apply to any event), while the event-agnostic top-level fields still apply. Pure and total.
 - **`mergeHookOutputs(outputs)`** — fold the results of every hook that matched one point: permission precedence **deny > ask > allow**, halt sticky on the first `continue:false`, block reasons joined with `\n\n`, `additionalContext`/`systemMessages` accumulated in order.
+- **`createDetachedRuns()`** — quiescence tracking for the emit-shaped points, which run detached (no seam awaits them). The bridge tracks each run chain — the hook run PLUS its continuation — and registers `drain()` as its effect disposer: drain fires the tracker's abort `signal` (so a still-running hook process is killed via `runHook`, not awaited out to its timeout), then resolves once every tracked chain has settled. `fiber.dispose()` resolving therefore means no detached hook work is left to fire into a disposed context ([defensive patterns](../../../docs/defensive-patterns.md): dispose must reach quiescence).
 
 ## `hook/*` session events
 
-Declaration-merged into `SessionEventMap` (log-only, like `compact/*` — NOT a `SurfaceEventType`, no `surfaceOp`): `hook/invoked` (a hook command ran) and `hook/result` (its outcome, paired by `handlerId`, with `appendHookResult` owning the decision rule). Payloads and per-event JSDoc are in the generated [persistence log event catalog](../../../docs/persistence-catalog/log-events.md); `stderrSummary` is truncated to the record's `stderrSummaryMaxChars` (the bridge's config, reference default `DEFAULT_STDERR_SUMMARY_MAX_CHARS` = 500; omitted when empty).
+Declaration-merged into `SessionEventMap` (log-only, like `compact/*` — NOT a `SurfaceEventType`, no `surfaceOp`): `hook/invoked` (a hook command ran) and `hook/result` (its outcome, paired by `handlerId`, with `appendHookResult` owning the decision rule). Payloads and per-event JSDoc are in the generated [persistence log event catalog](../../../docs/persistence-catalog.md); `stderrSummary` is truncated to the record's `stderrSummaryMaxChars` (the bridge's config, reference default `DEFAULT_STDERR_SUMMARY_MAX_CHARS` = 500; omitted when empty).
 
 Like every event they must sit inside an open turn. The mid-turn points (`PreToolUse`/`PostToolUse`/`UserPromptSubmit`/`Stop`) fire inside the loop's open turn by construction; `SessionStart` gets no `hook/*` record (its injected `context/message` is the durable evidence) — see the hooks RFC.
 

packages/hooks/hook-protocol/src/codec.ts

@@ -75,6 +75,12 @@ function permissionDecisionOf(value: string | undefined): HookOutput['decision']
  * (`decision`/`reason`/`continue`/`stopReason`/`systemMessage`)
  * are unaffected. Omit `expectedEventName` (or pass a matching one) to apply the
  * block as-is — a caller that doesn't key by event opts out of the check.
+ *
+ * @param exitCode - the process exit code; `undefined` when the hook could not be spawned at all.
+ * @param stdout - the captured stdout stream; consulted for structured JSON only on a 0 exit.
+ * @param stderr - the captured stderr stream; becomes the blocking `reason` on exit 2.
+ * @param expectedEventName - the event the hook is firing for; omit to apply a `hookSpecificOutput` block as-is.
+ * @returns the dialect-neutral decoded outcome.
  */
 export function parseHookOutput(exitCode: number | undefined, stdout: string, stderr: string, expectedEventName?: string): HookOutput {
   const trimmedErr = stderr.trim()

packages/hooks/hook-protocol/src/detached.ts

@@ -0,0 +1,71 @@
+/**
+ * Quiescence tracking for a bridge's DETACHED hook runs. The waterfall-shaped
+ * hook points (`UserPromptSubmit`, `PreToolUse`, …) are awaited by their seams,
+ * but the emit-shaped points (`SessionStart`, `SubagentStart`, `SubagentStop`)
+ * run fire-and-forget: no seam awaits them, so without tracking a bridge's
+ * disposal could strand a live hook process and let a late continuation fire
+ * into a disposed context (docs/defensive-patterns.md: dispose must reach
+ * quiescence). A bridge creates one tracker in `apply()`, passes
+ * {@link DetachedRuns.signal} to each detached {@link runHook} call, wraps the
+ * full run chain (the hook run PLUS its `.then` continuation) in
+ * {@link DetachedRuns.track}, and registers {@link DetachedRuns.drain} as its
+ * disposer.
+ *
+ * @module @deepseek-ai/dsh-hook-protocol/detached
+ */
+
+/** In-flight registry for one bridge's detached hook runs; see the module doc for the wiring contract. */
+export interface DetachedRuns {
+  /**
+   * The abort signal every tracked run must hand to {@link runHook} (via its
+   * `signal` option). {@link drain} fires it so a still-running hook process is
+   * killed rather than awaited out to its timeout (default 10 minutes).
+   */
+  readonly signal: AbortSignal
+  /**
+   * Register one detached run until it settles. Pass the FULL chain — the hook
+   * run and its continuation/error handler — so {@link drain} waits for the
+   * side effects (an inject, a warn), not just the process exit. A rejected
+   * chain is absorbed here (settlement bookkeeping only), but rejection
+   * handling is still the caller's job: an untracked `.catch` is what turns a
+   * failure into a logged warning instead of silence.
+   * @param run - the detached run chain to hold until settled.
+   */
+  track(run: Promise<unknown>): void
+  /**
+   * Abort {@link signal}, then resolve once every tracked chain has settled —
+   * including chains tracked while the drain is in progress. The bridge
+   * registers this as its effect disposer; cordis awaits it, so
+   * `fiber.dispose()` resolving means the bridge's detached work is quiescent.
+   * A run tracked AFTER drain resolves is not awaited by anyone — by then the
+   * bridge's listeners are disposed, so nothing can start one.
+   * @returns resolves when all tracked runs have settled.
+   */
+  drain(): Promise<void>
+}
+
+/**
+ * Create a {@link DetachedRuns} tracker (one per bridge `apply()`); settled
+ * runs are pruned so a long-lived session does not accumulate them.
+ * @returns the tracker.
+ */
+export function createDetachedRuns(): DetachedRuns {
+  const inflight = new Set<Promise<unknown>>()
+  const controller = new AbortController()
+  return {
+    signal: controller.signal,
+    track(run: Promise<unknown>): void {
+      inflight.add(run)
+      const settled = (): void => { inflight.delete(run) }
+      void run.then(settled, settled)
+    },
+    async drain(): Promise<void> {
+      controller.abort(new Error('hook bridge disposed'))
+      // Re-check after each wave: a chain can be tracked while a prior wave is
+      // settling; loop until the registry is observed empty.
+      while (inflight.size > 0) {
+        await Promise.allSettled([...inflight])
+      }
+    },
+  }
+}

packages/hooks/hook-protocol/src/events.ts

@@ -66,6 +66,9 @@ export const DEFAULT_STDERR_SUMMARY_MAX_CHARS = 500
  * `undefined` when empty, cut at `maxChars` with an ellipsis when over. The
  * bound is a parameter — like `runHook`'s `defaultTimeoutMs`, each bridge owns
  * the config default and passes it in.
+ * @param stderr - the hook's raw captured stderr.
+ * @param maxChars - the character cap for the summary (the bridge's config value).
+ * @returns the trimmed, capped summary, or `undefined` when stderr is blank.
  */
 export function summarizeStderr(stderr: string, maxChars: number): string | undefined {
   const t = stderr.trim()
@@ -73,7 +76,11 @@ export function summarizeStderr(stderr: string, maxChars: number): string | unde
   return t.length > maxChars ? t.slice(0, maxChars) + '…' : t
 }
 
-/** Append a `hook/invoked` provenance event to `session`. */
+/**
+ * Append a `hook/invoked` provenance event to `session`.
+ * @param session - the session whose open turn records the event.
+ * @param invocation - the invocation identity; an absent `matcher` is omitted from the payload.
+ */
 export function appendHookInvoked(session: Session, invocation: HookInvocation): void {
   session.append('hook/invoked', {
     turn: invocation.turn,
@@ -91,6 +98,8 @@ export function appendHookInvoked(session: Session, invocation: HookInvocation):
  * else `'pass'`; `stderrSummary` is the trimmed stderr truncated to
  * `record.stderrSummaryMaxChars` characters (omitted when empty); `exitCode`
  * is omitted when the hook never ran.
+ * @param session - the session whose open turn records the event.
+ * @param record - the outcome to record: the decoded output plus the summary cap and duration.
  */
 export function appendHookResult(session: Session, record: HookResultRecord): void {
   const { output } = record

packages/hooks/hook-protocol/src/index.ts

@@ -15,6 +15,8 @@
  *   session-event helpers (declaration-merged into `SessionEventMap`);
  *   `appendHookResult` derives the durable `decision`/`stderrSummary` from the
  *   {@link HookOutput} so the shared event's semantics live in one place.
+ * - {@link createDetachedRuns} — quiescence tracking for the fire-and-forget
+ *   hook points: disposal aborts and drains a bridge's detached runs.
  *
  * Each bridge owns what genuinely DIFFERS: building the per-event stdin payload
  * (CC vs Codex field sets), the dialect's env/substitution, and mapping the
@@ -38,3 +40,5 @@ export { mergeHookOutputs } from './merge.ts'
 export type { MergedDecision, MergedHookOutcome } from './merge.ts'
 export { appendHookInvoked, appendHookResult, DEFAULT_STDERR_SUMMARY_MAX_CHARS, summarizeStderr } from './events.ts'
 export type { HookInvocation, HookResultRecord } from './events.ts'
+export { createDetachedRuns } from './detached.ts'
+export type { DetachedRuns } from './detached.ts'

packages/hooks/hook-protocol/src/matcher.ts

@@ -34,6 +34,10 @@ const CLAUDE_LITERAL = /^[A-Za-z0-9_|]+$/
  * pattern exact-matches the query (splitting `|` into alternatives); every other
  * `claude` pattern and ALL `codex` patterns are tested as an unanchored regex.
  * An invalid regex matches nothing (never throws).
+ * @param matcher - the configured pattern; absent/empty/`'*'` are the match-all sentinels.
+ * @param query - the candidate value (a tool name, a session source, …).
+ * @param mode - the dialect deciding literal-vs-regex interpretation of the pattern.
+ * @returns `true` when the pattern selects the query; `false` on a non-match or an invalid regex.
  */
 export function matchesMatcher(matcher: string | undefined, query: string, mode: MatcherMode): boolean {
   if (isMatchAll(matcher)) return true

packages/hooks/hook-protocol/src/merge.ts

@@ -71,6 +71,8 @@ function decisionForRank(maxRank: number): MergedDecision {
  * into one {@link MergedHookOutcome} by the precedence rules above. An empty list
  * yields a neutral outcome (`decision: 'none'`, no stop, empty context) — the
  * caller treats that as "no hook had anything to say".
+ * @param outputs - every matched hook's decoded output, in hook order.
+ * @returns the single folded outcome the bridge maps onto its seam.
  */
 export function mergeHookOutputs(outputs: HookOutput[]): MergedHookOutcome {
   let maxRank = 0

packages/hooks/hook-protocol/src/runner.ts

@@ -70,6 +70,11 @@ export interface RunHookResult {
  * `exitCode: undefined`, so the caller's merge logic treats it as a
  * non-blocking error rather than crashing the turn. `now` is injected for
  * testable durations.
+ * @param bash - the executor seam the command runs through.
+ * @param hook - the configured command; its `timeoutSec` (wire unit: seconds) overrides the default timeout.
+ * @param options - the invocation's payload, env, cwd, signal, stdin framing, and default timeout.
+ * @param now - millisecond clock used for the reported duration.
+ * @returns the decoded output plus the run's wall-clock duration.
  */
 export async function runHook(
   bash: BashExecutor,

packages/hooks/hook-protocol/tests/detached.spec.ts

@@ -0,0 +1,68 @@
+import { describe, expect, it } from 'vitest'
+import { createDetachedRuns } from '@deepseek-ai/dsh-hook-protocol'
+
+/** A promise settled from outside, so a test controls exactly when a tracked run finishes. */
+function deferred(): { promise: Promise<void>; resolve: () => void; reject: (error: Error) => void } {
+  let resolve!: () => void
+  let reject!: (error: Error) => void
+  const promise = new Promise<void>((res, rej) => { resolve = res; reject = rej })
+  return { promise, resolve, reject }
+}
+
+describe('createDetachedRuns', () => {
+  it('starts with an unfired signal; drain fires it (so still-running hook processes get killed)', async () => {
+    const detached = createDetachedRuns()
+    expect(detached.signal.aborted).toBe(false)
+    await detached.drain()
+    expect(detached.signal.aborted).toBe(true)
+    expect(String(detached.signal.reason)).toContain('hook bridge disposed')
+  })
+
+  it('drain with nothing tracked resolves immediately', async () => {
+    await expect(createDetachedRuns().drain()).resolves.toBeUndefined()
+  })
+
+  it('drain waits for a tracked run to settle', async () => {
+    const detached = createDetachedRuns()
+    const run = deferred()
+    detached.track(run.promise)
+    let drained = false
+    const draining = detached.drain().then(() => { drained = true })
+    // Give the drain every chance to (wrongly) resolve before the run settles.
+    await new Promise(resolve => setTimeout(resolve, 10))
+    expect(drained).toBe(false)
+    run.resolve()
+    await draining
+    expect(drained).toBe(true)
+  })
+
+  it('drain waits for a run tracked WHILE a prior wave was settling', async () => {
+    const detached = createDetachedRuns()
+    const first = deferred()
+    const second = deferred()
+    detached.track(first.promise)
+    // The late run enters the registry from the first run's own continuation —
+    // after drain() snapshotted its first wave.
+    void first.promise.then(() => { detached.track(second.promise) })
+    let drained = false
+    const draining = detached.drain().then(() => { drained = true })
+    first.resolve()
+    await new Promise(resolve => setTimeout(resolve, 10))
+    expect(drained).toBe(false)
+    second.resolve()
+    await draining
+    expect(drained).toBe(true)
+  })
+
+  it('a rejected tracked run is absorbed by the settlement bookkeeping (drain still resolves)', async () => {
+    const detached = createDetachedRuns()
+    const run = deferred()
+    detached.track(run.promise)
+    // The caller-side handler every bridge attaches; the tracker's own
+    // bookkeeping must not depend on it, but an UNHANDLED rejection would fail
+    // the test run, which is exactly the guarantee under test.
+    run.promise.catch(() => {})
+    run.reject(new Error('hook run boom'))
+    await expect(detached.drain()).resolves.toBeUndefined()
+  })
+})