Merge remote-tracking branch 'origin/master' into codex/basic-session-search
# Conflicts: # apps/web/tests/snapshots/question-composer/answered.expected.md # packages/client/connection/tests/fake-api.ts # packages/client/runtime/README.i18n.yaml # packages/client/runtime/tests/fake-api.ts # packages/client/ui-workspace/README.i18n.yaml # packages/client/ui-workspace/README.md # packages/client/ui-workspace/README.zh.md # packages/client/ui-workspace/src/client/WorkspaceBrowser.tsx # packages/client/ui-workspace/tests/apply.spec.ts # packages/host/apiproxy/README.i18n.yaml # packages/host/apiproxy/README.md # packages/host/apiproxy/README.zh.md # packages/host/apiproxy/src/api-proxy.ts # packages/host/apiproxy/src/api/index.ts # packages/host/apiproxy/src/api/sessions.schema.ts # packages/host/apiproxy/src/fetch/client.ts # packages/host/apiproxy/src/fetch/handler.ts # packages/host/apiproxy/tests/fetch-carrier.spec.ts # packages/host/apiproxy/tests/rpc-schemas.spec.ts # packages/session-query/session-query/tests/search-helpers.spec.ts
This commit is contained in:
Hypatia May
@@ -1,6 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-06-15-code-mode.md: 38d1ebdda089f1cfa1c5f3192fa2399b3a00102b
|
||||
2026-06-15-code-mode.zh.md: 2ddccb005b806fee0b1f4d6f79d6473f492a117d
|
||||
# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-06-15-code-mode.md
|
||||
2026-06-15-code-mode.md: b6a24ecd9700e32912b8112b59cbd8b6ab131eb5
|
||||
2026-06-15-code-mode.zh.md: 4d0a4cf8fa31cf9d9954e5bd95f823dfc0668444
|
||||
|
||||
@@ -130,4 +130,4 @@ Deployments switching to `'code'` must update any native-only `toolOrder`. Assem
|
||||
|
||||
**Serialized-only sub-dispatch.** `Promise.all` gains no wall-clock parallelism yet, only fewer round-trips; models may over-expect. The instructions state it; lifting it is tied to the same concurrency-safety metadata the native parallel-dispatch TODO needs.
|
||||
|
||||
**Budget metering reads the event loop, not a flag.** Busy-time polling (`eventLoopUtilization()`) is coarser than an exact CPU meter — a budget expires up to one poll interval late — and its correctness claim ("a pending dispatch cannot pause it") is load-bearing against a hostile program. Both sides are unit-tested (hot loop with a pending decoy dispatch dies at `computeMs`; idle-on-slow-binding survives to `maxWallMs`), and the poll interval is an internal constant, not config — nothing a deployment could mis-tune into a bypass.
|
||||
**Budget metering reads the event loop, not a flag.** Busy-time polling (`eventLoopUtilization()`) is coarser than an exact CPU meter — a budget expires up to one poll interval late — and its correctness claim ("a pending dispatch cannot pause it") is load-bearing against a hostile program. Both sides are unit-tested (hot loop with a pending decoy dispatch dies at `computeMs`; idle-on-slow-binding survives to `maxWallMs`), and the poll interval is an internal constant, not config — nothing a deployment could mis-tune into a bypass. `maxWallMs` is config, and it reaches `setTimeout`, which clamps a delay above `MAX_TIMER_DELAY_MS` (2^31-1 ms) to 1 ms; a positivity check alone therefore accepts a 25-day ceiling that expires on the first tick and times out every run. The worker runtime range-checks the field at load for that reason. `computeMs` needs no upper bound because it is compared against measured utilization instead of being handed to a timer.
|
||||
|
||||
@@ -130,4 +130,4 @@ SDK 指示模型编写一个异步的可擦除 TypeScript 函数体，通过 `aw
|
||||
|
||||
**仅序列化的子分发。** `Promise.all` 尚未获得挂钟并行性，仅减少往返次数；模型可能过度期望。说明中已声明；解除此限制与原生并行分发 TODO 所需的并发安全元数据绑定。
|
||||
|
||||
**预算计量读取事件循环，而非 flag。** 忙碌时间轮询（`eventLoopUtilization()`）比精确 CPU 计量更粗糙——预算到期最多延迟一个轮询间隔——且其正确性声明（「pending 的分发不能暂停它」）对恶意程序是承重的。两侧都有单元测试（带 pending 诱饵分发的热循环在 `computeMs` 处死亡；在慢绑定上空闲的程序存活到 `maxWallMs`），轮询间隔是内部常量而非配置——部署无法将其误调为绕过手段。
|
||||
**预算计量读取事件循环，而非 flag。** 忙碌时间轮询（`eventLoopUtilization()`）比精确 CPU 计量更粗糙——预算到期最多延迟一个轮询间隔——且其正确性声明（「pending 的分发不能暂停它」）对恶意程序是承重的。两侧都有单元测试（带 pending 诱饵分发的热循环在 `computeMs` 处死亡；在慢绑定上空闲的程序存活到 `maxWallMs`），轮询间隔是内部常量而非配置——部署无法将其误调为绕过手段。`maxWallMs` 是配置项，且会传入 `setTimeout`，后者会把超过 `MAX_TIMER_DELAY_MS`（2^31-1 ms）的延迟夹到 1 ms；因此仅有正数校验会放行一个 25 天的上限，它在第一个 tick 就到期，使每次运行都超时。worker 运行时正因如此在加载时对该字段做范围校验。`computeMs` 不需要上界，因为它对照的是实测占用率，而不是交给定时器。
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-06-18-compaction-capability-seam.md: a263b5e7d0245bd1279024a50e05b2f33edad521
|
||||
2026-06-18-compaction-capability-seam.zh.md: df0cf9d9131978e608d47124ba0f0db0343ee12a
|
||||
# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-06-18-compaction-capability-seam.md
|
||||
2026-06-18-compaction-capability-seam.md: 3c219b734e148b963fb5857de89c16f28c2bd402
|
||||
2026-06-18-compaction-capability-seam.zh.md: b2c7e9720b596705b60a284e6ccf1448a782b7fc
|
||||
|
||||
@@ -39,7 +39,7 @@ An earlier draft put the full algorithm (the retention walk, token-summing, text
|
||||
|
||||
Successful-call pressure cannot run at pre-step because final `agent/request` routing, provider output, tool results, buffered context, and steering do not exist there. Serial `agent/post-step(agent, turn, step, signal)` fires after those facts are durable and before `step/end`. `dsh-compact-basic` measures the canonical logged request through `ctx.tokenMeter`, so the next request sees any replacement without a speculative envelope override. Once pressure qualifies, optional `ctx.toolResultPrune` rewriting runs before summary selection; compact-basic remeasures the durable surface and skips summarization if pruning restores safe pressure.
|
||||
|
||||
Canonical provider context overflow takes a separate path. The failed step closes, `agent/request-error` receives the original request error and consecutive retry count, and compact-basic prunes before forcing one useful balanced reduction. It returns retry only if `session.surface.replaceGeneration` increases, including pruning-only progress when no summary range exists; the loop then opens a new numbered step and reconstructs its request from the durable log. No replacement, a recovery failure before any replacement, cancellation, an exhausted cap, or an unrelated error preserves the original provider failure. If pruning already advanced the generation before later summary work fails, recovery retries from that durable pruned surface unless cancellation or disposal wins. The complete lifecycle decision is in the [after-call recovery Agent Note](../architecture/2026-07-10-after-call-compaction-pressure-and-overflow-recovery.md).
|
||||
Canonical provider context overflow takes a separate path. The failed step closes and `agent/request-error` receives the original request error. Compact-basic owns its per-agent overflow count, prunes before forcing one useful balanced reduction, and returns `{ kind: 'retry' }` only if `session.surface.replaceGeneration` increases, including pruning-only progress when no summary range exists. The loop then closes the failed turn, opens a new numbered retry turn, and reconstructs its request from the durable log. No replacement, a recovery failure before any replacement, cancellation, an exhausted cap, or an unrelated error preserves the original provider failure. If pruning already advanced the generation before later summary work fails, recovery retries from that durable pruned surface unless cancellation or disposal wins. The complete lifecycle decision is in the [after-call recovery Agent Note](../architecture/2026-07-10-after-call-compaction-pressure-and-overflow-recovery.md).
|
||||
|
||||
```
|
||||
assistant/message → tool/result/context/steering
|
||||
|
||||
@@ -39,7 +39,7 @@ Status: implemented
|
||||
|
||||
成功调用的压力检查不能在步骤前运行，因为最终的 `agent/request` 路由、提供方输出、工具结果、缓冲上下文与 steering 当时尚不存在。串行的 `agent/post-step(agent, turn, step, signal)` 会在这些事实持久化后、`step/end` 之前触发。`dsh-compact-basic` 通过 `ctx.tokenMeter` 测量规范的已记录请求，因此下一个请求无需推测性覆盖信封即可看到任何替换。压力达到条件后，可选的 `ctx.toolResultPrune` 重写在摘要范围选择前运行；compact-basic 重新测量持久 surface，如果修剪恢复到安全压力便跳过摘要生成。
|
||||
|
||||
规范的提供方上下文溢出走另一条路径。失败步骤先关闭，`agent/request-error` 接收原始请求错误与连续重试次数，compact-basic 在强制执行一次有效且平衡的缩减前先修剪。仅当 `session.surface.replaceGeneration` 增加时，它才返回 retry；这包括没有摘要范围时仅修剪取得的进展。随后循环开启新的编号步骤，并从持久日志重建请求。没有替换、任何替换前的恢复失败、取消、耗尽的上限或无关错误都会保留原始提供方失败。如果修剪已经推进 generation，而后续摘要工作失败，恢复会从该持久的已修剪 surface 重试，除非取消或资源释放胜出。完整生命周期决策见[调用后恢复 Agent Note](../architecture/2026-07-10-after-call-compaction-pressure-and-overflow-recovery.md)。
|
||||
规范的提供方上下文溢出走另一条路径。失败步骤先关闭，`agent/request-error` 接收原始请求错误。compact-basic 自行持有按 agent 计的溢出次数，在强制执行一次有效且平衡的缩减前先修剪，且仅当 `session.surface.replaceGeneration` 增加时才返回 `{ kind: 'retry' }`；这包括没有摘要范围时仅修剪取得的进展。随后循环关闭失败轮次，开启新的编号重试轮次，并从持久日志重建请求。没有替换、任何替换前的恢复失败、取消、耗尽的上限或无关错误都会保留原始提供方失败。如果修剪已经推进 generation，而后续摘要工作失败，恢复会从该持久的已修剪 surface 重试，除非取消或资源释放胜出。完整生命周期决策见[调用后恢复 Agent Note](../architecture/2026-07-10-after-call-compaction-pressure-and-overflow-recovery.md)。
|
||||
|
||||
```
|
||||
assistant/message → tool/result/context/steering
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-06-24-workspace-context.md: 6acdb6241bcc57250e217cfc8856e8b0598d4622
|
||||
2026-06-24-workspace-context.zh.md: f165d08108931df21697c7895523f10ef816297f
|
||||
# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-06-24-workspace-context.md
|
||||
2026-06-24-workspace-context.md: f86e227be615c9b54e2a9013d3c7dca75d3975f0
|
||||
2026-06-24-workspace-context.zh.md: 154b5260955570e2de3c88d98286c5ea6afaa3b5
|
||||
|
||||
@@ -10,13 +10,13 @@ Repository guidance such as `AGENTS.md` belongs in a coding session's effective
|
||||
|
||||
Neighboring products establish useful conventions but differ in details. Codex treats `AGENTS.md` as native, Claude Code uses `CLAUDE.md` and familiar system-reminder-style user context, and opencode supports both names with one winner per directory plus lazy nested discovery. The harness needs cross-tool compatibility without loading duplicate or contradictory files from the same scope.
|
||||
|
||||
The lifecycle has two distinct classes of content. The initial applicable chain is stable enough to live in the request prefix and benefit from provider prefix caching. Nested files, edits, candidate switches, and removals happen after the session starts and belong in durable append-only history rather than the frozen prefix.
|
||||
The lifecycle has two distinct classes of content. The initial applicable chain is injected once before the first request. Nested files, edits, candidate switches, and removals happen later and join the same durable append-only history.
|
||||
|
||||
## Decision
|
||||
|
||||
The implementation lives in `packages/context/workspace-context` as `@deepseek-ai/dsh-workspace-context`. It is a request-context extension, not a core service or a filesystem backend. The shared demo spine and Host Runtime mount it from an explicit `{ maxBytes } | false` deployment choice; `dsh web` enables a 65,536-byte budget while the Host Runtime's headless consumer disables it. The plugin consumes `agent/session-prefix`, `tools/post-execute`, and the optional `ctx.fs` capability.
|
||||
The implementation lives in `packages/context/workspace-context` as `@deepseek-ai/dsh-workspace-context`. It is a request-context extension, not a core service or a filesystem backend. The shared demo spine and Host Runtime mount it from an explicit `{ maxBytes } | false` deployment choice; `dsh web` enables a 65,536-byte budget while the Host Runtime's headless consumer disables it. The plugin consumes `agent/step`, `tools/post-execute`, and the optional `ctx.fs` capability.
|
||||
|
||||
The plugin does not statically inject `fs`. Providerless product trees therefore boot normally and the plugin no-ops until a filesystem provider exists. All production reads go through that provider. Candidate probes resolve each path and stat the result, so a final-component symlink is followed to its target: a link to a regular file loads, while a missing path or a non-file target is a confirmed absence. Following repository-owned links across the trust boundary is a deliberate reversal of the original no-follow probe; the [instruction-symlink follow note](2026-07-21-follow-instruction-symlinks.md) owns that decision and its residual risk. The session-prefix signal and dynamic tool execution signal propagate through resolution, metadata probes, and streaming reads, so cancellation does not wait for an unrelated filesystem scan. A resolve or stat exception is classified as unavailable: it skips only that candidate and is never interpreted as the deletion of an already-loaded scope.
|
||||
The plugin does not statically inject `fs`. Providerless product trees therefore boot normally and the plugin no-ops until a filesystem provider exists. All production reads go through that provider. Candidate probes resolve each path and stat the result, so a final-component symlink is followed to its target: a link to a regular file loads, while a missing path or a non-file target is a confirmed absence. Following repository-owned links across the trust boundary is a deliberate reversal of the original no-follow probe; the [instruction-symlink follow note](2026-07-21-follow-instruction-symlinks.md) owns that decision and its residual risk. The step signal and dynamic tool execution signal propagate through resolution, metadata probes, and streaming reads, so cancellation does not wait for an unrelated filesystem scan. A resolve or stat exception is classified as unavailable: it skips only that candidate and is never interpreted as the deletion of an already-loaded scope.
|
||||
|
||||
### File Names And Precedence
|
||||
|
||||
@@ -26,13 +26,13 @@ Candidate entries are same-directory file names. Empty entries, `.`/`..`, and en
|
||||
|
||||
The user-global file is fixed at `$DSH_HOME/AGENTS.md`, is not affected by either candidate list, and has no local overlay. `$DSH_HOME` defaults to `~/.dsh`, matching the harness-level home role of `~/.codex` or `~/.claude` rather than introducing a plugin-specific home. Tilde expansion and the default live in `dsh-paths` so future harness features share the same convention.
|
||||
|
||||
### Baseline Prefix
|
||||
### Baseline Injection
|
||||
|
||||
On the first request of an agent-loop instance, the plugin contributes one user-role message through `agent/session-prefix`. It loads the user-global file first, then finds the project root by walking upward from `agent.session.header.cwd` to a configured root marker (default `.git`), then loads one candidate from each directory from the root to the cwd. A `.git` file and a `.git` directory are both valid markers, covering linked worktrees and submodules. Without a marker, the cwd itself is the root.
|
||||
At the first `agent/step` of an agent-loop instance, the plugin injects one sourced user-role message before the request is derived. It loads the user-global file first, then finds the project root by walking upward from `agent.session.header.cwd` to a configured root marker (default `.git`), then loads one candidate from each directory from the root to the cwd. A `.git` file and a `.git` directory are both valid markers, covering linked worktrees and submodules. Without a marker, the cwd itself is the root.
|
||||
|
||||
The plugin prepends its contribution before `await next()` returns, so session-prefix contributions appear in plugin registration order. In the product spine workspace instructions are registered before a skills catalog and therefore appear first. The loop deep-freezes the composed prefix, logs it in `EpochHeader.messagePrefix`, and reuses it verbatim for that instance. It is request state, not `Session.deriveMessages()` history.
|
||||
The injection becomes a durable `user/message` with a typed `workspace-instructions` source. Its `baseline: true` marker distinguishes the complete startup or resume baseline from later deltas, and its change list persists the included scopes and content digests. In the product spine workspace instructions are registered before the skills catalog, so their `agent/step` listener injects first. The loop drains both messages before deriving the first request.
|
||||
|
||||
A resumed agent creates a new loop instance and recomposes the baseline from current files, with the new prefix anchored by the resume request header. This permits current baseline content on resume without mutating a prefix already used by an earlier instance.
|
||||
A resumed agent creates a new loop instance and injects a baseline composed from current files before its first request. This permits current baseline content on resume without mutating an earlier history event. A resume and a hot plugin remount both face a log that may already hold a baseline; they are told apart by `agent/session-start`, which a startup or resume emits before the first step while a remount attaches to an already-live session and never sees it. A remount retains the existing baseline only when its typed event remains in the current visible surface, and still rebuilds scope and provider-version tracking from current files. If compaction has shadowed that event, the remount injects a current baseline. A resume always re-composes.
|
||||
|
||||
The baseline is a user-role `<system-reminder>` with `Instructions from: <path>` sections and explicit authority and precedence language. This familiar model-facing frame avoids a harness-specific XML vocabulary. Project paths are root-relative and the user-global path is `~/.dsh/AGENTS.md` for the default home or `$DSH_HOME/AGENTS.md` for a configured home. A literal `</system-reminder>` inside file content is escaped. The package README owns the exact current [prompt shape](../../../../packages/context/workspace-context/README.md#prompt-shape).
|
||||
|
||||
@@ -42,21 +42,21 @@ After a successful first-party `read`, `write`, or `edit` call, the `tools/post-
|
||||
|
||||
A content edit appends `Updated instructions from: <path>`, states that the new content replaces the previous content, and includes the complete current file. If precedence changes from one candidate to another, the message also names the previous path and says it no longer applies. If no candidate remains, the plugin appends `Instructions removed: <path>` and states that the previously loaded instructions no longer apply.
|
||||
|
||||
Dynamic messages carry their complete system-reminder framing in `content`, and every `context/message` reaches the model verbatim as a user-role message (there is no core wrapper to opt out of). `context/message.meta` carries opaque JSON state that is persisted but never rendered to the model.
|
||||
Baseline and dynamic messages carry their complete system-reminder framing in `content`, and every sourced `user/message` reaches the model verbatim (there is no core wrapper to opt out of). The typed `workspace-instructions` source carries persisted state that is never rendered to the model.
|
||||
|
||||
Shell commands are not discovery triggers. Local bash calls start fresh shells, and inferring reached paths from arbitrary command strings would require shell semantics the prompt plugin does not own.
|
||||
|
||||
### Duplicate Suppression And Change Detection
|
||||
|
||||
Every dynamic workspace context event stores versioned metadata with `{ action, scope, path, digest? }`, where `digest` is SHA-1 over the loaded content. The model-facing prompt has no HTML comments, hidden markers, or headings that are parsed back into state.
|
||||
Every workspace context event stores versioned metadata with `{ action, scope, path, digest? }`, where `digest` is SHA-1 over the loaded content. Baselines additionally carry `baseline: true`. The model-facing prompt has no HTML comments, hidden markers, or headings that are parsed back into state.
|
||||
|
||||
At reconciliation time the plugin scans plugin-owned `context/message` events and derives the latest state for each visible scope. A short per-session pending map begins only after the immutable top-level `tools/result` proves an `additionalContexts` entry survived every post-execute listener, then covers the interval before the loop appends that context to the log. Each entry records the open `{ turn, step }`: an equal durable `context/message` at or after its sequence boundary confirms and removes it, while a matching `step/end` arriving first means the loop discarded its context buffer, so the plugin removes both the pending entry and its version-cache fast path. A nested Code Mode result stages its changes under the parent's opaque execution token so repeated sub-dispatches in one run do not duplicate them; the parent result rolls that provisional state back and commits only contexts retained by outer policy.
|
||||
At reconciliation time the plugin scans workspace-sourced `user/message` events and derives the latest state for each visible scope. A short per-session pending map begins only after the immutable top-level `tools/result` proves an `additionalContexts` entry survived every post-execute listener, then covers the interval before the loop appends that context to the log. Each entry records the open `{ turn, step }`: an equal durable `user/message` at or after its sequence boundary confirms and removes it, while a matching `step/end` arriving first means the loop discarded its context buffer, so the plugin removes both the pending entry and its version-cache fast path. A nested Code Mode result stages its changes under the parent's opaque execution token so repeated sub-dispatches in one run do not duplicate them; the parent result rolls that provisional state back and commits only contexts retained by outer policy.
|
||||
|
||||
An unchanged path and digest is suppressed. A logged removal is a tombstone, so a reappearing candidate becomes a new `set`. Resume works from persisted metadata. If compaction removes an instruction event from the visible surface, that state no longer suppresses a later load, matching the fact that the model can no longer see it. Only changes actually included under the byte budget enter metadata or pending state, so an omitted file remains eligible on a later touch.
|
||||
|
||||
The frozen baseline keeps an in-memory path/digest map for comparison. A later successful filesystem touch appends baseline edits or removals as dynamic messages; it never rewrites the prefix. During resumed prefix composition the plugin also reconciles visible dynamic scopes, so nested changes made while the agent was offline can append an update before the first resumed request.
|
||||
The initial baseline's typed changes are comparison state only while its event remains in the visible session surface. A later successful filesystem touch re-adds an unchanged baseline scope after compaction, or appends baseline edits or removals as dynamic messages; it never rewrites the original event. The in-memory scope marker and provider-version cache only select and accelerate probes, so neither can suppress context the model no longer sees. During resumed baseline preparation the plugin also reconciles visible dynamic scopes, so nested changes made while the agent was offline can append an update before the first resumed request.
|
||||
|
||||
There is intentionally no watcher. Detection occurs at the next successful structured filesystem touch or resumed prefix composition. A provider failure produces no removal; absence is only accepted when all configured candidates in that scope were probed successfully.
|
||||
There is intentionally no watcher. Detection occurs at the next successful structured filesystem touch or resumed baseline preparation. A provider failure produces no removal; absence is only accepted when all configured candidates in that scope were probed successfully.
|
||||
|
||||
### Byte Budget And Bounded Reads
|
||||
|
||||
@@ -68,7 +68,7 @@ There is intentionally no watcher. Detection occurs at the next successful struc
|
||||
|
||||
**Use a global `ctx.systemPrompt.section()`.** Rejected because one Cordis context can host sessions with different cwd values, while repository-owned text is lower-authority context rather than top-authority provider system content.
|
||||
|
||||
**Inject the baseline on every `agent/pre-step`.** Rejected because repeated history injection wastes tokens, complicates duplicate state, and prevents a structurally stable provider prefix. Prefix composition gives a frozen, logged, per-instance baseline while dynamic append-only messages handle changes.
|
||||
**Inject the baseline on every `agent/step`.** Rejected because repeated history injection wastes tokens and complicates duplicate state. A per-mount session guard gives one visible baseline event while it remains on the surface; dynamic append-only messages handle changes and compaction re-arming.
|
||||
|
||||
**Load both `AGENTS.md` and `CLAUDE.md` in one directory.** Rejected because repositories in transition commonly duplicate guidance across both files. Ordered candidates make precedence explicit and configurable.
|
||||
|
||||
@@ -78,7 +78,7 @@ There is intentionally no watcher. Detection occurs at the next successful struc
|
||||
|
||||
## Consequences
|
||||
|
||||
Workspace guidance is isolated per session and shared by the demo front doors, Web Host, and every tool presentation mode. Initial instructions benefit from stable prefix caching, while nested and changed content remains durable and replayable. The generic session/agent context contract carries JSON metadata propagated through prompt-submit and post-tool `additionalContexts` arrays without flattening entries.
|
||||
Workspace guidance is isolated per session and shared by the demo front doors, Web Host, and every tool presentation mode. Initial, nested, and changed instructions are durable and replayable. The generic session/agent context contract carries typed source data through injected messages and post-tool `additionalContexts` arrays without flattening entries.
|
||||
|
||||
Repository text remains untrusted input. Lower-authority user-role framing, explicit precedence language, and delimiter escaping reduce risk but do not eliminate prompt injection. Following a candidate symlink to its target widens that surface to off-tree content, so the permission and sandbox layers that confine `ctx.fs` to trusted roots are the boundary that treats workspace files as data rather than authority (the [instruction-symlink follow note](2026-07-21-follow-instruction-symlinks.md) owns the residual risk).
|
||||
|
||||
|
||||
@@ -10,13 +10,13 @@ Status: implemented
|
||||
|
||||
相邻产品形成了值得借鉴的约定，但具体做法各不相同。Codex 原生使用 `AGENTS.md`；Claude Code 使用 `CLAUDE.md`，并采用熟悉的 system-reminder 风格用户上下文；opencode 同时支持这两个名称，每个目录只选一个胜出者，并延迟发现嵌套文件。harness 需要跨工具兼容，同时避免从同一作用域加载重复或互相矛盾的文件。
|
||||
|
||||
生命周期中有两类截然不同的内容。初始适用文件链足够稳定，可以放入请求前缀并受益于提供方前缀缓存。嵌套文件、编辑、候选项切换和移除都发生在会话启动后，应进入持久的仅追加历史，而不是被冻结的前缀。
|
||||
生命周期中有两类截然不同的内容。初始适用文件链在第一次请求前一次性注入。嵌套文件、编辑、候选项切换和移除发生在其后，进入同一份持久的仅追加历史。
|
||||
|
||||
## 决策
|
||||
|
||||
该实现在 `packages/context/workspace-context` 中，包（package）名为 `@deepseek-ai/dsh-workspace-context`。它是请求上下文扩展，不是核心服务或文件系统后端。共享 demo 主干与 Host Runtime 根据显式的 `{ maxBytes } | false` 部署选择挂载它；`dsh web` 启用 65,536 字节预算，Host Runtime 的 headless 消费方则禁用它。该插件使用 `agent/session-prefix`、`tools/post-execute` 和可选的 `ctx.fs` 功能。
|
||||
该实现在 `packages/context/workspace-context` 中，包（package）名为 `@deepseek-ai/dsh-workspace-context`。它是请求上下文扩展，不是核心服务或文件系统后端。共享 demo 主干与 Host Runtime 根据显式的 `{ maxBytes } | false` 部署选择挂载它；`dsh web` 启用 65,536 字节预算，Host Runtime 的 headless 消费方则禁用它。该插件使用 `agent/step`、`tools/post-execute` 和可选的 `ctx.fs` 功能。
|
||||
|
||||
插件不会静态注入 `fs`。因此，不带提供方的产品树仍能正常启动；在文件系统提供方出现之前，插件保持无操作。所有生产读取都通过该提供方完成。候选项探测会解析每个路径并对结果执行 stat，因此会跟随最终路径组件的符号链接至其目标：指向普通文件的链接会被加载，缺失路径或非文件目标则确认为不存在。允许仓库拥有的链接跨越信任边界，是对最初不跟随探测方式的刻意反转；[跟随指令符号链接记录](2026-07-21-follow-instruction-symlinks.md)负责说明该决策及其残余风险。会话前缀信号与动态工具执行信号会贯穿解析、元数据探测和流式读取，因此取消不会等待无关的文件系统扫描。解析或 stat 异常归类为不可用：它只跳过该候选项，绝不被解释为已经加载的作用域被删除。
|
||||
插件不会静态注入 `fs`。因此，不带提供方的产品树仍能正常启动；在文件系统提供方出现之前，插件保持无操作。所有生产读取都通过该提供方完成。候选项探测会解析每个路径并对结果执行 stat，因此会跟随最终路径组件的符号链接至其目标：指向普通文件的链接会被加载，缺失路径或非文件目标则确认为不存在。允许仓库拥有的链接跨越信任边界，是对最初不跟随探测方式的刻意反转；[跟随指令符号链接记录](2026-07-21-follow-instruction-symlinks.md)负责说明该决策及其残余风险。步骤信号与动态工具执行信号会贯穿解析、元数据探测和流式读取，因此取消不会等待无关的文件系统扫描。解析或 stat 异常归类为不可用：它只跳过该候选项，绝不被解释为已经加载的作用域被删除。
|
||||
|
||||
### 文件名与优先级
|
||||
|
||||
@@ -26,13 +26,13 @@ Status: implemented
|
||||
|
||||
用户全局文件固定为 `$DSH_HOME/AGENTS.md`，不受任一候选列表影响，也没有本地覆盖层。`$DSH_HOME` 默认为 `~/.dsh`，与 `~/.codex` 或 `~/.claude` 在 harness 层的 home 角色一致，而不会引入插件专用 home。波浪号展开与默认值位于 `dsh-paths` 中，以便未来的 harness 功能共享同一约定。
|
||||
|
||||
### 基线前缀
|
||||
### 基线注入
|
||||
|
||||
agent loop（智能体循环）实例的第一次请求会让插件通过 `agent/session-prefix` 提供一条 user 角色消息。它先加载用户全局文件，再从 `agent.session.header.cwd` 向上遍历至配置的根标记（默认为 `.git`）以确定项目根目录，随后从根目录至 cwd 的每级目录各加载一个候选项。`.git` 文件与 `.git` 目录都是有效标记，因而能覆盖链接 worktree 和 submodule。找不到标记时，cwd 本身就是根目录。
|
||||
在 agent loop（智能体循环）实例的第一个 `agent/step`，插件会在派生请求前注入一条带来源的 user 角色消息。它先加载用户全局文件，再从 `agent.session.header.cwd` 向上遍历至配置的根标记（默认为 `.git`）以确定项目根目录，随后从根目录至 cwd 的每级目录各加载一个候选项。`.git` 文件与 `.git` 目录都是有效标记，因而能覆盖链接 worktree 和 submodule。找不到标记时，cwd 本身就是根目录。
|
||||
|
||||
插件会在 `await next()` 返回前前置其贡献，因此会话前缀贡献按插件注册顺序出现。在产品主干中，工作区指令的注册先于 skill 目录，所以它排在前面。循环会深度冻结组合后的前缀，将其记录在 `EpochHeader.messagePrefix` 中，并在该实例内逐字复用。它是请求状态，不是 `Session.deriveMessages()` 历史。
|
||||
该注入成为一条持久 `user/message`，并携带带类型的 `workspace-instructions` 来源。其 `baseline: true` 标记将完整的启动或恢复基线与后续增量区分开来，变更列表则持久保存已纳入的作用域和内容 digest。在产品主干中，工作区指令的注册先于 skill 目录，所以其 `agent/step` 监听器先注入。循环会在派生第一次请求前 drain 这两条消息。
|
||||
|
||||
恢复 agent 会创建新的循环实例，并使用当前文件重新组合基线；新的前缀由恢复请求 header 锚定。这样，恢复时可以使用当前基线内容，而无需修改先前实例已经使用过的前缀。
|
||||
恢复 agent 会创建新的循环实例，并在其第一次请求前注入由当前文件组合的基线。这样，恢复时可以使用当前基线内容，而无需修改先前的历史事件。恢复与插件热重挂都会面对日志中可能已存在基线的情况；二者通过 `agent/session-start` 区分：启动或恢复会在第一步前发出该事件，而热重挂附着到一个已存活的会话、永远不会看到它。只有当基线的类型化事件仍在当前可见表层中时，热重挂才保留既有基线，同时仍会根据当前文件重建 scope 与提供方版本跟踪。如果压缩（compaction）已遮蔽该事件，热重挂会注入当前基线。恢复则始终重新组合。
|
||||
|
||||
基线是一条 user 角色的 `<system-reminder>`，包含 `Instructions from: <path>` 章节，以及明确的权威性与优先级说明。这种熟悉的模型可见框架避免引入 harness 专用的 XML 词汇。项目路径相对于根目录；使用默认 home 时，用户全局路径为 `~/.dsh/AGENTS.md`，使用已配置 home 时则为 `$DSH_HOME/AGENTS.md`。文件内容中的字面量 `</system-reminder>` 会被转义。包 README 负责规定当前准确的[提示词形态](../../../../packages/context/workspace-context/README.md#prompt-shape)。
|
||||
|
||||
@@ -42,21 +42,21 @@ agent loop（智能体循环）实例的第一次请求会让插件通过 `agent
|
||||
|
||||
内容编辑会追加 `Updated instructions from: <path>`，说明新内容取代先前内容，并包含当前的完整文件。如果优先级从一个候选项变为另一个，消息还会指出先前路径并说明它不再适用。如果没有候选项保留，插件会追加 `Instructions removed: <path>`，并说明先前加载的指令不再适用。
|
||||
|
||||
动态消息在 `content` 中携带完整的 system-reminder 框架；每个 `context/message` 都作为 user 角色消息逐字抵达模型，核心层不会再添加可选择退出的包装。`context/message.meta` 携带不透明 JSON 状态，该状态会持久化，但绝不会渲染给模型。
|
||||
基线消息和动态消息都在 `content` 中携带完整的 system-reminder 框架；每条带来源的 `user/message` 都逐字抵达模型，核心层不会再添加可选择退出的包装。带类型的 `workspace-instructions` 来源携带持久化状态，该状态绝不会渲染给模型。
|
||||
|
||||
shell 命令不会触发发现。本地 bash 调用会启动全新的 shell，而从任意命令字符串推断已到达路径，需要实现提示词插件并不拥有的 shell 语义。
|
||||
|
||||
### 重复抑制与变更检测
|
||||
|
||||
每个动态工作区上下文事件都会存储带版本的元数据，其形态为 `{ action, scope, path, digest? }`；`digest` 是对已加载内容计算的 SHA-1。模型可见提示词中没有 HTML 注释、隐藏标记，也没有会被解析回状态的标题。
|
||||
每个工作区上下文事件都会存储带版本的元数据，其形态为 `{ action, scope, path, digest? }`；`digest` 是对已加载内容计算的 SHA-1。基线还会额外携带 `baseline: true`。模型可见提示词中没有 HTML 注释、隐藏标记，也没有会被解析回状态的标题。
|
||||
|
||||
协调时，插件扫描自身拥有的 `context/message` 事件，并派生每个可见作用域的最新状态。一个简短的逐会话待处理映射只会在不可变的顶层 `tools/result` 证明某个 `additionalContexts` 条目经过所有 post-execute 监听器后仍然保留时开始记录；随后，它覆盖循环将该上下文追加到日志之前的间隔。每个条目记录开启状态的 `{ turn, step }`：如果相同的持久 `context/message` 出现在其序列边界或之后，该条目得到确认并被移除；如果匹配的 `step/end` 先到达，则说明循环丢弃了上下文缓冲区，插件会同时移除待处理条目及其版本缓存快速路径。嵌套的 Code Mode 结果会把变更暂存在父级的不透明执行 token 下，确保一次运行中的重复子分发不会产生重复项；父级结果会回滚这份临时状态，并且只提交外层策略保留的上下文。
|
||||
协调时，插件扫描带工作区来源的 `user/message` 事件，并派生每个可见作用域的最新状态。一个简短的逐会话待处理映射只会在不可变的顶层 `tools/result` 证明某个 `additionalContexts` 条目经过所有 post-execute 监听器后仍然保留时开始记录；随后，它覆盖循环将该上下文追加到日志之前的间隔。每个条目记录开启状态的 `{ turn, step }`：如果相同的持久 `user/message` 出现在其序列边界或之后，该条目得到确认并被移除；如果匹配的 `step/end` 先到达，则说明循环丢弃了上下文缓冲区，插件会同时移除待处理条目及其版本缓存快速路径。嵌套的 Code Mode 结果会把变更暂存在父级的不透明执行 token 下，确保一次运行中的重复子分发不会产生重复项；父级结果会回滚这份临时状态，并且只提交外层策略保留的上下文。
|
||||
|
||||
路径和 digest 均未变化时会被抑制。日志中的移除操作是一条墓碑记录，因此重新出现的候选项会成为新的 `set`。恢复操作从持久化元数据继续工作。如果压缩（compaction）从可见表面移除某条指令事件，该状态不再抑制后续加载，这与模型已经无法看见它的事实一致。只有真正纳入字节预算的变更才会进入元数据或待处理状态，因此被省略的文件在之后的触碰中仍有资格加载。
|
||||
路径和 digest 均未变化时会被抑制。日志中的移除操作是一条墓碑记录，因此重新出现的候选项会成为新的 `set`。恢复操作从持久化元数据继续工作。如果压缩从可见表面移除某条指令事件，该状态不再抑制后续加载，这与模型已经无法看见它的事实一致。只有真正纳入字节预算的变更才会进入元数据或待处理状态，因此被省略的文件在之后的触碰中仍有资格加载。
|
||||
|
||||
被冻结的基线会保留一个内存中的 path／digest map 以供比较。后续成功的文件系统触碰会把基线编辑或移除操作追加为动态消息，绝不重写前缀。恢复时重新组合前缀的过程中，插件还会协调可见的动态作用域，因此 agent 离线期间发生的嵌套变更可以在第一次恢复请求前追加更新。
|
||||
只有当初始基线事件仍在可见会话表层中时，其类型化变更才用作比较状态。后续成功的文件系统触碰会在压缩后重新添加未变化的基线 scope，或把基线编辑或移除操作追加为动态消息；它绝不重写原始事件。内存中的 scope 标记和提供方版本 cache 只用于选择探测对象并加速探测，因此二者都不能抑制模型已无法看见的上下文。恢复时准备基线的过程中，插件还会协调可见的动态作用域，因此 agent 离线期间发生的嵌套变更可以在第一次恢复请求前追加更新。
|
||||
|
||||
系统刻意不使用文件监视器。检测发生在下一次成功的结构化文件系统触碰或恢复时的前缀组合。提供方失败不会产生移除；只有该作用域中的全部已配置候选项都成功完成探测后，系统才接受「不存在」这一结论。
|
||||
系统刻意不使用文件监视器。检测发生在下一次成功的结构化文件系统触碰或恢复时的基线准备。提供方失败不会产生移除；只有该作用域中的全部已配置候选项都成功完成探测后，系统才接受「不存在」这一结论。
|
||||
|
||||
### 字节预算与有界读取
|
||||
|
||||
@@ -68,7 +68,7 @@ shell 命令不会触发发现。本地 bash 调用会启动全新的 shell，
|
||||
|
||||
**使用全局 `ctx.systemPrompt.section()`。** 不予采纳，因为同一个 Cordis 上下文可以承载 cwd 不同的多个会话，而仓库所有的文本属于低权威用户上下文，不是最高权威的提供方系统内容。
|
||||
|
||||
**在每次 `agent/pre-step` 时注入基线。** 不予采纳，因为重复注入历史会浪费 token、使重复状态复杂化，并妨碍提供方前缀保持结构稳定。前缀组合提供冻结、已记录且逐实例的基线，动态仅追加消息则负责变更。
|
||||
**在每次 `agent/step` 时注入基线。** 不予采纳，因为重复注入历史会浪费 token，并使重复状态复杂化。逐挂载会话防护会在基线事件仍留在表面期间提供一条可见基线事件；动态仅追加消息负责处理变更和压缩后的重新启用。
|
||||
|
||||
**在一个目录中同时加载 `AGENTS.md` 和 `CLAUDE.md`。** 不予采纳，因为正在迁移的仓库通常会在两个文件中重复指引。按顺序排列的候选项让优先级显式且可配置。
|
||||
|
||||
@@ -78,7 +78,7 @@ shell 命令不会触发发现。本地 bash 调用会启动全新的 shell，
|
||||
|
||||
## 影响
|
||||
|
||||
工作区指引按会话隔离，并由 demo 前端、Web Host 与每一种工具展示模式共享。初始指令受益于稳定的前缀缓存，嵌套与变更内容则保持持久且可回放。通用的 session／agent 上下文契约通过 prompt-submit 与工具执行后的 `additionalContexts` 数组携带 JSON 元数据，而不会把条目展平。
|
||||
工作区指引按会话隔离，并由 demo 前端、Web Host 与每一种工具展示模式共享。初始、嵌套与变更指令都保持持久且可回放。通用的 session／agent 上下文契约通过注入消息与工具执行后的 `additionalContexts` 数组携带带类型的来源数据，而不会把条目展平。
|
||||
|
||||
仓库文本仍是不受信任的输入。低权威 user 角色框架、显式优先级说明和分隔符转义可以降低风险，但无法消除提示词注入。跟随候选符号链接到目标，会把该接口扩大至树外内容；因此，把 `ctx.fs` 限制在可信根目录内的权限与沙箱层才是真正的边界，它们让系统把工作区文件当作数据而不是权威（[跟随指令符号链接记录](2026-07-21-follow-instruction-symlinks.md)负责说明残余风险）。
|
||||
|
||||
|
||||
@@ -1,51 +0,0 @@
|
||||
# Agent Note: Ask-user question capability
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-06-25-ask-user-question.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
The agent sometimes cannot proceed safely from model inference alone: it needs the human to choose a path, confirm a risky/default action, or provide missing information. Before this change, the only way to get that answer was for the model to ask in assistant text and then stop, which broke the normal tool-call loop: the agent had no structured way to pause, no option metadata for UIs, no abort/error taxonomy, and no way for non-stdio front doors to present the question consistently.
|
||||
|
||||
This is a user-facing capability, but it also crosses package boundaries. A model-facing tool needs a provider-neutral request vocabulary; each UI surface needs to decide how to show and collect the answer; the agent loop should remain unchanged because a tool call already has the right async shape.
|
||||
|
||||
## Decision
|
||||
|
||||
Introduce `dsh-user-interaction` as the provider-neutral interface package for `ctx.userInteraction`, colocated with the model-facing consumer `dsh-tool-ask-user` under `packages/ui`. The grouping is intentional: asking a human is a UI-backed product affordance, not part of the providerless core spine. The seam still owns the stable request/answer/error vocabulary, while UI product surfaces provide the concrete provider that collects the answer. The tool registers `ask_user_question`, forwards `{ questions, agent, signal }`, and returns the provider-computed structured answers as the tool result.
|
||||
|
||||
The model-facing request vocabulary is deliberately aligned with the product-research schema: `ask_user_question({ questions: [{ id, question, header?, options?: [{ label, description? }], multi_select? }] })`. `id` is supplied per question and echoed in the result so a batch can be routed without relying on question text. `label` is both user-facing display text and the selected value returned to the model; there is no separate `value`, no `recommended`, no `allow_custom`, and no `desc` alias.
|
||||
|
||||
Providers return `{ answers: [{ id, selected, custom? }] }`. `selected` is always an array of selected option labels, so single-select and `multi_select` answers share one result shape. `custom` carries a free-text "Other" answer; optionless questions collect `custom` directly. When `custom` is present, it overrides any selected choices and `selected` is empty. A provider that supports partial completion represents a deliberately skipped item with the existing `{ id, selected: [] }` shape, preserving the other answers without extending the tool result vocabulary.
|
||||
|
||||
`UserInteractionError` extends `HarnessError`, so failures such as `NO_PROVIDER`, `ASK_ABORTED`, or missing request ownership survive `ctx.tools.execute()` as machine-routable `{ name, code }` tool errors. This matches the structured-error taxonomy and lets the model or a wrapping plugin distinguish "user cancelled" from a generic thrown exception.
|
||||
|
||||
## UI mappings
|
||||
|
||||
`dsh web` mounts `dsh-client-ui-question`, whose host half opts the Web product into the model-facing tool and whose browser half registers a `question` entry in the conversation-owned keyed composer slot. `createApiProxy` implements the Web provider with a process-memory pending table keyed by a host-minted rpcId. It registers the wait before broadcasting `question/requested`, replays the same id on every mux reopen, validates the session and complete answer batch before claiming it, and broadcasts `question/resolved` after answer, cancellation, abort, or disposal. Claiming deletes the entry synchronously, so the first valid response wins and duplicate or late responses return `not-pending`.
|
||||
|
||||
The Web composer shows one question at a time while retaining every request in the session object layer. It supports single-select, multi-select, optionless or explicit custom answers, description text, and a visual recommendation badge without selecting the recommendation automatically. Single-select choices advance to the next item immediately, and Enter submits when every item is answered or explicitly skipped; Enter during IME composition only confirms the input candidate. The footer skips only the current item and preserves earlier drafts; the close control rejects the whole tool call with `ASK_CANCELLED`. The normal composer returns only after the host's resolved frame removes the pending item.
|
||||
|
||||
`dsh-tui` renders each question as a keyboard overlay, shows option descriptions, supports single- and multi-select choices plus free-form custom answers, and rejects pending questions on abort, provider disposal, or terminal shutdown. Batched and simultaneous requests are queued so one overlay owns keyboard focus at a time.
|
||||
|
||||
An ACP elicitation mapping existed while the bridge was an editor UI; [ACP as an automation-only protocol](../simplification/2026-07-23-acp-automation-only-protocol.md) removed that third mapping.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
**Assistant text followed by a stopped turn.** The model could ask the user in plain assistant text and then stop. That loses the structured option metadata, gives UIs no provider-neutral way to render a choice, and forces the next human answer to arrive as a new user prompt rather than as the result of the operation that needed the answer.
|
||||
|
||||
**Core-owned ask-user packages.** The first implementation split the seam and the model-facing tool across `packages/core` and `packages/ui`, but both names describe one UI-backed human-interaction affordance. The seam remains provider-neutral, but it is not providerless core infrastructure like sessions, tools, or the agent registry. Keeping `dsh-user-interaction` and `dsh-tool-ask-user` together under `packages/ui` makes the package map match the product boundary: apps and bridges provide the human-answer provider, and the stdio app opts into the model-facing tool.
|
||||
|
||||
**Use a permission request for general questions.** Permission requests authorize tool execution; `ask_user_question` gathers information with optional free-form answers. Reusing the permission channel would collapse two different product concepts.
|
||||
|
||||
**A loop-level pause primitive.** The agent loop already knows how to await a tool call and resume from a tool result. Adding a new loop special case would duplicate that async shape and make every loop implementation learn about a UI concern.
|
||||
|
||||
## Consequences
|
||||
|
||||
The feature gives the model a powerful pause primitive, so prompt guidance matters. The tool description tells the model to ask concise questions and use options when possible. Product policy can later wrap `tools/execute` to restrict when the tool is allowed, but the loop should not special-case it.
|
||||
|
||||
`dsh-user-interaction` and `dsh-tool-ask-user` both live in `packages/ui` because they form one product-facing human-interaction capability. `agent-core` does not load either the tool or a provider. `dsh-tui-demo` opts into the seam, TUI provider, and model-facing tool. `dsh web` boots the seam/provider in the host runtime and exposes the tool through the selected Web question plugin. The ACP automation app mounts neither the seam nor the tool.
|
||||
|
||||
## Testing
|
||||
|
||||
Unit coverage pins provider registration/disposal, duplicate-provider rejection, abort-before-provider, empty-question rejection, structured tool errors through `ctx.tools.execute()`, batched answers, multi-select answers, custom answers, explicit per-item skips, and the model schema including the removal of `value`, `recommended`, `allow_custom`, and `desc`. TUI tests cover option descriptions, queued requests, shutdown/abort cleanup, optionless free-form input, invalid choices, duplicate multi-select selections, and batched question flows. Web tests pin stable-id replay, response validation, first-wins settlement, duplicate and late responses, whole-request cancellation versus owner abort, single-select advance, IME-safe Enter submission, per-item skip preservation, composer takeover, structured batch submission, and restoration of the normal composer.
|
||||
@@ -1,51 +0,0 @@
|
||||
# Agent Note: ask-user 提问能力
|
||||
|
||||
Status: implemented
|
||||
|
||||
[English](2026-06-25-ask-user-question.md) | 中文
|
||||
|
||||
## 问题
|
||||
|
||||
agent（智能体）有时仅凭模型推理（inference）无法安全地继续执行：它需要人类选择路径、确认有风险的或默认的操作，或者提供缺失的信息。在此变更之前，获取答案的唯一方式是模型在 assistant 文本中提问然后停止，这打断了正常的工具调用循环：agent 没有结构化的暂停方式，没有供 UI 使用的选项元数据，没有中止/错误分类体系，也没有让非 stdio 前端一致地呈现问题的途径。
|
||||
|
||||
这是一个面向用户的能力，但它也跨越了包（package）边界。面向模型的工具需要一套提供方无关的请求词汇；每个 UI 界面需要决定如何展示和收集答案；agent loop（智能体循环）应保持不变，因为工具调用本身已具备正确的异步形状。
|
||||
|
||||
## 决策
|
||||
|
||||
引入 `dsh-user-interaction` 作为 `ctx.userInteraction` 的提供方无关接口包，与面向模型的消费方 `dsh-tool-ask-user` 一同放在 `packages/ui` 下。这一分组是有意为之的：向人类提问是一种由 UI 支撑的产品功能，不属于无提供方的核心主干。seam 仍然拥有稳定的请求/应答/错误词汇，而 UI 产品界面提供收集答案的具体提供方。该工具注册 `ask_user_question`，转发 `{ questions, agent, signal }`，并将提供方计算出的结构化答案作为工具结果返回。
|
||||
|
||||
面向模型的请求词汇有意与产品调研 schema 对齐：`ask_user_question({ questions: [{ id, question, header?, options?: [{ label, description? }], multi_select? }] })`。`id` 按问题提供并在结果中回传，使批量请求无需依赖问题文本即可路由。`label` 既是面向用户的显示文本，也是返回给模型的选中值；没有单独的 `value`，没有 `recommended`，没有 `allow_custom`，也没有 `desc` 别名。
|
||||
|
||||
提供方返回 `{ answers: [{ id, selected, custom? }] }`。`selected` 始终是选中选项 label 的数组，因此单选和 `multi_select` 的答案共享同一种结果形状。`custom` 承载自由文本的「其他」答案；无选项的问题直接收集 `custom`。当 `custom` 存在时，它覆盖任何已选择的选项，`selected` 为空。支持部分完成的提供方使用现有的 `{ id, selected: [] }` 形状表示某项被有意跳过，在不扩展工具结果词汇的前提下保留其他答案。
|
||||
|
||||
`UserInteractionError` 继承 `HarnessError`，因此 `NO_PROVIDER`、`ASK_ABORTED` 或请求归属缺失等失败会以机器可路由的 `{ name, code }` 工具错误形式通过 `ctx.tools.execute()` 传出。这与结构化错误分类体系一致，使模型或包装插件能够区分「用户取消」与一般的抛出异常。
|
||||
|
||||
## UI 映射
|
||||
|
||||
`dsh web` 挂载 `dsh-client-ui-question`：其 host 侧使 Web 产品选择性加载面向模型的工具，浏览器侧则在 conversation 拥有的具名输入区 slot 中注册 `question` 项。`createApiProxy` 使用以 host 生成的 rpcId 为键的进程内 pending 表实现 Web 提供方。它先注册等待项，再广播 `question/requested`；每次 mux 重开时以相同 id 重放；在受理前校验会话和完整答案批次；并在回答、取消、中止或资源释放后广播 `question/resolved`。受理会同步删除该条目，因此首个有效响应胜出，重复或迟到的响应返回 `not-pending`。
|
||||
|
||||
Web 输入区一次显示一个问题，同时在会话对象层保留每个请求。它支持单选、多选、无选项问题或显式自定义答案、描述文本与可视化推荐标记，但不会自动选中推荐项。选择单选项后会立即进入下一项；当所有项都已回答或显式跳过时，按 Enter 提交；IME 组字期间按 Enter 只会确认输入候选项。页脚只跳过当前项并保留先前的草稿；关闭控件以 `ASK_CANCELLED` 拒绝整个工具调用。常规输入区只有在 host 的 resolved 帧移除待处理项后才会恢复。
|
||||
|
||||
`dsh-tui` 将每个问题渲染为键盘叠层，展示选项描述，支持单选、多选和自由格式自定义答案，并在中止、提供方 dispose（资源释放）或终端关闭时拒绝待处理的问题。批量请求和并发请求都会排队，确保同一时刻只有一个叠层占用键盘焦点。
|
||||
|
||||
在桥接层还是编辑器 UI 时曾存在一个 ACP（Agent Client Protocol）elicitation 映射；[ACP 作为仅面向自动化的协议](../simplification/2026-07-23-acp-automation-only-protocol.md)移除了这第三个映射。
|
||||
|
||||
## 曾考虑的替代方案
|
||||
|
||||
**Assistant 文本后跟一个停止的轮次。** 模型可以在纯 assistant 文本中向用户提问然后停止。这会丢失结构化选项元数据，UI 没有提供方无关的方式来渲染选择，且下一条人类回答只能作为新的 user 提示词到达，而非作为需要答案的那次操作的结果。
|
||||
|
||||
**核心拥有的 ask-user 包。** 最初实现将 seam 和面向模型的工具分别放在 `packages/core` 和 `packages/ui`，但两者描述的是同一个由 UI 支撑的人机交互功能。seam 仍然是提供方无关的，但它不是像会话、工具或 agent 注册表那样的无提供方核心基础设施。将 `dsh-user-interaction` 和 `dsh-tool-ask-user` 一起放在 `packages/ui` 下，使包的划分与产品边界一致：应用和 bridge 提供人类答案的提供方，stdio 应用选择性加载面向模型的工具。
|
||||
|
||||
**用权限请求处理通用提问。** 权限请求是对工具执行的授权；`ask_user_question` 是带可选自由格式答案的信息收集。复用权限通道会混淆两个不同的产品概念。
|
||||
|
||||
**循环级别的暂停原语。** agent loop 已经知道如何等待工具调用并从工具结果恢复。添加新的循环特殊分支会重复这一异步形状，并迫使每个循环实现都了解一个 UI 关注点。
|
||||
|
||||
## 后果
|
||||
|
||||
该功能赋予模型一个强大的暂停原语，因此提示词引导很重要。工具描述告诉模型：提问要简洁，尽可能使用选项。产品策略后续可以包装 `tools/execute` 来限制工具何时可用，但循环不应对其做特殊处理。
|
||||
|
||||
`dsh-user-interaction` 和 `dsh-tool-ask-user` 都位于 `packages/ui`，因为它们共同构成一个面向产品的人机交互能力。`agent-core` 不加载工具或提供方。`dsh-tui-demo` 选择性加载 seam、TUI 提供方和面向模型的工具。`dsh web` 在 host 运行时启动 seam/提供方，并通过选定的 Web question 插件暴露该工具。ACP 自动化应用既不挂载 seam 也不挂载该工具。
|
||||
|
||||
## 测试
|
||||
|
||||
单元覆盖率固定了以下场景：提供方注册/释放、重复提供方拒绝、提供方就绪前中止、空问题拒绝、通过 `ctx.tools.execute()` 传出的结构化工具错误、批量答案、多选答案、自定义答案、显式按项跳过，以及模型 schema（包括移除 `value`、`recommended`、`allow_custom` 和 `desc`）。TUI 测试覆盖选项描述、排队请求、关闭/中止清理、无选项自由格式输入、无效选择、重复多选和批量问题流。Web 测试固定稳定 id 重放、响应校验、首个响应胜出的结算、重复和迟到响应、整个请求的取消与拥有方中止的区别、单选后前进、IME 安全的 Enter 提交、按项跳过保留、输入区接管、结构化批量提交，以及常规输入区的恢复。
|
||||
@@ -1,6 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-06-30-hook-bridges.md: c207b6901155925215548676364e903f5de2f29b
|
||||
2026-06-30-hook-bridges.zh.md: d396279d7ed1991536da2cea39e2aec5e50960c2
|
||||
# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-06-30-hook-bridges.md
|
||||
2026-06-30-hook-bridges.md: 99c6b1941a10e198ec3028f5fe505dabfff9abbe
|
||||
2026-06-30-hook-bridges.zh.md: 11ed3a5d177661271b30f1a58d034caa577b5348
|
||||
|
||||
@@ -6,7 +6,7 @@ English | [中文](2026-06-30-hook-bridges.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
The harness's extension surface is its typed interception seams ([the interception-seams Agent Note](2026-06-30-interception-seams.md)): a "native hook" is just an ordinary cordis plugin subscribing to `agent/session-start`, `agent/prompt-submit`, `tools/pre-execute`, `tools/post-execute`, `agent/turn-continuation`, `subagent/start`, `subagent/end`. But users arrive with **existing** Claude Code (CC) and Codex hook configs — a `hooks.json` (or a settings file's `hooks` key) full of shell-command hooks — and want those to run unmodified. This Agent Note introduces the two **bridge plugins** that translate that external shell-hook protocol onto the typed seams, built on the shared wire-protocol library ([the hook-protocol-lib Agent Note](2026-06-30-hook-protocol-lib.md)).
|
||||
The harness's extension surface is its typed interception seams ([the interception-seams Agent Note](2026-06-30-interception-seams.md)): a "native hook" is just an ordinary cordis plugin subscribing to `agent/session-start`, `agent/prompt-submit`, `tools/pre-execute`, `tools/post-execute`, `agent/turn-stopping`, `subagent/start`, or `subagent/end`. But users arrive with **existing** Claude Code (CC) and Codex hook configs — a `hooks.json` (or a settings file's `hooks` key) full of shell-command hooks — and want those to run unmodified. This Agent Note introduces the two **bridge plugins** that translate that external shell-hook protocol onto the typed seams, built on the shared wire-protocol library ([the hook-protocol-lib Agent Note](2026-06-30-hook-protocol-lib.md)).
|
||||
|
||||
The framing that shapes the whole design: **a bridge is a compatibility adapter, not a power tool.** Anything a bridge does (block a tool, inject context, force continuation, observe a subagent) a native cordis plugin does more powerfully — typed returns, full `ctx`, no serialization boundary. The bridge's reason to exist is to run the explicitly supported subset of external CC/Codex command hooks. That keeps each bridge thin: parse the config, pick a matcher mode, build the per-event payload, call `runHook` + `mergeHookOutputs` from the shared lib, and map the neutral outcome onto a seam Decision. The package READMEs own the exact current unsupported-event and partial-field inventory against the official protocols.
|
||||
|
||||
@@ -27,7 +27,7 @@ Each bridge maps the neutral `MergedHookOutcome` from the shared lib onto the se
|
||||
| `agent/prompt-submit` | `deny`→`block`; context-only→delegate+fold | `block`→`block`; context-only→delegate+fold |
|
||||
| `tools/pre-execute` | `deny`→`deny`; `ask`→`ask` | `block`→`deny` (no allow/ask) |
|
||||
| `tools/post-execute` | `deny`→`block`+feedback; context-only→delegate+fold | same |
|
||||
| `agent/turn-continuation` | blocking Stop → `continue` (reason = next-step steering) | same |
|
||||
| `agent/turn-stopping` | blocking Stop → next-step steering | same |
|
||||
| `subagent/start` (emit) | additionalContext → inject into a live in-process child; a remote child has no local injection target | unsupported by this bridge |
|
||||
| `subagent/end` (emit) | observe-only | unsupported by this bridge |
|
||||
|
||||
@@ -35,7 +35,9 @@ The CC bridge's `ask` result is a real permission path, not a terminal bridge de
|
||||
|
||||
### Context source is always the plugin (the mislabel guard)
|
||||
|
||||
`agent.inject()` defaults a missing `MessageSource` to `{ kind: 'user' }`, so every bridge `inject()` and `HookContext` passes `{ kind: 'plugin', plugin: 'hooks-claude' | 'hooks-codex' }`. Unit coverage pins the resulting `context/message.source` as the plugin rather than the user.
|
||||
Every bridge `inject()` and additional-context input explicitly passes `{ kind: 'plugin', plugin: 'hooks-claude' | 'hooks-codex' }`. Unit coverage pins the resulting `user/message.source` as the plugin rather than the user.
|
||||
|
||||
`UserPromptSubmit` runs during admission, before any turn opens. It therefore writes no turn-scoped `hook/invoked` / `hook/result` pair: a block leaves no transcript, while allowed additional context is durably represented by its sourced `user/message`. The Codex payload still receives the candidate next `turn_id`; rejection does not consume that number.
|
||||
|
||||
### Adding context is not a veto — delegate, then prepend
|
||||
|
||||
@@ -57,13 +59,13 @@ Hooks run in the agent's session workspace, so relative paths target the user's
|
||||
|
||||
- **Tool-input rewrite.** A CC/Codex `updatedInput` is logged + warned, not honored — input rewrite is a deferred consistency-design problem ([the pre-tool-input-rewrite Agent Note](../../proposed/feature/2026-06-30-pre-tool-input-rewrite.md)), because the pre-execution args are read by `tool/call` audit + `assistant/message` history + tool presentation, so an honest rewrite is a design unit, not a field.
|
||||
- **Stop loop-guard** (`TODO(stop-loop-guard)`). Claude Code supplies `stop_hook_active` and overrides a hook after eight consecutive blocks; Codex supplies `stop_hook_active` but documents no equivalent cap. Both bridges always report `false`, so a Stop hook that unconditionally blocks force-continues every step — a hook author must self-limit until state tracking lands.
|
||||
- **Hook `continue:false` (hard halt).** A hook can ask to halt the whole run (CC/Codex `continue:false`); the shared merge folds it into `MergedHookOutcome.stop`/`stopReason`, but no bridge acts on it (`TODO(hook-continue-false)`) — the interception seams have no "hard-halt the agent" primitive yet (a Decision blocks/steers a single point, not the run). Deferred with the loop-guard work; the halt request is recorded in the `hook/result` log, and the hook keeps its per-point effect (decision/context) meanwhile.
|
||||
- **Hook `continue:false` (hard halt).** A hook can ask to halt the whole run (CC/Codex `continue:false`); the shared merge folds it into `MergedHookOutcome.stop`/`stopReason`, but no bridge acts on it (`TODO(hook-continue-false)`) — the interception seams have no "hard-halt the agent" primitive yet (a Decision blocks/steers a single point, not the run). Deferred with the loop-guard work; mid-turn requests record the halt in `hook/result`, and the hook keeps its per-point effect (decision/context) meanwhile.
|
||||
- **Config discovery.** The path is explicit in `cordis.yml` and process-level (see above); the full multi-layer CC/Codex precedence walk, per-session project-local discovery, and the trust/hash model are not reimplemented (`TODO(per-session-hook-config)`).
|
||||
- **Session-start / subagent-start context is best-effort (`TODO(session-start-gating)`).** Both hooks run detached from startup, so their context is injected when ready but may miss the first request or a short-lived child. Guaranteeing first-request delivery requires an awaited startup seam.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
**Concurrent per-point hook execution.** The reference engines run a point's matched hooks concurrently and fold the results. These bridges run them **serially** (`await` per hook inside the match loop) and fold with the same most-restrictive merge. Serial is deliberate: it keeps each hook's `hook/invoked`/`hook/result` pair adjacent and in a deterministic order in the session log, and the fold is order-independent for the decision (`deny > ask > allow`) so the outcome matches. The cost is latency (hook *N* waits for hook *N−1*) and that per-hook timeouts are not overlapped — acceptable for the hook counts real configs use; revisit if a config ever fans out enough for the wall-clock to matter.
|
||||
**Concurrent per-point hook execution.** The reference engines run a point's matched hooks concurrently and fold the results. These bridges run them **serially** (`await` per hook inside the match loop) and fold with the same most-restrictive merge. Serial is deliberate: for turn-scoped points it keeps each hook's `hook/invoked`/`hook/result` pair adjacent and in deterministic order, and the fold is order-independent for the decision (`deny > ask > allow`) so the outcome matches. The cost is latency (hook *N* waits for hook *N−1*) and that per-hook timeouts are not overlapped — acceptable for the hook counts real configs use; revisit if a config ever fans out enough for the wall-clock to matter.
|
||||
|
||||
## Consequences
|
||||
|
||||
|
||||
@@ -6,7 +6,7 @@ Status: implemented
|
||||
|
||||
## 问题
|
||||
|
||||
harness 的扩展面是其类型化的拦截 seam（见[拦截 seam Agent Note](2026-06-30-interception-seams.md)）：所谓「原生钩子」不过是一个普通的 Cordis 插件，订阅 `agent/session-start`、`agent/prompt-submit`、`tools/pre-execute`、`tools/post-execute`、`agent/turn-continuation`、`subagent/start`、`subagent/end`。但用户带着**既有的** Claude Code（CC）和 Codex 钩子配置到来，一个 `hooks.json`（或 settings 文件中的 `hooks` 键）里满是 shell 命令钩子，并希望它们原样运行。本 Agent Note 引入两个**桥接插件**，将外部 shell 钩子协议翻译到类型化 seam 上，构建于共享的协议格式（wire format）库之上（见 [hook-protocol-lib Agent Note](2026-06-30-hook-protocol-lib.md)）。
|
||||
harness 的扩展面是其类型化的拦截 seam（见[拦截 seam Agent Note](2026-06-30-interception-seams.md)）：所谓「原生钩子」不过是一个普通的 Cordis 插件，订阅 `agent/session-start`、`agent/prompt-submit`、`tools/pre-execute`、`tools/post-execute`、`agent/turn-stopping`、`subagent/start` 或 `subagent/end`。但用户带着**既有的** Claude Code（CC）和 Codex 钩子配置到来，一个 `hooks.json`（或 settings 文件中的 `hooks` 键）里满是 shell 命令钩子，并希望它们原样运行。本 Agent Note 引入两个**桥接插件**，将外部 shell 钩子协议翻译到类型化 seam 上，构建于共享的协议格式（wire format）库之上（见 [hook-protocol-lib Agent Note](2026-06-30-hook-protocol-lib.md)）。
|
||||
|
||||
贯穿整个设计的定位：**桥接是兼容性适配器，不是高级工具。** 桥接能做的事（阻止工具、注入上下文、强制继续、观察 subagent），原生 Cordis 插件都能做得更强——类型化返回值、完整 `ctx`、无序列化边界。桥接存在的理由是运行外部 CC/Codex 命令钩子中被明确支持的子集。这使每个桥接保持精简：解析配置、选择匹配模式、构建每事件的 payload、调用共享库的 `runHook` + `mergeHookOutputs`，再将中性结果映射为 seam Decision。各包的 README 维护着当前不支持的事件和部分字段的完整清单，以官方协议为参照。
|
||||
|
||||
@@ -27,7 +27,7 @@ harness 的扩展面是其类型化的拦截 seam（见[拦截 seam Agent Note](
|
||||
| `agent/prompt-submit` | `deny`→`block`；仅上下文→delegate+fold | `block`→`block`；仅上下文→delegate+fold |
|
||||
| `tools/pre-execute` | `deny`→`deny`；`ask`→`ask` | `block`→`deny`（无 allow/ask） |
|
||||
| `tools/post-execute` | `deny`→`block`+feedback；仅上下文→delegate+fold | 同上 |
|
||||
| `agent/turn-continuation` | 阻塞的 Stop → `continue`（reason = 下一步 steering（中途引导）） | 同上 |
|
||||
| `agent/turn-stopping` | 阻塞的 Stop → 下一步 steering（中途引导） | 同上 |
|
||||
| `subagent/start`（emit） | additionalContext → 注入到存活的进程内 subagent；远程 subagent 无本地注入目标 | 本桥接不支持 |
|
||||
| `subagent/end`（emit） | 仅观察 | 本桥接不支持 |
|
||||
|
||||
@@ -35,7 +35,9 @@ CC 桥接的 `ask` 结果是一条真正的权限路径，而非终态桥接决
|
||||
|
||||
### 上下文来源始终是插件（误标签防护）
|
||||
|
||||
`agent.inject()` 在缺少 `MessageSource` 时默认为 `{ kind: 'user' }`，因此每个桥接的 `inject()` 和 `HookContext` 都传入 `{ kind: 'plugin', plugin: 'hooks-claude' | 'hooks-codex' }`。单元测试覆盖率固定验证结果中的 `context/message.source` 为插件而非用户。
|
||||
每个桥接的 `inject()` 和 additional-context 输入都显式传入 `{ kind: 'plugin', plugin: 'hooks-claude' | 'hooks-codex' }`。单元测试覆盖率固定验证结果中的 `user/message.source` 为插件而非用户。
|
||||
|
||||
`UserPromptSubmit` 在准入阶段运行，早于任何轮次开启。因此它不写入任何轮次范围的 `hook/invoked` / `hook/result` 对：阻止不会留下 transcript（文本记录），而被允许的额外上下文由其带来源的 `user/message` 持久呈现。Codex payload 仍会收到候选的下一个 `turn_id`；拒绝不会消耗该编号。
|
||||
|
||||
### 添加上下文不是否决——先 delegate，再 prepend
|
||||
|
||||
@@ -57,13 +59,13 @@ Claude Code 始终导出 `CLAUDE_PROJECT_DIR`，常见的未修改钩子引用 `
|
||||
|
||||
- **工具输入重写。** CC/Codex 的 `updatedInput` 被记录日志并发出警告，但不予执行——输入重写是一个推迟的一致性设计问题（见 [pre-tool-input-rewrite Agent Note](../../proposed/feature/2026-06-30-pre-tool-input-rewrite.md)），因为 pre-execution 参数被 `tool/call` 审计、`assistant/message` 历史和工具展示共同读取，诚实的重写是一个设计单元，而非一个字段。
|
||||
- **Stop 循环防护**（`TODO(stop-loop-guard)`）。Claude Code 提供 `stop_hook_active` 并在连续八次阻塞后覆盖钩子；Codex 提供 `stop_hook_active` 但未记录等效上限。两个桥接始终报告 `false`，因此一个无条件阻塞的 Stop 钩子会在每一步强制继续——在状态追踪落地之前，钩子作者必须自行限制。
|
||||
- **钩子 `continue:false`（硬停止）。** 钩子可以请求终止整个运行（CC/Codex `continue:false`）；共享合并将其折叠为 `MergedHookOutcome.stop`/`stopReason`，但没有桥接对其采取行动（`TODO(hook-continue-false)`）——拦截 seam 尚无「硬停止 agent」原语（Decision 阻塞/引导的是单个点，而非整个运行）。与循环防护工作一同推迟；停止请求记录在 `hook/result` 日志中，钩子在此期间保留其逐点效果（决策/上下文）。
|
||||
- **钩子 `continue:false`（硬停止）。** 钩子可以请求终止整个运行（CC/Codex `continue:false`）；共享合并将其折叠为 `MergedHookOutcome.stop`/`stopReason`，但没有桥接对其采取行动（`TODO(hook-continue-false)`）——拦截 seam 尚无「硬停止 agent」原语（Decision 阻塞/引导的是单个点，而非整个运行）。与循环防护工作一同推迟；轮中请求会将停止请求记录在 `hook/result` 中，钩子在此期间保留其逐点效果（决策/上下文）。
|
||||
- **配置发现。** 路径在 `cordis.yml` 中显式指定且为进程级（见上文）；完整的多层 CC/Codex 优先级遍历、按会话的项目本地发现以及信任/hash 模型未被重新实现（`TODO(per-session-hook-config)`）。
|
||||
- **Session-start / subagent-start 上下文为尽力而为（`TODO(session-start-gating)`）。** 两个钩子以 detached 方式运行于启动过程之外，因此其上下文在就绪时注入，但可能错过首个请求或短命的 subagent。要保证首请求送达，需要一个 awaited 的启动 seam。
|
||||
|
||||
## 曾考虑的替代方案
|
||||
|
||||
**每点钩子并发执行。** 参考引擎对一个点匹配到的钩子并发运行并折叠结果。本桥接**串行**运行（匹配循环内每个钩子 `await`），并以相同的最严格合并策略折叠。串行是刻意的：它使每个钩子的 `hook/invoked`/`hook/result` 对在会话日志中相邻且顺序确定，而折叠对决策是顺序无关的（`deny > ask > allow`），因此结果一致。代价是延迟（钩子 *N* 等待钩子 *N−1*）以及每钩子超时不重叠——对真实配置中的钩子数量可以接受；如果某配置的扇出大到影响总耗时，再重新评估。
|
||||
**每点钩子并发执行。** 参考引擎对一个点匹配到的钩子并发运行并折叠结果。本桥接**串行**运行（匹配循环内每个钩子 `await`），并以相同的最严格合并策略折叠。串行是刻意的：对轮次范围的拦截点，它使每个钩子的 `hook/invoked`/`hook/result` 对相邻且顺序确定，而折叠对决策是顺序无关的（`deny > ask > allow`），因此结果一致。代价是延迟（钩子 *N* 等待钩子 *N−1*）以及每钩子超时不重叠——对真实配置中的钩子数量可以接受；如果某配置的扇出大到影响总耗时，再重新评估。
|
||||
|
||||
## 后果
|
||||
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-06-30-interception-seams.md: 81799f4c6e3e7a4c6b9605cd97f5728b99d11995
|
||||
2026-06-30-interception-seams.zh.md: 65ae16842c632641e7ac65908162f4784dc6e1e0
|
||||
# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-06-30-interception-seams.md
|
||||
2026-06-30-interception-seams.md: 4658983e1f098ecd199eecec4408e7c2f134cbf7
|
||||
2026-06-30-interception-seams.zh.md: 13b7c56829412773111fcf6d75cc717c51d49c7b
|
||||
|
||||
@@ -16,9 +16,9 @@ The canonical surface separates transformable policy, around-dispatch control, a
|
||||
|
||||
**Agent events** (`dsh-agent`):
|
||||
- `agent/session-start(agent, source)` — emit, once before turn 1, carrying a `SessionStartSource` (`startup` for a fresh/forked create, `resume` for a reloaded persisted session; `clear`/`compact` reserved). A pure notification — it CANNOT block startup (a deliberate gap: a bridge logs/injects, it does not gate startup). A listener seeds context via `agent.inject()`.
|
||||
- `agent/prompt-submit(agent, content, source, signal, next) → PromptDecision` — waterfall, fired for the turn's single claimed queued message before the `user/message` append. The explicit turn signal is placed before the final `next`; `allow` optionally rewrites the prompt `content` or attaches separately sourced `additionalContexts[]`, while `block` appends a durable `prompt/blocked` and rejects that zero-step turn.
|
||||
- `agent/prompt-submit(agent, content, source, signal, next) → PromptDecision` — waterfall, fired for one claimed queued message before the loop opens a turn or appends `user/message`. The explicit admission signal is placed before the final `next`; `allow` optionally rewrites the prompt `content` or attaches separately sourced `additionalContexts[]`, while `block` discards the candidate without creating session history.
|
||||
|
||||
**`agent/turn-continuation`** receives and returns a `ContinuationDecision`. A `{action:'continue', reason?}` may carry model-facing content and source recorded as next-step steering in the same turn — the typed twin of the `/goal` step-end-steer pattern. It is not a `context/message`, so its type does not offer durable context metadata.
|
||||
**`agent/turn-stopping`** is an awaited notification at the natural stop boundary. A listener that needs another step calls `agent.steer()` with explicitly sourced model-facing content; the loop then re-reads the outbox and either continues or closes the turn.
|
||||
|
||||
### The tool pipeline gives each phase one kind of authority
|
||||
|
||||
@@ -33,15 +33,13 @@ Every call follows `tools/pre-execute` → guards → `tools/execute` → dispat
|
||||
|
||||
Core dispatch and the tool body sit inside normalization boundaries, so tool, listener, invalid canonical value, renderer/projector, non-JSON presentation, and identity-shape failures resolve as JSON-safe `isError` results rather than escaping the turn. A post-execute listener can therefore inspect a thrown tool; definition-owned final content invariants also cover outer pipeline and candidate-materialization failures; and a final observer sees the execution-local canonical value beside exactly the presentation fields the session log can persist. The [canonical tool-output contract](../architecture/2026-07-20-canonical-tool-output-contract.md) owns the value/projection and durability rules.
|
||||
|
||||
**`TurnEndReason.rejected`** (`dsh-session`): a zero-step turn whose claimed prompt was blocked by `prompt-submit`.
|
||||
|
||||
### Three load-bearing loop decisions
|
||||
|
||||
1. **Open the turn before prompt policy.** A blocked prompt becomes a zero-step `rejected` turn, preserving enclosure and giving ACP a durable terminal event. The veto records `prompt/blocked` with the original prompt and reason, while every allowed `additionalContexts` entry is injected into the open turn. Each claimed ordinary-send item is the sole message in its turn under the [one-send-one-turn simplification](../simplification/2026-07-17-one-send-one-turn.md); a pre-start drop creates no turn.
|
||||
1. **Run prompt policy before opening the turn.** A blocked prompt creates no turn or durable event. On allow, the loop stages the rewritten prompt followed by every returned `additionalContexts` entry, opens the turn, and drains that outbox before the first step. Each claimed ordinary-send item is the sole direct prompt in its turn under the [one-send-one-turn simplification](../simplification/2026-07-17-one-send-one-turn.md).
|
||||
|
||||
2. **Post-tool `additionalContexts` and asynchronous injections enter the active-batch FIFO and append when that batch settles.** `content`/`feedback` shape the result `execute()` returns, but each context is a separate `context/message`, and a single step or composite tool can produce many. Appending context immediately would interleave `result(c1) → context → result(c2)` or place nested context before its outer result, breaking tool-call/result adjacency. `ToolRunContext.deferContext()` therefore collects nested-dispatch context through failures, `execute()` surfaces the ordered array on `ToolExecutionResult`, and the loop accepts it into the same FIFO as `agent.inject()` calls made during execution. The FIFO appends after every recorded result when the batch settles, including before an interrupted turn closes. An accepted outer call preserves deferred contexts before decision contexts; an outer block discards deferred contexts and exposes only contexts explicitly supplied by the blocking decision.
|
||||
2. **Post-tool `additionalContexts` and asynchronous injections enter the active-batch FIFO and append when that batch settles.** `content`/`feedback` shape the result `execute()` returns, but each context is a separate sourced `user/message`, and a single step or composite tool can produce many. Appending context immediately would interleave `result(c1) → context → result(c2)` or place nested context before its outer result, breaking tool-call/result adjacency. `ToolRunContext.deferContext()` therefore collects nested-dispatch context through failures, `execute()` surfaces the ordered array on `ToolExecutionResult`, and the loop accepts it into the same FIFO as `agent.inject()` calls made during execution. The FIFO appends after every recorded result when the batch settles, including before an interrupted turn closes. An accepted outer call preserves deferred contexts before decision contexts; an outer block discards deferred contexts and exposes only contexts explicitly supplied by the blocking decision.
|
||||
|
||||
3. **A forced `continue` `reason` is enqueued through the steering channel**, so the next step's top-of-loop drain records it as steering for the continued turn — next-*step* steering within the SAME turn, not a next-*turn* prompt (matching the existing `hasSteering` force-continue override).
|
||||
3. **A stopping listener requests continuation through the steering channel**, so the next step's top-of-loop drain records it as steering for the continued turn — next-*step* steering within the SAME turn, not a next-*turn* prompt.
|
||||
|
||||
### Pre-tool input rewrite is a separate consistency decision
|
||||
|
||||
@@ -49,7 +47,7 @@ Core dispatch and the tool body sit inside normalization boundaries, so tool, li
|
||||
|
||||
### Boundaries
|
||||
|
||||
The seam package does **not** declare `hook/*` session events (the durable hook-invocation log); those belong to `dsh-hook-protocol`, because a native plugin uses typed decisions without an external hook log. The native-plugin integration test (`packages/core/agent-loop/tests/interception.spec.ts`) composes the seams through the real loop with no `hook/*` protocol. Compaction (`PreCompact`/`PostCompact`), Notification, and Codex `PermissionRequest` remain outside this decision. The [approval seam](2026-07-06-approval-seam.md) resolves `ask` decisions through `ctx.approval`, while terminal monotonic stopping is owned separately by `agent/turn-stop`.
|
||||
The seam package does **not** declare `hook/*` session events (the durable hook-invocation log); those belong to `dsh-hook-protocol`, because a native plugin uses typed decisions without an external hook log. The native-plugin integration test (`packages/core/agent-loop/tests/interception.spec.ts`) composes the seams through the real loop with no `hook/*` protocol. Compaction (`PreCompact`/`PostCompact`), Notification, and Codex `PermissionRequest` remain outside this decision. The [approval seam](2026-07-06-approval-seam.md) resolves `ask` decisions through `ctx.approval`; terminal monotonic stopping is expressed by tool-result data, while `agent/turn-stopping` is the last chance to steer another step.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
@@ -58,4 +56,4 @@ The seam package does **not** declare `hook/*` session events (the durable hook-
|
||||
|
||||
## Consequences
|
||||
|
||||
The canonical interception surface is uniformly typed without giving every extension the same power: hooks return decisions, execution wrappers wrap, terminal guards only deny, and final observers only observe. The loop owns session-start, prompt-submit, post-tool context buffering, and continuation; `dsh-tools` owns identity sealing and the five-phase execution pipeline. Their contracts are documented in [architecture.md](../../../../docs/architecture.md), package READMEs, [core interception decisions](../../../../docs/core-data-structures/core.md#interception-decisions), and [tool structures](../../../../docs/core-data-structures/tools.md). The ACP bridge maps `rejected` turns to its `cancelled` codec value, while hook-driven snapshots verify the observable bridge behavior end to end.
|
||||
The canonical interception surface is uniformly typed without giving every extension the same power: hooks return decisions, execution wrappers wrap, terminal guards only deny, and final observers only observe. The loop owns session-start, pre-turn prompt admission, post-tool context buffering, and stopping; `dsh-tools` owns identity sealing and the five-phase execution pipeline. Their contracts are documented in [architecture.md](../../../../docs/architecture.md), package READMEs, [core interception decisions](../../../../docs/core-data-structures/core.md#interception-decisions), and [tool structures](../../../../docs/core-data-structures/tools.md). The ACP bridge settles an admission rejection as `cancelled` after the agent becomes idle with no owned turn, while hook-driven snapshots verify the observable bridge behavior end to end.
|
||||
|
||||
@@ -16,9 +16,9 @@ harness 需要一套钩子子系统：用户像 Claude Code（CC）和 Codex 那
|
||||
|
||||
**Agent 事件**（`dsh-agent`）：
|
||||
- `agent/session-start(agent, source)` ——emit，在第 1 轮次之前触发一次，携带 `SessionStartSource`（`startup` 表示全新/fork 创建，`resume` 表示重新加载的持久化会话；`clear`/`compact` 保留）。纯通知，不能阻塞启动（这是有意的空白：桥接可以记录/注入，但不管控启动）。监听器通过 `agent.inject()` 注入上下文。
|
||||
- `agent/prompt-submit(agent, content, source, signal, next) → PromptDecision` ——waterfall，在轮次唯一取得所有权的排队消息追加为 `user/message` 之前触发。显式轮次 signal 位于最后的 `next` 之前；`allow` 可以重写提示词 `content` 或附加来源各自独立的 `additionalContexts[]`，而 `block` 会追加一条持久的 `prompt/blocked`，并拒绝这个零步骤轮次。
|
||||
- `agent/prompt-submit(agent, content, source, signal, next) → PromptDecision` ——waterfall，针对一条取得所有权的排队消息触发，早于循环开启轮次或追加 `user/message`。显式准入 signal 位于最后的 `next` 之前；`allow` 可以重写提示词 `content` 或附加来源各自独立的 `additionalContexts[]`，而 `block` 会丢弃该候选消息，不产生会话历史。
|
||||
|
||||
**`agent/turn-continuation`** 接收并返回一个 `ContinuationDecision`。`{action:'continue', reason?}` 可携带面向模型的内容和来源，记录为同一轮次内的下一步 steering（中途引导）——与 `/goal` step-end-steer 模式互为类型化孪生。它不是 `context/message`，因此其类型不提供持久上下文元数据。
|
||||
**`agent/turn-stopping`** 是自然停止边界上的一次 awaited 通知。需要再执行一步的监听器调用 `agent.steer()`，传入来源显式的面向模型的内容（steering，中途引导）；循环随后重新读取 outbox，继续执行或关闭轮次。
|
||||
|
||||
### 工具流水线为每个阶段赋予一种权限
|
||||
|
||||
@@ -33,15 +33,13 @@ harness 需要一套钩子子系统：用户像 Claude Code（CC）和 Codex 那
|
||||
|
||||
核心调度与工具体位于规范化边界内部，因此工具、监听器、无效规范值、渲染器／投影器、非 JSON 呈现和身份形状错误均解析为 JSON 安全的 `isError` 结果，而非逃逸出轮次。post-execute 监听器因此可以检查一个抛出异常的工具；由定义拥有的最终内容不变式也会覆盖外层流水线与候选结果实体化失败；最终观测者会同时看到执行期间的规范值，以及会话日志能够持久化的确切呈现字段。[规范工具输出契约](../architecture/2026-07-20-canonical-tool-output-contract.md)定义值／投影与持久性规则。
|
||||
|
||||
**`TurnEndReason.rejected`**（`dsh-session`）：取得所有权的提示词被 `prompt-submit` 阻止的零步骤轮次。
|
||||
|
||||
### 三个承重的循环决策
|
||||
|
||||
1. **在提示词策略之前开启轮次。** 被阻止的提示词成为零步骤的 `rejected` 轮次，保持封闭性并为 ACP（Agent Client Protocol）提供持久的终结事件。否决记录 `prompt/blocked`（含原始提示词和原因），而每个允许的 `additionalContexts` 条目都注入到已开启的轮次中。依照[一次 send 对应一个轮次的简化](../simplification/2026-07-17-one-send-one-turn.md)，每个取得所有权的 ordinary-send 条目都是其轮次中的唯一消息；启动前丢弃不会创建轮次。
|
||||
1. **在开启轮次之前运行提示词策略。** 被阻止的提示词不会创建轮次，也不产生持久事件。允许时，循环先暂存重写后的提示词，再暂存每个返回的 `additionalContexts` 条目，然后开启轮次并在第一个步骤之前排空该 outbox。依照[一次 send 对应一个轮次的简化](../simplification/2026-07-17-one-send-one-turn.md)，每个取得所有权的 ordinary-send 条目都是其轮次中唯一的直接提示词。
|
||||
|
||||
2. **工具执行后的 `additionalContexts` 与异步注入进入活跃批次 FIFO，并在该批次结算时追加。** `content`/`feedback` 塑造 `execute()` 返回的结果，但每项上下文都是独立的 `context/message`，而单个步骤或组合工具可以产生许多上下文。立即追加上下文会产生 `result(c1) → context → result(c2)` 的交错，或把嵌套上下文放在外层结果之前，破坏工具调用/结果邻接性。因此 `ToolRunContext.deferContext()` 会在失败路径上也收集嵌套调度上下文，`execute()` 在 `ToolExecutionResult` 上暴露有序数组，循环再把它接纳到与执行期间 `agent.inject()` 调用相同的 FIFO 中。FIFO 在批次结算时，于每个已记录结果之后追加，其中也包括被中断轮次关闭之前。被接受的外层调用将 deferred contexts 保留在 decision contexts 之前；被外层阻止时则丢弃 deferred contexts，只暴露阻止 decision 显式提供的上下文。
|
||||
2. **工具执行后的 `additionalContexts` 与异步注入进入活跃批次 FIFO，并在该批次结算时追加。** `content`/`feedback` 塑造 `execute()` 返回的结果，但每项上下文都是一条独立的带来源 `user/message`，而单个步骤或组合工具可以产生许多上下文。立即追加上下文会产生 `result(c1) → context → result(c2)` 的交错，或把嵌套上下文放在外层结果之前，破坏工具调用/结果邻接性。因此 `ToolRunContext.deferContext()` 会在失败路径上也收集嵌套调度上下文，`execute()` 在 `ToolExecutionResult` 上暴露有序数组，循环再把它接纳到与执行期间 `agent.inject()` 调用相同的 FIFO 中。FIFO 在批次结算时，于每个已记录结果之后追加，其中也包括被中断轮次关闭之前。被接受的外层调用将 deferred contexts 保留在 decision contexts 之前；被外层阻止时则丢弃 deferred contexts，只暴露阻止 decision 显式提供的上下文。
|
||||
|
||||
3. **强制 `continue` 的 `reason` 通过 steering 通道入队**，使得下一步骤在循环顶部排空时将其记录为当前轮次的 steering——同一轮次内的下一*步骤* steering，而非下一*轮次*的提示词（与现有的 `hasSteering` 强制继续覆盖一致）。
|
||||
3. **stopping 监听器通过 steering 通道请求继续执行**，使得下一步骤在循环顶部排空时将其记录为当前轮次的 steering——同一轮次内的下一*步骤* steering，而非下一*轮次*的提示词。
|
||||
|
||||
### 工具执行前输入重写是一个独立的一致性决策
|
||||
|
||||
@@ -49,7 +47,7 @@ harness 需要一套钩子子系统：用户像 Claude Code（CC）和 Codex 那
|
||||
|
||||
### 边界
|
||||
|
||||
seam 包**不**声明 `hook/*` 会话事件（持久的钩子调用日志）；那些属于 `dsh-hook-protocol`，因为原生插件使用类型化 decision 而无需外部钩子日志。原生插件集成测试（`packages/core/agent-loop/tests/interception.spec.ts`）通过真实循环组合这些 seam，不涉及 `hook/*` 协议。压缩（compaction）（`PreCompact`/`PostCompact`）、Notification 和 Codex `PermissionRequest` 不在本决策范围内。[审批 seam](2026-07-06-approval-seam.md) 通过 `ctx.approval` 解析 `ask` decision，而终结性的单调停止由 `agent/turn-stop` 独立负责。
|
||||
seam 包**不**声明 `hook/*` 会话事件（持久的钩子调用日志）；那些属于 `dsh-hook-protocol`，因为原生插件使用类型化 decision 而无需外部钩子日志。原生插件集成测试（`packages/core/agent-loop/tests/interception.spec.ts`）通过真实循环组合这些 seam，不涉及 `hook/*` 协议。压缩（compaction）（`PreCompact`/`PostCompact`）、Notification 和 Codex `PermissionRequest` 不在本决策范围内。[审批 seam](2026-07-06-approval-seam.md) 通过 `ctx.approval` 解析 `ask` decision；终结性的单调停止由工具结果数据表达，而 `agent/turn-stopping` 是引导再执行一步的最后机会。
|
||||
|
||||
## 曾考虑的替代方案
|
||||
|
||||
@@ -58,4 +56,4 @@ seam 包**不**声明 `hook/*` 会话事件（持久的钩子调用日志）；
|
||||
|
||||
## 后果
|
||||
|
||||
规范拦截表面具有统一的类型化，同时不给每个扩展相同的权力：钩子返回 decision，执行包装层做包装，终结 guard 只能拒绝，最终观测者只能观测。循环负责 session-start、prompt-submit、工具执行后上下文缓冲和 continuation；`dsh-tools` 负责身份封存与五阶段执行流水线。它们的契约记录在 [architecture.md](../../../../docs/architecture.md)、各包 README、[核心拦截 decision](../../../../docs/core-data-structures/core.md#interception-decisions) 与[工具结构](../../../../docs/core-data-structures/tools.md)中。ACP 桥接将 `rejected` 轮次映射为其 `cancelled` 编解码值，而钩子驱动的快照端到端验证可观测的桥接行为。
|
||||
规范拦截表面具有统一的类型化，同时不给每个扩展相同的权力：钩子返回 decision，执行包装层做包装，终结 guard 只能拒绝，最终观测者只能观测。循环负责 session-start、轮次前的提示词准入、工具执行后上下文缓冲和 stopping；`dsh-tools` 负责身份封存与五阶段执行流水线。它们的契约记录在 [architecture.md](../../../../docs/architecture.md)、各包 README、[核心拦截 decision](../../../../docs/core-data-structures/core.md#interception-decisions) 与[工具结构](../../../../docs/core-data-structures/tools.md)中。ACP 桥接在 agent 空闲且不再拥有轮次后，将准入拒绝结算为 `cancelled`，而钩子驱动的快照端到端验证可观测的桥接行为。
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-05-skill-system.md: 95772067b3b224f15fd290876c297eccbe5ab97e
|
||||
2026-07-05-skill-system.zh.md: 589410076b6da1b108b01f55bf217001b2a8404f
|
||||
# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-07-05-skill-system.md
|
||||
2026-07-05-skill-system.md: 4fc621a9fdfa8042ebf3eb1975f0930cb1bf116c
|
||||
2026-07-05-skill-system.zh.md: 0dbebd211a1fa9e434d3f0a189c936f2b1c76574
|
||||
|
||||
@@ -12,9 +12,9 @@ DeepSeek Harness uses the same primitive so project-specific review, plugin-auth
|
||||
|
||||
## Decision
|
||||
|
||||
`@deepseek-ai/dsh-skill` is the pure provider registry (`ctx.skills`), `@deepseek-ai/dsh-skill-local` is the shipped local filesystem provider, and `@deepseek-ai/dsh-tool-skill` owns the session-prefix catalog and model-facing loader tool. `dsh-agent-spine-demo` loads the registry, local provider, and consumer by default so TUI, headless, and ACP apps get the same behavior while embedded or remote providers contribute skills without changing the registry or consumer. Its `skills` config forwards `registry`, `local`, and `tool` branches to those owners.
|
||||
`@deepseek-ai/dsh-skill` is the pure provider registry (`ctx.skills`), `@deepseek-ai/dsh-skill-local` is the shipped local filesystem provider, and `@deepseek-ai/dsh-tool-skill` owns the durable session catalog and model-facing loader tool. `dsh-agent-spine-demo` loads the registry, local provider, and consumer by default so TUI, headless, and ACP apps get the same behavior while embedded or remote providers contribute skills without changing the registry or consumer. Its `skills` config forwards `registry`, `local`, and `tool` branches to those owners.
|
||||
|
||||
Provider plugins register synchronously during `apply()`. Provider membership is direct effect-owned state: registration and disposal invalidate completed catalogs synchronously, and discovery reads the current provider map on demand rather than observing registry-change events. Provider catalogs return ranked candidates from awaited `list()` calls, where remote providers perform initialization, authentication, and discovery while honoring the lookup abort signal. The registry validates each candidate, resolves same-name skills first-wins by rank, provider registration order, and provider-local order, then sorts summaries by skill name for deterministic consumers. It caches only completed catalog snapshots and retries when a provider/runtime revision changes during discovery, so an unload cannot freeze a stale, unresolvable skill into a session prefix. Runtime `ctx.skills.register(...)` remains a convenience for embedded in-process skills and uses project-over-user priority; `runtime` is reserved as the registry-owned provider name.
|
||||
Provider plugins register synchronously during `apply()`. Provider membership is direct effect-owned state: registration and disposal invalidate completed catalogs synchronously, and discovery reads the current provider map on demand rather than observing registry-change events. Provider catalogs return ranked candidates from awaited `list()` calls, where remote providers perform initialization, authentication, and discovery while honoring the lookup abort signal. The registry validates each candidate, resolves same-name skills first-wins by rank, provider registration order, and provider-local order, then sorts summaries by skill name for deterministic consumers. It caches only completed catalog snapshots and retries when a provider/runtime revision changes during discovery, so an unload cannot freeze a stale, unresolvable skill into a session catalog. Runtime `ctx.skills.register(...)` remains a convenience for embedded in-process skills and uses project-over-user priority; `runtime` is reserved as the registry-owned provider name.
|
||||
|
||||
The local provider scans cwd-sensitive project roots, custom roots, and user roots in first-wins rank order: project `.dsh`, project `.agents`, `customSkillDirs`, user `.dsh`, then user `.agents`. The user `.dsh/skills` scan skips `.system` so a system-owned directory is not treated as normal user content. DeepSeek Harness does not ship built-in system skills; embedded or remote providers supply additional skills when configured.
|
||||
|
||||
@@ -22,7 +22,7 @@ Each skill is either `<name>/SKILL.md` or `<name>.md` with YAML frontmatter. `na
|
||||
|
||||
Local skill filesystem I/O goes through `ctx.fs` when a filesystem service is loaded: project-root lookup probes `.git` with `resolve` and `stat`, root discovery uses `listDir`, and skill reads use `readText`. The Node filesystem remains a fallback for minimal contexts that mount `dsh-skill-local` without the fs seam. Missing roots, unreadable or malformed skill files, and transient provider `list()` failures degrade to warn-and-skip so one bad source does not make every agent request fail; malformed candidates still fail fast because they are provider contract violations.
|
||||
|
||||
`dsh-tool-skill` contributes one user-role `<system-reminder>` catalog through [`agent/session-prefix`](2026-07-07-session-prefix.md). The catalog contains sorted skill name and description only; it excludes bodies, paths, sources, providers, and routing hints. Descriptions are whitespace-normalized, XML-escaped, and capped by `catalogDescriptionMaxLength`, whose default is `500` and minimum is `3`. The session-prefix seam freezes the request-only catalog per loop instance and records it in the request header, preserving reconstructability without adding it to durable history. Full skill bodies are never included in the catalog.
|
||||
`dsh-tool-skill` injects one durable user-role `<system-reminder>` catalog as a sourced `user/message` at the session's first `agent/step`, and only when that agent's tool view resolves this plugin's exact `skill` registration. The catalog contains sorted skill name and description only; it excludes bodies, paths, sources, providers, and routing hints. Descriptions are whitespace-normalized, XML-escaped, and capped by `catalogDescriptionMaxLength`, whose default is `500` and minimum is `3`. Full skill bodies are never included in the catalog. (The catalog originally rode the request-only [session-prefix seam](../../archived/feature/2026-07-07-session-prefix.md), archived; the [unified sourced-message decision](../architecture/2026-07-22-unified-send-and-coalesced-user-messages.md) moved it into durable history.)
|
||||
|
||||
The `skill({ name })` tool loads one full skill for the current agent cwd and returns a tool result containing `<skill_content name="...">`, `<skill_resources>`, and `<skill_instructions>`. `resourceBase` supplies a directory, URL, or opaque provider-managed base for explicitly referenced scripts, references, and assets; resources load only as needed, without directory enumeration. An unresolved name reports that the skill is unknown or no longer available; invalid names and skills marked `disableModelInvocation` retain distinct tool errors. The tool result is the model-visible disclosure path.
|
||||
|
||||
@@ -36,7 +36,7 @@ The data structures and catalog/tool contract are documented in [skills.md](../.
|
||||
|
||||
**Put local filesystem scanning directly inside `ctx.skills`.** Rejected because coding agents, web agents, and future plugin ecosystems need different skill sources. A provider registry mirrors the subagent seam: the registry owns conflict resolution and consumers, while implementations own loading.
|
||||
|
||||
**Use a system-prompt section.** Rejected because the rendered system prompt is a single string, while the catalog is a user-role `<system-reminder>` message with request-only lifecycle requirements. [`agent/session-prefix`](2026-07-07-session-prefix.md) is the selected mechanism: it places the catalog ahead of derived history and records the composed message in the request header.
|
||||
**Use a system-prompt section.** Rejected because the rendered system prompt is a single string, while the catalog is a user-role `<system-reminder>` message. The [request-only session-prefix seam](../../archived/feature/2026-07-07-session-prefix.md) (archived) was the original mechanism; after the unified sourced-message decision removed that seam, the catalog became a durable sourced injection with the same message shape.
|
||||
|
||||
**Materialize built-in DSH authoring skills under `~/.dsh/skills/.system`.** Rejected because bundled skills do not write user home on startup, and embedded or remote providers supply configured skills.
|
||||
|
||||
@@ -46,7 +46,7 @@ The data structures and catalog/tool contract are documented in [skills.md](../.
|
||||
|
||||
## Consequences
|
||||
|
||||
The agent-core spine includes one session-prefix contributor, one local provider, and one model-facing tool. Skill discovery is cwd-sensitive, so callers that create agents with different session cwd values can observe different project skill overrides by design.
|
||||
The agent-core spine includes one catalog contributor, one local provider, and one model-facing tool. Skill discovery is cwd-sensitive, so callers that create agents with different session cwd values can observe different project skill overrides by design.
|
||||
|
||||
The catalog is deterministic for a fixed root set and runtime registration revision, but disk changes are not watched; discovery is memoized until runtime registration invalidates the cache or the process restarts.
|
||||
|
||||
|
||||
@@ -12,9 +12,9 @@ DeepSeek Harness 使用同一原语，使项目特定的评审、插件编写和
|
||||
|
||||
## 决策
|
||||
|
||||
`@deepseek-ai/dsh-skill` 是纯提供方注册表（`ctx.skills`），`@deepseek-ai/dsh-skill-local` 是随附的本地文件系统提供方，`@deepseek-ai/dsh-tool-skill` 负责会话前缀目录与面向模型的 loader 工具。`dsh-agent-spine-demo` 默认加载注册表、本地提供方和消费方，使 TUI、headless 与 ACP（Agent Client Protocol）应用获得相同行为，同时嵌入式或远程提供方可在不修改注册表或消费方的前提下贡献 skill。其 `skills` 配置将 `registry`、`local` 和 `tool` 分支分别转发给对应的所有者。
|
||||
`@deepseek-ai/dsh-skill` 是纯提供方注册表（`ctx.skills`），`@deepseek-ai/dsh-skill-local` 是随附的本地文件系统提供方，`@deepseek-ai/dsh-tool-skill` 负责持久化会话目录与面向模型的 loader 工具。`dsh-agent-spine-demo` 默认加载注册表、本地提供方和消费方，使 TUI、headless 与 ACP（Agent Client Protocol）应用获得相同行为，同时嵌入式或远程提供方可在不修改注册表或消费方的前提下贡献 skill。其 `skills` 配置将 `registry`、`local` 和 `tool` 分支分别转发给对应的所有者。
|
||||
|
||||
提供方插件在 `apply()` 期间同步注册。提供方成员资格是由直接 effect 持有的状态：注册与 dispose（资源释放）同步地使已完成的目录失效，发现操作按需读取当前提供方映射而非监听注册表变更事件。提供方目录从等待的 `list()` 调用返回排序后的候选项，远程提供方在此过程中执行初始化、认证和发现，同时遵守查找的 abort 信号。注册表校验每个候选项，按排名、提供方注册顺序和提供方内部顺序以先到先得方式解决同名 skill 冲突，然后按 skill 名称排序摘要以保证消费方获得确定性结果。它仅缓存已完成的目录快照，并在发现过程中提供方/运行时修订版本发生变化时重试，因此卸载操作不会将一个陈旧且不可解析的 skill 冻结到会话前缀中。运行时 `ctx.skills.register(...)` 仍作为嵌入式进程内 skill 的便捷方式保留，使用 project 优先于 user 的优先级；`runtime` 保留为注册表拥有的提供方名称。
|
||||
提供方插件在 `apply()` 期间同步注册。提供方成员资格是由直接 effect 持有的状态：注册与 dispose（资源释放）同步地使已完成的目录失效，发现操作按需读取当前提供方映射而非监听注册表变更事件。提供方目录从等待的 `list()` 调用返回排序后的候选项，远程提供方在此过程中执行初始化、认证和发现，同时遵守查找的 abort 信号。注册表校验每个候选项，按排名、提供方注册顺序和提供方内部顺序以先到先得方式解决同名 skill 冲突，然后按 skill 名称排序摘要以保证消费方获得确定性结果。它仅缓存已完成的目录快照，并在发现过程中提供方/运行时修订版本发生变化时重试，因此卸载操作不会将一个陈旧且不可解析的 skill 冻结到会话目录中。运行时 `ctx.skills.register(...)` 仍作为嵌入式进程内 skill 的便捷方式保留，使用 project 优先于 user 的优先级；`runtime` 保留为注册表拥有的提供方名称。
|
||||
|
||||
本地提供方按先到先得的排名顺序扫描 cwd 敏感的项目根目录、自定义根目录和用户根目录：项目 `.dsh`、项目 `.agents`、`customSkillDirs`、用户 `.dsh`，然后是用户 `.agents`。用户 `.dsh/skills` 扫描跳过 `.system`，以免系统拥有的目录被当作普通用户内容处理。DeepSeek Harness 不随附内置系统 skill；嵌入式或远程提供方在配置后提供额外 skill。
|
||||
|
||||
@@ -22,7 +22,7 @@ DeepSeek Harness 使用同一原语，使项目特定的评审、插件编写和
|
||||
|
||||
本地 skill 的文件系统 I/O 在加载了文件系统服务时通过 `ctx.fs` 进行：项目根目录查找使用 `resolve` 和 `stat` 探测 `.git`，根目录发现使用 `listDir`，skill 读取使用 `readText`。Node 文件系统作为后备，供在不挂载 fs seam 的最小上下文中加载 `dsh-skill-local` 时使用。缺失的根目录、不可读或格式错误的 skill 文件、以及提供方 `list()` 的瞬态失败均降级为警告并跳过，使一个坏源不会导致所有 agent 请求失败；格式错误的候选项仍然快速失败，因为它们违反了提供方契约。
|
||||
|
||||
`dsh-tool-skill` 通过 [`agent/session-prefix`](2026-07-07-session-prefix.md) 贡献一个 user-role `<system-reminder>` 目录。该目录仅包含排序后的 skill 名称与描述；不包含正文、路径、来源、提供方和路由提示。描述经过空白规范化、XML 转义，并受 `catalogDescriptionMaxLength` 上限约束，其默认值为 `500`，最小值为 `3`。session-prefix seam 将仅用于请求的目录按 loop 实例冻结，并记录在请求头中，在不将其加入持久化历史的前提下保持可重建性。完整的 skill 正文从不包含在目录中。
|
||||
`dsh-tool-skill` 在会话的第一个 `agent/step` 注入一个持久化的 user-role `<system-reminder>` 目录，作为带来源的 `user/message`，且仅当该 agent 的工具视图解析到本插件精确的 `skill` 注册时才注入。该目录仅包含排序后的 skill 名称与描述；不包含正文、路径、来源、提供方和路由提示。描述经过空白规范化、XML 转义，并受 `catalogDescriptionMaxLength` 上限约束，其默认值为 `500`，最小值为 `3`。完整的 skill 正文从不包含在目录中。（目录最初通过仅请求的[会话前缀 seam](../../archived/feature/2026-07-07-session-prefix.md)（已归档）传递；[统一带来源消息的决策](../architecture/2026-07-22-unified-send-and-coalesced-user-messages.md)将其移入持久化历史。）
|
||||
|
||||
`skill({ name })` 工具为当前 agent cwd 加载一个完整 skill，返回包含 `<skill_content name="...">`、`<skill_resources>` 和 `<skill_instructions>` 的工具结果。`resourceBase` 提供一个目录、URL 或不透明的提供方管理的基路径，用于显式引用的脚本、参考资料和资产；资源仅按需加载，不进行目录枚举。无法解析的名称报告该 skill 未知或不再可用；无效名称和标记了 `disableModelInvocation` 的 skill 保留不同的工具错误。工具结果是面向模型的可见披露路径。
|
||||
|
||||
@@ -36,7 +36,7 @@ DeepSeek Harness 使用同一原语，使项目特定的评审、插件编写和
|
||||
|
||||
**将本地文件系统扫描直接放入 `ctx.skills`。** 否决，因为编码 agent、Web agent 和未来的插件生态需要不同的 skill 来源。提供方注册表与 subagent seam 镜像：注册表拥有冲突解决和消费方，实现拥有加载。
|
||||
|
||||
**使用系统提示词段落。** 否决，因为渲染后的系统提示词是单一字符串，而目录是一条具有仅请求生命周期要求的 user-role `<system-reminder>` 消息。[`agent/session-prefix`](2026-07-07-session-prefix.md) 是选定的机制：它将目录置于派生历史之前，并将组合后的消息记录在请求头中。
|
||||
**使用系统提示词段落。** 否决，因为渲染后的系统提示词是单一字符串，而目录是一条 user-role `<system-reminder>` 消息。[仅请求的会话前缀 seam](../../archived/feature/2026-07-07-session-prefix.md)（已归档）是最初的机制；统一带来源消息的决策移除该 seam 后，目录改为具有相同消息形状的持久化带来源注入。
|
||||
|
||||
**在 `~/.dsh/skills/.system` 下物化内置 DSH 编写 skill。** 否决，因为打包的 skill 不应在启动时写入用户主目录，嵌入式或远程提供方在配置后提供 skill。
|
||||
|
||||
@@ -46,7 +46,7 @@ DeepSeek Harness 使用同一原语，使项目特定的评审、插件编写和
|
||||
|
||||
## 后果
|
||||
|
||||
agent-core 主干包含一个 session-prefix 贡献者、一个本地提供方和一个面向模型的工具。Skill 发现是 cwd 敏感的，因此以不同会话 cwd 值创建 agent 的调用方可以按设计观察到不同的项目 skill 覆盖。
|
||||
agent-core 主干包含一个目录贡献者、一个本地提供方和一个面向模型的工具。Skill 发现是 cwd 敏感的，因此以不同会话 cwd 值创建 agent 的调用方可以按设计观察到不同的项目 skill 覆盖。
|
||||
|
||||
目录对于固定的根目录集合和运行时注册修订版本是确定性的，但不监视磁盘变化；发现结果被缓存，直到运行时注册使缓存失效或进程重启。
|
||||
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-06-sandbox.md: c6883873192f15ba2982436e156d8795396c0148
|
||||
2026-07-06-sandbox.zh.md: d84df9b06b15dd296801073d381603f34cfd2878
|
||||
# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-07-06-sandbox.md
|
||||
2026-07-06-sandbox.md: b9c410e4498a565a7414085fddfb74856592da66
|
||||
2026-07-06-sandbox.zh.md: 876870a12f1205b4f2cc2c8c13b5e7ec815a1569
|
||||
|
||||
@@ -165,7 +165,7 @@ What shipped pins — the tiers in Testing hold each:
|
||||
- A resumed session's overrides apply with no catch-up state; a default changed while the process was down is narrated before the session's first new request, attributed to the operator.
|
||||
- Two concurrent sessions never see each other's state or notices.
|
||||
- Two concurrent project sessions in one Cordis context resolve independent workspace roots; bash and fs writes succeed inside the calling session's cwd and fail against its neighbor's cwd.
|
||||
- `agent-loop` is untouched — everything rides `systemPrompt.section`, `SessionEventMap` merging, `agent.inject()`, `agent/pre-step`, `agent/prompt-submit`, and capability-owned policy resolution.
|
||||
- `agent-loop` is untouched — everything rides `systemPrompt.section`, `SessionEventMap` merging, `agent.inject()`, `agent/step`, `agent/prompt-submit`, and capability-owned policy resolution.
|
||||
|
||||
Costs and accepted limits:
|
||||
|
||||
|
||||
@@ -128,7 +128,7 @@ fs/web/todo 在进程内执行，因此它们的沙箱语义是各自 seam 层
|
||||
|
||||
- **第二个消费方**——`subagent-acp` 可选地约束子 agent（按调用策略；默认无约束——子 agent 必须写入自己的持久化）。
|
||||
- **更多环境**——环境一致的能力组示例（如 bash+fs 对一个容器）。
|
||||
- **Windows 链**——`PLATFORM_CHAINS.win32` 保留为空（失败关闭）；填充它意味着来自 AppContainer/restricted-token 家族的约束 runner，从其自己的仓库按 `node-addon-landlock-run` 模板交付，加上其 profile 方言和拒绝/runner 失败签名。改为包装第三方 landstrip runner 的方案[经考虑后已驳回](../../rejected/feature/2026-07-26-evaluate-landstrip-for-windows-sandbox-rung.md)——对安全不变式而言，它还远未经过实战检验。
|
||||
- **Windows 链**——`PLATFORM_CHAINS.win32` 保留为空（失败关闭）；填充它意味着来自 AppContainer/restricted-token 家族的约束 runner，从其自己的仓库按 `node-addon-landlock-run` 模板交付，加上其 profile 方言和拒绝/runner 失败签名。改为包装第三方 landstrip runner 的方案[经考虑后已驳回](../../rejected/feature/2026-07-26-evaluate-landstrip-for-windows-sandbox-rung.md)——它所经受的实战检验还不足以承载安全不变式。
|
||||
|
||||
## 曾考虑的替代方案
|
||||
|
||||
@@ -165,7 +165,7 @@ fs/web/todo 在进程内执行，因此它们的沙箱语义是各自 seam 层
|
||||
- 恢复的会话的覆盖直接生效，无需追赶状态；进程停止期间变更的默认值在会话的首个新请求前被叙述，归因于运维人员。
|
||||
- 两个并发会话永远看不到彼此的状态或通知。
|
||||
- 同一个 Cordis 上下文中的两个并发项目会话解析各自独立的工作区根目录；bash 和 fs 写入在调用方会话的 cwd 内成功，对其相邻会话的 cwd 则失败。
|
||||
- `agent-loop` 未被触及——一切搭载 `systemPrompt.section`、`SessionEventMap` 合并、`agent.inject()`、`agent/pre-step`、`agent/prompt-submit` 和由能力拥有的策略解析。
|
||||
- `agent-loop` 未被触及——一切搭载 `systemPrompt.section`、`SessionEventMap` 合并、`agent.inject()`、`agent/step`、`agent/prompt-submit` 和由能力拥有的策略解析。
|
||||
|
||||
代价与已接受的限制：
|
||||
|
||||
|
||||
@@ -1,43 +0,0 @@
|
||||
# Agent Note: The session prefix — request-only messages in front of the derived history
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-07-07-session-prefix.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
A plugin often owns a session-stable opener the model must always see — a skills catalog, an AGENTS.md digest, a workspace baseline. Before this seam the harness offered two homes, and both are wrong for that content. The system prompt is one rendered string: message-shaped content (a user-role `<system-reminder>` envelope, a multi-message primer) does not fit it, and providers weight conversation messages differently from system text. Durable history (`agent.inject()`, a `context/message` at session start) makes the opener permanent: every `deriveMessages()` consumer replays it, the compaction retention walk owns it, forks bake it in stale, and a resume cannot refresh it — a catalog captured at session birth outlives the world it described.
|
||||
|
||||
The obvious third option — let a plugin edit the request's `messages` on the way out — is banned by [the reconstructable-requests Agent Note](../architecture/2026-07-05-reconstructable-requests.md): every loop-built request is a pure function of the session log, so whatever channel carries the opener must log exactly what it sends. What was missing was a request-only message channel with a durable record.
|
||||
|
||||
## Decision
|
||||
|
||||
`agent/session-prefix` is a waterfall on the agent event map ([`packages/core/agent/src/types.ts`](../../../../packages/core/agent/src/types.ts)): listeners receive a frozen empty seed and return an extension (the canonical contribution is a prepend, `[mine, ...await next()]`, which yields registration order on the wire). The loop ([`packages/core/agent-loop/src/loop.ts`](../../../../packages/core/agent-loop/src/loop.ts)) fires it once per loop instance, lazily before the instance's first `agent/pre-step`; the composed list is deep-cloned, deep-frozen, cached on the instance, and placed in front of the ENTIRE derived history — directly after the provider's system slot — on every request the instance sends ([wire order](../../../../docs/core-data-structures/core.md#the-request-envelope-llmcallconfig-and-the-logged-header)).
|
||||
|
||||
Three properties carry the design:
|
||||
|
||||
- **Request-only, header-logged.** `deriveMessages()` never returns the prefix; its one durable record is `EpochHeader.messagePrefix` on the instance's anchoring `request/header` snapshot — the channel the reconstructable-requests Agent Note already owns for the request's non-history half, so no new session event exists. The [`dsh-agent-loop/invariant`](../../../../packages/core/agent-loop/src/invariant.ts) companion recomputes `messagePrefix + boundary derivation` against every loop-built request; an unlogged prefix cannot reach the wire when that contribution is enabled.
|
||||
- **Frozen per instance.** Reuse is structural, not disciplined: the cached product cannot change mid-session, so the provider's prompt cache holds by construction and the prefix extends the cacheable region at zero marginal cost per step. A process restart or `ctx.agents.resume()` is a new instance: it recomposes, and any drift lands attributably on the `'resume'` header snapshot. This is the routing rule the seam creates: session-frozen openers ride the prefix; content that changes mid-session rides the append-only history channels (`agent.inject()` or tool/prompt-submit `additionalContexts` — [the interception-seams Agent Note](2026-06-30-interception-seams.md)), each a durable `context/message` paid once and prefix-cached thereafter.
|
||||
- **Exact in the durable request envelope.** Composition precedes the instance's first `agent/pre-step` and request boundary. The first routed request logs the current prefix on its header, so post-step token pressure reads the exact prefix together with the actual prompt, tools, and routed model; no compaction-only parameter is carried through the generic pre-step seam. A composition interrupted by cancel/dispose is discarded, never cached: an abort-aware listener's degraded fallback cannot leak into later requests, and the next turn recomposes under a live signal.
|
||||
|
||||
Because composition runs before the boundary snapshot, a composing listener's session append joins the CURRENT request's derived history. Compaction structurally cannot touch the prefix (or the system prompt): it rewrites surface nodes, and header state never enters the surface.
|
||||
|
||||
## Testing
|
||||
|
||||
[Interception tests](../../../../packages/core/agent-loop/tests/interception.spec.ts) pin compose-once reuse without changed headers, prepend order, empty-prefix omission, immutability, composition before pre-step, and the prefix on the routed header; [cancellation tests](../../../../packages/core/agent-loop/tests/cancel.spec.ts) pin discard and recomposition. Session, invariant, token-meter, and compaction tests cover header round trips, request reconstruction, and durable prefix-aware pressure accounting. Snapshot normalization preserves prefix counts, while the [pinned-header scenario](../../archived/testing/2026-07-06-pin-request-header-content-in-one-scenario.md) owns content and the default example remains prefix-free. The provider-independent seam needs no dedicated e2e; the with-key [request-cache e2e](../../../../packages/core/agent-loop/tests/request-cache.e2e.ts) covers its cache economics.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
- **Per-request `before`/`after` slots recomputed every step** (the shape first proposed: a waterfall firing on every request, contributing frozen `before` messages ahead of the history and fresh `after` messages behind it) — rejected. A per-step `before` recompose invites drift that must be logged as a full changed header, and an `after` slot sits behind the growing history, so its tokens re-pay on every request and everything after it is uncacheable. Measured against the alternatives, every current update pattern is served cheaper by a durable append (paid once, cache-read thereafter), and the only content with no home was the session-stable opener — which wants freezing, not recomputation.
|
||||
- **A system-prompt section** (`system-prompt/assemble`) — rejected for this content: the assembly renders to the single `system` string, so message-shaped openers do not fit, and the system prompt is deliberately re-assembled per step (with a full changed header when it changes) while the opener wants instance-frozen semantics.
|
||||
- **A durable history opener** (`inject()` at session start) — rejected: permanent history is the failure mode in the problem statement — replayed everywhere, compactable, stale across resumes.
|
||||
- **Compose per turn instead of per instance** — rejected: a turn-boundary recompose either desyncs silently from the log or forces a changed header, and it busts the provider cache exactly as often as it fires; the legitimate refresh point is the instance boundary, where the `'resume'` snapshot already records drift attributably.
|
||||
- **Carry prompt/prefix through `agent/pre-step` for provisional pressure** — rejected because it couples a generic lifecycle seam to one consumer and still misses later request routing and tools; post-step replay reads every request-envelope field from its durable routed header.
|
||||
- **A dedicated session event carrying the prefix** — rejected: the header events are the request's non-history record by design; a second event would be a second home for the same fact and another codec to keep total.
|
||||
|
||||
## Consequences
|
||||
|
||||
- `agent/pre-step` stays a generic `(agent, turn, step, signal)` checkpoint. Compaction receives no prefix parameter; `ctx.tokenMeter` folds the prefix from the canonical routed header at post-step.
|
||||
- A contributor whose content changes mid-session is not re-read until the next instance — by design. A deployment needing mid-session catalog updates routes the change notice through the append-only history channels and pays one durable `context/message`.
|
||||
- The dropped `after` slot leaves no request-only channel near the request tail; nothing in the repo needs one, and adding it back would re-open the every-step re-pay cost the design exists to avoid.
|
||||
- An empty composition is canonical absence: no-contributor deployments log no extra header bytes and their requests are the bare derivation.
|
||||
@@ -1,43 +0,0 @@
|
||||
# Agent Note: 会话前缀——派生历史之前的仅请求消息
|
||||
|
||||
Status: implemented
|
||||
|
||||
[English](2026-07-07-session-prefix.md) | 中文
|
||||
|
||||
## 问题
|
||||
|
||||
插件经常拥有一段会话级别稳定的开场内容，模型必须始终看到它：技能目录、AGENTS.md 摘要、工作区基线。在引入本 seam 之前，harness 为这类内容提供了两个归属位置，但两者都不合适。系统提示词是一个渲染后的单一字符串：消息形态的内容（user 角色的 `<system-reminder>` 信封、多消息引导序列）放不进去，而且提供方对会话消息和系统文本的权重处理不同。持久化历史（`agent.inject()`、会话启动时的 `context/message`）使开场内容变为永久：每个 `deriveMessages()` 消费方都会回放它，压缩（compaction）的保留遍历拥有它，fork 会将其以陈旧状态固化，恢复也无法刷新它——会话诞生时捕获的目录会比它所描述的世界活得更久。
|
||||
|
||||
显而易见的第三种选项——让插件在请求发出途中编辑 `messages`——被[可重建请求 Agent Note](../architecture/2026-07-05-reconstructable-requests.md)禁止：每个由循环构建的请求都是会话日志的纯函数，因此无论哪个通道承载开场内容，都必须精确记录它所发送的内容。缺失的是一个带有持久记录的仅请求消息通道。
|
||||
|
||||
## 决策
|
||||
|
||||
`agent/session-prefix` 是 agent 事件映射上的一个 waterfall（瀑布式事件）（[`packages/core/agent/src/types.ts`](../../../../packages/core/agent/src/types.ts)）：监听器接收一个冻结的空种子并返回扩展（规范的贡献方式是前置插入 `[mine, ...await next()]`，在协议格式上产生注册顺序）。agent loop（智能体循环）（[`packages/core/agent-loop/src/loop.ts`](../../../../packages/core/agent-loop/src/loop.ts)）在每个循环实例中触发一次，惰性地在实例首次 `agent/pre-step` 之前执行；组合后的列表被深拷贝、深冻结、缓存在实例上，并在该实例发出的每个请求中置于整个派生历史之前——紧接在提供方的 system 槽位之后（[协议格式顺序](../../../../docs/core-data-structures/core.md#the-request-envelope-llmcallconfig-and-the-logged-header)）。
|
||||
|
||||
三个属性承载了这一设计：
|
||||
|
||||
- **仅请求，记录在 header 中。** `deriveMessages()` 从不返回前缀；它唯一的持久记录是实例锚定的 `request/header` 快照上的 `EpochHeader.messagePrefix`——可重建请求 Agent Note 已为请求的非历史部分拥有的通道，因此不引入新的会话事件。配套的 [`dsh-agent-loop/invariant`](../../../../packages/core/agent-loop/src/invariant.ts)对每个循环构建的请求重新计算 `messagePrefix + boundary derivation`；启用该贡献时，未记录的前缀无法到达协议格式。
|
||||
- **按实例冻结。** 复用是结构性的，而非靠纪律保证：缓存的产物在会话中途不可变，因此提供方的提示词缓存从构造上成立，前缀以每步零边际成本扩展了可缓存区域。进程重启或 `ctx.agents.resume()` 产生新实例：它重新组合，任何漂移都可追溯地落在 `'resume'` header 快照上。这就是本 seam 创建的路由规则：会话冻结的开场内容走前缀；会话中途变化的内容走仅追加历史通道（`agent.inject()` 或工具/prompt-submit 的 `additionalContexts`——[拦截 seam Agent Note](2026-06-30-interception-seams.md)），每条都是一次性支付的持久 `context/message`，之后被前缀缓存覆盖。
|
||||
- **在持久请求信封中保持精确。** 组合先于实例的首次 `agent/pre-step` 和请求边界。第一个已路由请求会把当前前缀记录在其 header 上，因此步骤后的 token 压力会将精确前缀与实际提示词、工具和已路由模型一起读取；通用的步骤前检查点 seam 不携带压缩专属参数。被取消/dispose 中断的组合会被丢弃，永不缓存：感知中止的监听器的降级回退不会泄漏到后续请求中，下一轮次在活信号下重新组合。
|
||||
|
||||
由于组合在边界快照之前运行，组合监听器的会话追加会加入当前请求的派生历史。压缩在结构上不可能触及前缀（或系统提示词）：它重写的是表面节点，而 header 状态从不进入表面。
|
||||
|
||||
## 测试
|
||||
|
||||
[拦截测试](../../../../packages/core/agent-loop/tests/interception.spec.ts)固定了以下行为：没有变更 header 时的组合一次复用、前置插入顺序、空前缀省略、不可变性、组合在步骤前检查点之前完成，以及已路由 header 上的前缀；[取消测试](../../../../packages/core/agent-loop/tests/cancel.spec.ts)固定了丢弃与重新组合。Session、不变式、token-meter 和压缩测试覆盖 header 往返、请求重建与持久前缀感知的压力核算。快照归一化保留前缀计数，[固定 header 场景](../../archived/testing/2026-07-06-pin-request-header-content-in-one-scenario.md)拥有内容，默认示例保持无前缀。与提供方无关的 seam 无需专门 e2e；带密钥的 [request-cache e2e](../../../../packages/core/agent-loop/tests/request-cache.e2e.ts) 覆盖了其缓存经济性。
|
||||
|
||||
## 曾考虑的替代方案
|
||||
|
||||
- **每请求 `before`/`after` 槽位，每步重新计算**（最初提出的形态：一个每请求触发的 waterfall，贡献冻结的 `before` 消息置于历史之前、新鲜的 `after` 消息置于历史之后）：否决。每步重新组合 `before` 会引入漂移，必须记录为完整的变更 header；`after` 槽位位于不断增长的历史之后，其 token 在每个请求中重复支付，且其后的所有内容不可缓存。对照各替代方案衡量，当前所有更新模式都能通过持久追加更廉价地满足（支付一次，此后缓存读取），而唯一没有归属的内容是会话稳定的开场——它需要的是冻结，而非重新计算。
|
||||
- **系统提示词分段**（`system-prompt/assemble`）：对此类内容否决。assembly 渲染为单一 `system` 字符串，消息形态的开场放不进去；且系统提示词被设计为每步重新组装（变化时带完整的变更 header），而开场内容需要按实例冻结的语义。
|
||||
- **持久化历史开场**（会话启动时 `inject()`）：否决。永久历史正是问题陈述中的失败模式——到处被回放、可被压缩、在恢复后仍保持陈旧状态。
|
||||
- **按轮次组合而非按实例组合**：否决。轮次边界的重新组合要么与日志静默失同步，要么强制产生变更 header；且它每次触发都会破坏提供方缓存。合理的刷新点是实例边界，`'resume'` 快照已在那里可追溯地记录漂移。
|
||||
- **通过 `agent/pre-step` 携带提示词/前缀，用于临时压力估算**：否决，因为它把通用生命周期 seam 耦合到一个消费方，而且仍会遗漏更晚的请求路由和工具；步骤后的回放会从持久的已路由 header 读取请求信封的每个字段。
|
||||
- **专用会话事件承载前缀**：否决。header 事件按设计就是请求的非历史记录；第二个事件会为同一事实提供第二个归属，并多出一个需要保持完整的编解码器。
|
||||
|
||||
## 后果
|
||||
|
||||
- `agent/pre-step` 保持通用的 `(agent, turn, step, signal)` 检查点。压缩不接收 prefix 参数；`ctx.tokenMeter` 在步骤后从规范的已路由 header 折叠前缀。
|
||||
- 贡献者的内容在会话中途变化时，直到下一个实例才会被重新读取——这是设计意图。需要会话中途目录更新的部署，应将变更通知路由到仅追加历史通道，支付一条持久 `context/message`。
|
||||
- 被放弃的 `after` 槽位意味着请求尾部附近没有仅请求通道；仓库中没有任何功能需要它，且恢复它会重新引入本设计旨在避免的每步重复支付成本。
|
||||
- 空组合即为规范缺失：无贡献者的部署不记录额外的 header 字节，其请求就是裸派生。
|
||||
@@ -1,76 +0,0 @@
|
||||
# Agent Note: Repeat-tool-call guard plugin
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-07-08-repeat-tool-guard.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
A model stuck in a loop re-issues the same tool call with byte-identical arguments — re-running a failing grep, re-reading an unchanged file, polling a command that already gave its answer — and each round trip burns tokens, wall-clock, and (for paid APIs) money without adding information. The harness has nothing that notices: the loop has no step budget, no plugin tracks call repetition, and the model only escapes when it happens to vary its own behavior. The failure mode is real and cheap to detect — [pi-repeat-tool-guard](https://github.com/Kingwl/pi-repeat-tool-guard) ships exactly this as a pi coding-agent extension: count consecutive identical calls and, past a threshold, append a `<system-reminder>` telling the model to stop repeating itself and change course.
|
||||
|
||||
The harness already has every seam the pi extension uses, and better ones: [the interception-seams Agent Note](2026-06-30-interception-seams.md) gives `tools/post-execute` a sanctioned way to attach model-facing context to a finished call, the loop buffers and injects that context with call/result adjacency preserved, and injected context is a logged `context/message` — so a native guard satisfies the model-visible ⟺ logged rule with no new session event. What was missing was only the plugin itself.
|
||||
|
||||
## Decision
|
||||
|
||||
The guard is a loop-hygiene plugin, not a model-facing tool. It counts consecutive calls to the same tool with identical canonical arguments and injects advisory reminders at configured thresholds. It never delays, blocks, or rewrites a call; the model decides whether to retry differently or finish.
|
||||
|
||||
The plugin is `@deepseek-ai/dsh-repeat-tool-guard` at `packages/guard/repeat-tool-guard/`, opening the `guard/` group for loop-hygiene plugins (single-package groups have precedent: [the todo-write Agent Note](2026-06-29-todo-write-tool.md) shipped `todo/tool-todo`). It registers two listeners and holds state in a `WeakMap` keyed by the live `Agent` object — the tool registry is a context-level singleton whose waterfalls interleave every agent's calls (subagents run on the same context), so per-agent keying is correctness, not polish; weak object keys also make a disposal-only cleanup listener unnecessary.
|
||||
|
||||
- **`tools/post-execute` (waterfall)** — the one detection point. The listener receives `(exec, result)` together, so counting and reminder delivery need no cross-event pending map (the pi extension needs one only because its `tool_call`/`tool_result` hooks are separate events). It always delegates via `next()` and, when a threshold is hit, prepends a reminder to the downstream decision's `additionalContexts` — the observe-and-enrich posture [the hooks bridges](2026-06-30-hook-bridges.md) already use, honoring the waterfall contract. Counting happens here rather than in `tools/pre-execute` because post-execute also runs for denied calls (`ToolRegistry.execute` routes a deny through the same pipeline), and a model hammering a denied call is exactly the loop worth breaking.
|
||||
- **`agent/prompt-submit` (waterfall)** — pure reset hook: delegate via `next()`, clear the submitting agent's chain. A user interjection changes the context; repetition across it is not a loop.
|
||||
|
||||
### Detection semantics
|
||||
|
||||
The chain key is `(tool name, canonical arguments)`; a call identical to the previous tracked call increments the agent's consecutive counter, a different tracked call resets it to 1. Canonicalization is a deep key-sort plus `JSON.stringify`: `ToolExecution.arguments` is by construction the loop's `JSON.parse` output (or the raw string fallback for malformed argument JSON, which is itself a comparable value), so the pi original's bigint/circular/`undefined` handling has no inputs here and is deliberately dropped.
|
||||
|
||||
Two deliberate rules, both documented in [the package README](../../../../packages/guard/repeat-tool-guard/README.md) because they are behavior a reader would otherwise guess at:
|
||||
|
||||
- **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 — and it is the pi extension's (undocumented) semantics, kept on purpose and written down.
|
||||
- **Calls without an agent are ignored.** A direct `ctx.tools.execute()` caller (tests, non-loop consumers) has no model to remind and no live agent object to key on.
|
||||
|
||||
### Reminder delivery
|
||||
|
||||
Reminders ride `additionalContexts` as their own entries (source `{kind: 'plugin', plugin: 'repeat-tool-guard'}` — the label is load-bearing per `HookContext`), never a `content` replacement: the `tool/result` event stays the tool's own output for audit, and the loop appends buffered contexts as `context/message`s after the step's results, which the session renders as tagged synthetic-user envelopes and derived history replays. Thresholds escalate: the first configured threshold gets a short "you are repeating yourself, analyze the previous result" nudge; each later threshold gets the detailed form naming the tool, the repeat count, and the canonical arguments (head-truncated at `argumentsPreviewChars`, default 500 — a looping `write`-sized payload must not ride into the next request unbounded; the chain key always compares the full canonical string), and stating that the calls made no progress. The pi original hardcodes the gentle text to the literal count 3; the guard keys it to `thresholds[0]`, fixing that bug in the port. A downstream hook bridge contribution remains a separate array entry, so both plugins retain their source, envelope, and metadata.
|
||||
|
||||
### 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` is validated at load and throws on an empty list, a non-integer, a value below 2, or a duplicate — misconfiguration fails loud, replacing the pi original's silent fall-back to defaults. `include`/`exclude` entries support `*` wildcards. Patterns are predicates over whatever tools exist at call time, not references to a registry entry, so an entry matching no currently registered tool is NOT an error — unlike `toolOrder`'s referent check, `exclude: [mcp_*]` must stay valid in a deployment that loads no MCP tools.
|
||||
|
||||
## Testing
|
||||
|
||||
- **Unit:** A real loop with a scripted adapter covers counting and reset rules, untracked transparency, disposal cleanup, per-agent isolation, canonical argument key order, escalation, denied calls, no-agent execution, wildcard escaping, invalid config, and downstream block or replacement decisions at per-file 100% coverage.
|
||||
- **Snapshot:** The keyless `repeat-tool-guard` scenario makes five identical `todo_write` calls and pins the gentle third-call and detailed fifth-call reminders in both ACP output and the session log. The plugin is loaded in the live example but remains inert in other scenarios.
|
||||
- **E2e:** None; the plugin is deterministic and provider-independent, and its seam contracts are covered by their owners.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
- **Append the reminder into the tool result** (`accept` with replaced `content` — the pi extension's mechanism, which patches result content because that is the only channel its API offers) — rejected: it makes the logged `tool/result` lie about what the tool returned, and `additionalContexts` is the separate sanctioned channel for post-execute commentary, with loop-level buffering that preserves call/result adjacency.
|
||||
- **Count in `tools/pre-execute` with a pending-reminder map** (the pi two-phase shape) — rejected: post-execute alone sees `(exec, result)` together and also fires for denied calls, so one listener with no cross-event state covers strictly more attempts with less machinery.
|
||||
- **Escalate to `block` at the highest threshold** — rejected for the initial scope: a blocked call punishes legitimate identical repeats (polling a long-running terminal, re-checking a file the agent expects to change), and an advisory reminder keeps the model in control. Revisit with evidence; the decision shape (`PostToolDecision`) already supports it.
|
||||
- **A per-deployment external hook via the CC/Codex bridges** (a `PostToolUse` script) — rejected as the answer: it works for one deployment, but a shipped, unit-tested, `cordis.yml`-configurable plugin is the harness-native form, without per-call subprocess cost.
|
||||
- **A loop-level step or repetition budget in `agent-loop`** — rejected: "plugins, not loop changes"; a hard step budget is a blunter, orthogonal control that would need its own proposal.
|
||||
- **Fuzzy/near-identical detection** (normalized paths, similar-but-not-equal arguments) — rejected: exact match after canonicalization is cheap, deterministic, and explainable to the model; similarity thresholds invite false positives and need evidence before they earn complexity.
|
||||
- **Placing the package in `core/`** — rejected: core is the product spine; a behavioral guard is an optional leaf plugin, and the `todo/` precedent is a small dedicated group per plugin family.
|
||||
|
||||
## Consequences
|
||||
|
||||
- The reminder is advisory by design: idempotent polling patterns that repeat identical calls on purpose still receive nudges past the thresholds, and the pressure valves are config (`thresholds`, `exclude`) plus reminder text that explicitly allows finishing when enough evidence has been gathered. Each trigger costs reminder tokens on the next request; thresholds bound the frequency.
|
||||
- Chain state is in-memory only: a session resumed from persistence starts with a fresh chain, so a loop spanning a resume draws its reminders later than a live one — accepted, the guard is a heuristic nudge, not a logged invariant, and persisting counter state would buy little for real complexity.
|
||||
- When multiple post-execute producers attach context on one call, each contribution stays a separate `HookContext`; ordering follows waterfall nesting and each entry retains its own provenance.
|
||||
- Implementing the snapshot tier surfaced a hidden assumption in the suite kit: the fixture guard equated "authored model scenario" with "override-driven". The `Scenario` table now carries an explicit `overridden` flag, and the sidecar's presence is checked BOTH ways against it (an unregistered stray sidecar would silently replace the derived script) — the suite kit is stricter than it was before this plugin existed.
|
||||
|
||||
## Deferred
|
||||
|
||||
- Compaction does not reset chains: a compacted history changes what the model sees, but the repetition risk usually survives compaction.
|
||||
- Escalating to `block` at a high threshold is not implemented; `PostToolDecision` already supports it if evidence arrives.
|
||||
- Subagent chains stay isolated per agent; no sharing mechanism exists until a concrete case appears.
|
||||
@@ -1,76 +0,0 @@
|
||||
# Agent Note: 重复工具调用守卫插件
|
||||
|
||||
Status: implemented
|
||||
|
||||
[English](2026-07-08-repeat-tool-guard.md) | 中文
|
||||
|
||||
## 问题
|
||||
|
||||
模型陷入循环时，会以字节级相同的参数反复发起同一个工具调用——重新运行一条失败的 grep、重新读取一个未变化的文件、轮询一条已经给出答案的命令——每一轮往返都消耗 token、挂钟时间以及（对付费 API 而言）金钱，却不带来新信息。harness 目前没有任何机制能察觉这一点：循环没有步骤预算，没有插件追踪调用重复，模型只有在碰巧改变自身行为时才能跳出。这种失败模式真实存在且检测成本极低——[pi-repeat-tool-guard](https://github.com/Kingwl/pi-repeat-tool-guard) 正是以 pi coding-agent 扩展的形式提供了这一功能：统计连续相同调用次数，超过阈值后追加一条 `<system-reminder>` 告诉模型停止重复并换个方向。
|
||||
|
||||
harness 已经具备 pi 扩展所使用的全部 seam，而且更好：[拦截 seam Agent Note](2026-06-30-interception-seams.md)赋予 `tools/post-execute` 一种经过认可的方式，将面向模型的上下文附加到已完成的调用上；循环缓冲并注入该上下文，同时保持调用/结果的邻接关系；注入的上下文是一条已记录的 `context/message`——因此原生守卫无需新增会话事件即可满足「模型可见 ⟺ 已记录」规则。缺少的只是插件本身。
|
||||
|
||||
## 决策
|
||||
|
||||
该守卫是一个循环卫生插件，而非面向模型的工具。它统计对同一工具以相同规范化参数发起的连续调用次数，并在配置的阈值处注入建议性提醒。它从不延迟、阻止或改写调用；模型自行决定是换种方式重试还是结束。
|
||||
|
||||
插件为 `@deepseek-ai/dsh-repeat-tool-guard`，位于 `packages/guard/repeat-tool-guard/`，开辟 `guard/` 分组用于循环卫生插件（单包（package）分组有先例：[todo-write Agent Note](2026-06-29-todo-write-tool.md)发布了 `todo/tool-todo`）。它注册两个监听器，将状态保存在以存活 `Agent` 对象为键的 `WeakMap` 中——工具注册表是上下文级别的单例，其 waterfall（瀑布式事件）交错所有 agent（智能体）的调用（subagent 运行在同一个上下文上），因此按 agent 分键是正确性要求，而非锦上添花；弱对象键还使得纯清理用途的 disposal 监听器不再必要。
|
||||
|
||||
- **`tools/post-execute`（waterfall）**——唯一的检测点。监听器同时接收 `(exec, result)`，因此计数和提醒投递无需跨事件的 pending map（pi 扩展需要它，仅因为其 `tool_call`/`tool_result` 钩子是分开的事件）。它始终通过 `next()` 委托，当命中阈值时，将提醒前置到下游决策的 `additionalContexts`——这正是[钩子桥接](2026-06-30-hook-bridges.md)已采用的「观察并丰富」姿态，遵守 waterfall 契约。计数放在此处而非 `tools/pre-execute`，因为 post-execute 也会为被拒绝的调用触发（`ToolRegistry.execute` 将 deny 路由到同一条流水线），而模型反复敲击一个被拒绝的调用恰恰是值得打破的循环。
|
||||
- **`agent/prompt-submit`（waterfall）**——纯重置钩子：通过 `next()` 委托，清除提交 agent 的链。用户介入改变了上下文；跨越介入的重复不是循环。
|
||||
|
||||
### 检测语义
|
||||
|
||||
链的键是 `(tool name, canonical arguments)`；与前一个被追踪调用相同的调用递增该 agent 的连续计数器，不同的被追踪调用将其重置为 1。规范化方式为深度键排序加 `JSON.stringify`：`ToolExecution.arguments` 按构造就是循环中 `JSON.parse` 的输出（或格式错误的参数 JSON 的原始字符串回退，其本身也是可比较的值），因此 pi 原版对 bigint/循环引用/`undefined` 的处理在此没有输入，被有意去除。
|
||||
|
||||
两条刻意的规则，均记录在[包 README](../../../../packages/guard/repeat-tool-guard/README.md) 中，因为它们是读者否则只能猜测的行为：
|
||||
|
||||
- **未追踪的调用对链透明。** 被 `include`/`exclude` 排除的调用既不递增也不重置计数器，因此 `grep X → todo_write → grep X` 在 `todo_write` 被排除时仍计为两次连续的 `grep X`。这正是排除功能有用的原因——穿插在循环中的簿记工具不得为循环洗白——也是 pi 扩展的（未文档化的）语义，有意保留并明确写下。
|
||||
- **没有 agent 的调用被忽略。** 直接调用 `ctx.tools.execute()` 的调用方（测试、非循环消费方）没有可提醒的模型，也没有可作键的存活 agent 对象。
|
||||
|
||||
### 提醒投递
|
||||
|
||||
提醒作为独立条目搭载在 `additionalContexts` 上（source 为 `{kind: 'plugin', plugin: 'repeat-tool-guard'}`——依照 `HookContext`，该标签承载语义），绝不替换 `content`：`tool/result` 事件仍是工具自身的审计输出，循环则在步骤结果之后把缓冲的上下文追加为 `context/message`，会话将其渲染为带标签的合成 user 信封，并由派生历史回放。阈值逐级升级：第一个配置阈值获得简短的「你正在重复自己，请分析先前结果」提示；后续各阈值获得详细形式，包含工具、重复计数和规范参数（在头部截断到 `argumentsPreviewChars`，默认 500——循环中的 `write` 级 payload 不得无界地进入下一次请求；链键始终比较完整规范字符串），并说明这些调用没有取得进展。pi 原版把温和文本硬编码为字面计数 3；本守卫以 `thresholds[0]` 为键，修复了移植中的这一 bug。下游钩子桥贡献仍是独立数组条目，因此两个插件都保留各自的 source、信封与元数据。
|
||||
|
||||
### 配置
|
||||
|
||||
```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` 在加载时校验，遇到空列表、非整数、小于 2 的值或重复项时抛出异常——配置错误快速失败，取代 pi 原版的静默回退到默认值。`include`/`exclude` 条目支持 `*` 通配符。模式是对调用时实际存在的工具的谓词，而非对注册表条目的引用，因此匹配不到当前已注册工具的条目不是错误——与 `toolOrder` 的引用检查不同，`exclude: [mcp_*]` 在未加载 MCP 工具的部署中也必须保持有效。
|
||||
|
||||
## 测试
|
||||
|
||||
- **单元测试：** 使用脚本化适配器的真实循环，覆盖计数与重置规则、未追踪透明性、dispose（资源释放）清理、按 agent 隔离、规范化参数键序、升级、被拒绝的调用、无 agent 执行、通配符转义、无效配置，以及下游阻止或 replacement 决策，达到逐文件 100% 覆盖率。
|
||||
- **快照测试：** keyless 的 `repeat-tool-guard` 场景发起五次相同的 `todo_write` 调用，在 ACP 输出和会话日志中固定第三次调用的温和提醒与第五次调用的详细提醒。该插件在实时示例中加载，但在其他场景中保持静默。
|
||||
- **E2e 测试：** 无。该插件是确定性的且与提供方无关，其 seam 契约由各自的所有者覆盖。
|
||||
|
||||
## 曾考虑的替代方案
|
||||
|
||||
- **将提醒追加到工具结果中**（以替换 `content` 的方式 `accept`——pi 扩展的机制，它修补结果内容是因为那是其 API 提供的唯一通道）：否决。这会让已记录的 `tool/result` 对工具实际返回的内容撒谎，而 `additionalContexts` 是 post-execute 评注的独立认可通道，循环级缓冲保持了调用/结果的邻接关系。
|
||||
- **在 `tools/pre-execute` 中计数并使用 pending-reminder map**（pi 的两阶段形态）：否决。post-execute 单独就能同时看到 `(exec, result)` 且也为被拒绝的调用触发，因此一个监听器、无跨事件状态即可以更少的机制覆盖严格更多的尝试。
|
||||
- **在最高阈值升级为 `block`**：在初始范围内否决。阻止调用会惩罚合法的相同重复（轮询长时间运行的终端、重新检查 agent 预期会变化的文件），而建议性提醒让模型保持控制权。待有证据后重新审视；决策形状（`PostToolDecision`）已支持此选项。
|
||||
- **通过 CC/Codex 桥接的逐部署外部钩子**（一个 `PostToolUse` 脚本）：否决作为最终答案。它对单个部署有效，但一个已发布、有单元测试、可通过 `cordis.yml` 配置的插件才是 harness 原生的形式，且没有逐调用的子进程开销。
|
||||
- **在 `agent-loop` 中设置循环级步骤或重复预算**：否决。「用插件，不改循环」；硬性步骤预算是一种更粗粒度的正交控制，需要自己的提案。
|
||||
- **模糊/近似相同检测**（路径归一化、相似但不完全相同的参数）：否决。规范化后的精确匹配成本低、确定性强、且可向模型解释；相似度阈值引入误报风险，需要证据才能换取复杂度。
|
||||
- **将包放在 `core/`**：否决。core 是产品主干；行为守卫是可选的叶子插件，`todo/` 的先例是每个插件族一个小型专属分组。
|
||||
|
||||
## 后果
|
||||
|
||||
- 提醒在设计上是建议性的：有意重复相同调用的幂等轮询模式仍会在超过阈值后收到提示，减压阀是配置（`thresholds`、`exclude`）加上明确允许「在已收集足够证据时结束」的提醒文本。每次触发在下一次请求中增加提醒 token 的开销；阈值限制了触发频率。
|
||||
- 链状态仅存于内存：从持久化恢复的会话以全新的链开始，因此跨越恢复的循环比实时循环更晚收到提醒——可以接受，守卫是启发式提示而非已记录的不变式，持久化计数器状态带来的收益不值得其复杂度。
|
||||
- 当多个 post-execute 生产者在同一次调用上附加上下文时，每项贡献保持为独立的 `HookContext`；顺序遵循 waterfall 嵌套关系，每个条目保留自己的溯源信息。
|
||||
- 实现快照层时暴露了 suite kit 的一项隐藏假设：fixture guard 把「撰写的模型场景」等同于「由 override 驱动」。`Scenario` 表现在携带显式的 `overridden` 标志，并且 sidecar 是否存在会以双向方式与其核对（未注册的游离 sidecar 会静默替换派生脚本）——suite kit 比本插件出现前更严格。
|
||||
|
||||
## 延后事项
|
||||
|
||||
- 压缩（compaction）不重置链：压缩后的历史改变了模型所见的内容，但重复风险通常在压缩后仍然存在。
|
||||
- 在高阈值升级为 `block` 未实现；`PostToolDecision` 已支持此选项，待证据到来时启用。
|
||||
- subagent 的链按 agent 隔离；在出现具体用例之前不提供共享机制。
|
||||
@@ -1,6 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-08-self-referential-cordis-toolset.md: 80bffa3a2a959939f18fd1d3422607cf61895fc7
|
||||
2026-07-08-self-referential-cordis-toolset.zh.md: 2ec79037045fdb040cccf31699789abdd3a12db2
|
||||
# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-07-08-self-referential-cordis-toolset.md
|
||||
2026-07-08-self-referential-cordis-toolset.md: 40934fe0e2975c4e068df6ef8f31ed7921df3230
|
||||
2026-07-08-self-referential-cordis-toolset.zh.md: 13662b9359aa85895ce85391ebd5a5902cc451cc
|
||||
|
||||
@@ -12,19 +12,19 @@ First, model-written registration must be validated where it happens: a malforme
|
||||
|
||||
## Decision
|
||||
|
||||
The toolset ships as [`@deepseek-ai/dsh-tool-cordis`](../../../../packages/cordis/tool-cordis/README.md) — a new top-level `packages/cordis/` group — and is demoed by [`examples/cordis-agent`](../../../../examples/cordis-agent/README.md). It gives the model three tools over the live cordis runtime it is running inside: inspect it, mount model-written plugins into it, dispose them again.
|
||||
The toolset ships as [`@deepseek-ai/dsh-tool-cordis`](../../../../packages/cordis/tool-cordis/README.md) and is demoed by [`examples/cordis-agent`](../../../../examples/cordis-agent/README.md). It gives the model three tools over the live Cordis runtime in the current DSH process: inspect it, mount an in-memory temporary Plugin, and unmount that Plugin to quiescence.
|
||||
|
||||
The vm isolates accidental global pollution, and the context façade hides framework internals. Neither restricts the authority of exposed services: a mount can call `ctx.bash` to run commands with the host executor's privileges and can reach the real filesystem and web services. This is an opt-in development tool with bash-equivalent trust, not a security boundary or product default.
|
||||
The vm isolates accidental global pollution, and the context façade hides framework internals. Neither restricts the authority of exposed services: a temporary Plugin can call `ctx.bash` with the host executor's privileges and reach the real filesystem and web services. It runs in the shared DSH runtime and may affect other sessions in that process. This is an opt-in development tool with bash-equivalent trust, not a security boundary or product default.
|
||||
|
||||
### The three tools
|
||||
|
||||
| Tool | Contract |
|
||||
|---|---|
|
||||
| `cordis_inspect` | Read-only report over the live runtime, one Markdown section per `what` value (omit `what` for all sections). An exact `name` with `what: "api"` or `what: "events"` narrows to one source-documented target. Never mutates. |
|
||||
| `cordis_mount` | Evaluates `code` (the body of an async JavaScript function) in a `node:vm` sandbox; the code must `return` a cordis plugin, which is mounted as a child of the `cordis-dynamic` group fiber and tracked under a fresh id (`dyn-1`, `dyn-2`, …). |
|
||||
| `cordis_unmount` | Disposes one dynamic mount by id and returns only after disposal reaches quiescence — every registration the plugin made is unwound, not merely requested to stop. |
|
||||
| `cordis_inspect` | Read-only report over the live current-process runtime, one Markdown section per `what` value (omit `what` for all sections). `plugins` lists every live fiber; `temporary` lists only the temporary Plugins created by `cordis_mount`. An exact `name` with `what: "api"` or `what: "events"` narrows to one source-documented target. |
|
||||
| `cordis_mount` | Evaluates `code` now as an async JavaScript-function body in a `node:vm` sandbox and saves it nowhere. The returned Plugin is mounted under the internal `cordis-dynamic` group and tracked under a fresh process-local id (`dyn-1`, `dyn-2`, …). |
|
||||
| `cordis_unmount` | Unmounts one `cordis_mount` temporary Plugin by id and returns only after every owned tool, listener, service, timer, and effect reaches quiescence. It cannot remove Loader, configured, or installed Plugins. |
|
||||
|
||||
`cordis_inspect` sections: `services` (every provided ctx service and the owning fiber, non-active owners flagged), `plugins` (a flat list of every loaded plugin with its lifecycle state, from `ctx.registry` — what capabilities are loaded, deliberately not the tree shape), `tools` (what the model can call), `dynamic` (the mount table: id, name, state, provided services, awaited services), `api` (live service signatures + the type shapes they reference, from the generated catalog), and `events` (harness events with dispatch mode and signature). Broad `api` and `events` reports omit full JSDoc to stay compact; an exact `name` returns one service or event with its original method/declaration JSDoc. A name is invalid with other sections, unknown targets fail, and an API target must be live. The model-facing tool descriptions carry the operational rules the model needs at call time; [the generated tool catalog](../../../../docs/tool-catalog.md) is their exhaustive rendering.
|
||||
`cordis_inspect` sections are `services` (every provided ctx service and owning fiber), `plugins` (every live plugin fiber), `tools` (what the model can call), `temporary` (the `cordis_mount` subset with id, running/pending state, provided and awaited services, and lifetime), `api` (live service signatures and referenced types), and `events` (harness events with dispatch mode and signature). Temporary Plugins remain active across later turns and disappear after `cordis_unmount`, toolset unload, or DSH restart; they are never restored automatically. Broad `api` and `events` reports omit full JSDoc to stay compact; an exact `name` returns one service or event with its original method/declaration JSDoc. A name is invalid with other sections, unknown targets fail, and an API target must be live. The model-facing tool descriptions carry the operational rules needed at call time; [the generated tool catalog](../../../../docs/tool-catalog.md) is their exhaustive rendering.
|
||||
|
||||
### Sandbox semantics
|
||||
|
||||
@@ -36,9 +36,11 @@ Mount code crosses the vm boundary through three controls. Dual-realm `instanceo
|
||||
|
||||
The boundary normalizes unambiguous JSON-Schema forms into `ParameterSchemaSpec`, preserving `integer`, raw object openness, and required arrays. Direct DSL object nodes must declare `additionalProperties`; invalid vocabulary fails with the accepted alternatives. Parse, TypeScript, missing-return, Node-API, and duplicate-tool errors include the relevant source line or corrective contract without narrating implementation internals.
|
||||
|
||||
### The dynamic group and mount lifecycle
|
||||
### The internal group and temporary-Plugin lifecycle
|
||||
|
||||
All dynamic mounts are children of one `cordis-dynamic` group beneath the tool plugin, so ordinary fiber disposal handles reload and unload. Mounting awaits settlement; startup failure disposes the fiber before returning an error. A settled pending mount remains visible with its missing injections. `cordis_unmount` awaits the mount fiber's disposal.
|
||||
Every temporary Plugin is a child of one internal `cordis-dynamic` group beneath the tool plugin, so ordinary fiber disposal handles toolset reload and unload. `cordis_mount` awaits settlement; startup failure disposes the fiber before returning an error. A settled pending Plugin remains visible with its missing injections. `cordis_unmount` awaits the Plugin fiber's disposal.
|
||||
|
||||
Temporary Plugins exist only in process memory. They create no Plugin file, install no package, change no `cordis.yml` or personal/project configuration, do not survive restart, and have no automatic save, promote, or install path. Keeping an experiment means asking the Agent to implement a normal local, project, or repository Plugin through the regular development workflow.
|
||||
|
||||
### Cross-mount composition via provide/inject
|
||||
|
||||
@@ -52,9 +54,9 @@ Freshness is gated like every generated artifact: `pnpm run verify-cordis-api` (
|
||||
|
||||
### Configuration, rendering, and observability
|
||||
|
||||
The plugin exposes one config field, validated by schemastery and documented in [the config catalog](../../../../docs/config-catalog.md): `vmTimeoutMs` (default 5000), the millisecond bound on the synchronous portion of mount-code evaluation. Tool names, the `cordis-dynamic` group name, and the `dyn-` id prefix are structural vocabulary and stay fixed. All three tools render as `generic` cards per [the tool cookbook](../../../../docs/cookbook/adding-a-tool.md) (`cordis_inspect` a `read`, `cordis_mount` an `execute` carrying the code as `rawInput`, `cordis_unmount` a `delete`), with no `presentResult` overrides.
|
||||
The plugin exposes one config field, validated by schemastery and documented in [the config catalog](../../../../docs/config-catalog.md): `vmTimeoutMs` (default 5000), the millisecond bound on the synchronous portion of code evaluation. The current model-facing names are `cordis_inspect`, `cordis_mount`, and `cordis_unmount`; the internal `cordis-dynamic` group name and `dyn-` id prefix remain structural vocabulary. All three tools render as `generic` cards per [the tool cookbook](../../../../docs/cookbook/adding-a-tool.md): inspect is `read`, mount is `execute` carrying code as `rawInput`, and unmount is `delete`. Web conversation rows preserve those generic mechanics while giving the tools the action titles `Inspect`, `Mount temporary Plugin`, and `Unmount temporary Plugin` plus one shared Cordis accent; the mount row retains the shared JavaScript expansion and syntax highlighting.
|
||||
|
||||
Model-visible ⟺ logged holds with no new session event type: a mount or unmount is visible only through its own `tool/call` / `tool/result` pair, which the loop logs, and the changed tool set a mount induces is logged by the full changed request header the loop emits when schemas change between steps. There is deliberately no `cordis/mount` provenance event — it would duplicate what the tool-call pair records. Dynamic mounts are process-lifetime, not session state: resuming a persisted session rehydrates the conversation but does not re-mount plugins.
|
||||
Model-visible ⟺ logged holds with no new session event type: mount and unmount are visible through their logged `tool/call` / `tool/result` pairs, and any changed tool set is logged by the full changed request header emitted when schemas change between steps. Temporary Plugins are process memory, not session state: session resume rehydrates conversation history but never recreates them.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
|
||||
@@ -12,19 +12,19 @@ Status: implemented
|
||||
|
||||
## 决策
|
||||
|
||||
该工具集以 [`@deepseek-ai/dsh-tool-cordis`](../../../../packages/cordis/tool-cordis/README.md) 发布——一个新的顶层 `packages/cordis/` 分组——并由 [`examples/cordis-agent`](../../../../examples/cordis-agent/README.md) 演示。它为模型提供三个工具，操作模型自身运行其中的活跃 cordis 运行时：审视它、将模型编写的插件挂载进去、再将其释放。
|
||||
该工具集以 [`@deepseek-ai/dsh-tool-cordis`](../../../../packages/cordis/tool-cordis/README.md) 发布，并由 [`examples/cordis-agent`](../../../../examples/cordis-agent/README.md) 演示。它为模型提供三个工具，操作当前 DSH 进程中的活跃 Cordis 运行时：审视它、挂载一个仅存于内存的临时 Plugin，再将该 Plugin 卸载至完全停稳。
|
||||
|
||||
vm 隔离了意外的全局污染，上下文门面隐藏了框架内部细节。但二者都不限制已暴露服务的权限：一个挂载可以调用 `ctx.bash` 以宿主执行器的权限运行命令，也能访问真实的文件系统和网络服务。这是一个需要显式启用的开发工具，信任等级与 bash 相当，不是安全边界，也不是产品默认配置。
|
||||
vm 隔离了意外的全局污染，上下文门面隐藏了框架内部细节。但二者都不限制已暴露服务的权限：临时 Plugin 可以调用 `ctx.bash` 以宿主执行器的权限运行命令，也能访问真实的文件系统和网络服务。它运行在共享 DSH runtime 中，可能影响同一进程的其他 session。这是一个需要显式启用的开发工具，信任等级与 bash 相当，不是安全边界，也不是产品默认配置。
|
||||
|
||||
### 三个工具
|
||||
|
||||
| 工具 | 契约 |
|
||||
|---|---|
|
||||
| `cordis_inspect` | 对活跃运行时的只读报告，每个 `what` 值对应一个 Markdown 段落（省略 `what` 则输出全部段落）。精确的 `name` 搭配 `what: "api"` 或 `what: "events"` 可收窄到一个带源码文档的目标。从不产生变更。 |
|
||||
| `cordis_mount` | 在 `node:vm` 沙箱中执行 `code`（一个异步 JavaScript 函数的函数体）；代码必须 `return` 一个 cordis 插件，该插件作为 `cordis-dynamic` 分组 fiber 的子节点挂载，并以一个新 id（`dyn-1`、`dyn-2`……）跟踪。 |
|
||||
| `cordis_unmount` | 按 id 释放一个动态挂载，并等到释放达到完全停稳后才返回——该插件所做的每一项注册都被撤销，而不仅仅是请求停止。 |
|
||||
| `cordis_inspect` | 当前进程活跃运行时的只读报告，每个 `what` 值对应一个 Markdown 段落（省略 `what` 则输出全部段落）。`plugins` 列出全部存活 fiber，`temporary` 只列 `cordis_mount` 创建的临时 Plugin。精确 `name` 搭配 `what: "api"` 或 `what: "events"` 可收窄到一个带源码文档的目标。 |
|
||||
| `cordis_mount` | 立即在 `node:vm` 沙箱中把 `code` 作为异步 JavaScript 函数体求值，且不保存到任何位置。返回的 Plugin 挂在内部 `cordis-dynamic` 分组下，并用新的进程内 id（`dyn-1`、`dyn-2`……）跟踪。 |
|
||||
| `cordis_unmount` | 按 id 卸载一个 `cordis_mount` 临时 Plugin，并只在其自有工具、监听器、服务、定时器和其他 effect 完全停稳后返回。它不能删除 Loader、配置或已安装的 Plugin。 |
|
||||
|
||||
`cordis_inspect` 的段落：`services`（每个已提供的 ctx 服务及其所属 fiber，非活跃的所有者会被标记）、`plugins`（来自 `ctx.registry` 的所有已加载插件的扁平列表及其生命周期状态——展示加载了哪些能力，刻意不展示树形结构）、`tools`（模型可调用的工具）、`dynamic`（挂载表：id、名称、状态、提供的服务、等待的服务）、`api`（来自生成目录的活跃服务签名及其引用的类型形状）和 `events`（harness 事件及其分发模式和签名）。宽泛的 `api` 和 `events` 报告省略完整 JSDoc 以保持紧凑；精确 `name` 会返回一个服务或事件，以及其原始方法/声明 JSDoc。其他段落不能搭配 name，未知目标会失败，而 API 目标必须处于活跃状态。面向模型的工具描述携带了模型在调用时所需的操作规则；[生成的工具目录](../../../../docs/tool-catalog.md)是其完整呈现。
|
||||
`cordis_inspect` 的段落是 `services`（每个已提供的 ctx 服务及所属 fiber）、`plugins`（全部存活 Plugin fiber）、`tools`（模型可调用的工具）、`temporary`（`cordis_mount` 子集，包含 id、running／pending 状态、提供与等待的服务和生命周期）、`api`（活跃服务签名及其引用类型）和 `events`（harness 事件及分发模式和签名）。临时 Plugin 可跨后续 turn 保持活跃，并在 `cordis_unmount`、工具集卸载或 DSH 重启后消失；系统绝不会自动恢复它们。宽泛的 `api` 和 `events` 报告省略完整 JSDoc；精确 `name` 返回一个服务或事件及其原始 JSDoc。其他段落不能搭配 name，未知目标会失败，而 API 目标必须处于活跃状态。[生成的工具目录](../../../../docs/tool-catalog.md)完整呈现面向模型的调用契约。
|
||||
|
||||
### 沙箱语义
|
||||
|
||||
@@ -36,9 +36,11 @@ vm 隔离了意外的全局污染，上下文门面隐藏了框架内部细节
|
||||
|
||||
边界将无歧义的 JSON-Schema 形式规范化为 `ParameterSchemaSpec`，同时保留 `integer`、原始对象开放性和 required 数组。直接使用 DSL 的对象节点必须声明 `additionalProperties`；无效词汇会报错并给出可接受的替代方案。解析错误、TypeScript 错误、缺少 return、Node API 误用和重复工具名等错误信息包含相关源码行或纠正性契约，不叙述实现内部细节。
|
||||
|
||||
### 动态分组与挂载生命周期
|
||||
### 内部分组与临时 Plugin 生命周期
|
||||
|
||||
所有动态挂载都是工具插件下方 `cordis-dynamic` 分组的子节点，因此普通的 fiber 释放即可处理重载和卸载。挂载会等待 settlement；启动失败时在返回错误前释放 fiber。已 settle 但处于 pending 状态的挂载仍然可见，并列出其缺失的注入。`cordis_unmount` 等待挂载 fiber 的释放完成。
|
||||
每个临时 Plugin 都是工具插件下方内部 `cordis-dynamic` 分组的子节点，因此普通的 fiber 释放即可处理工具集重载和卸载。`cordis_mount` 会等待 settlement；启动失败时在返回错误前释放 fiber。已 settle 但处于 pending 状态的 Plugin 仍然可见，并列出其缺失的注入。`cordis_unmount` 等待 Plugin fiber 的释放完成。
|
||||
|
||||
临时 Plugin 只存在于进程内存中。它不会创建 Plugin 文件、安装 package、修改 `cordis.yml` 或个人／项目配置、跨重启存续，也不存在自动保存、转正式或安装路径。若要保留实验结果，应让 Agent 通过常规开发流程实现普通的本地、项目或仓库 Plugin。
|
||||
|
||||
### 通过 provide/inject 实现跨挂载组合
|
||||
|
||||
@@ -52,9 +54,9 @@ vm 隔离了意外的全局污染，上下文门面隐藏了框架内部细节
|
||||
|
||||
### 配置、渲染与可观测性
|
||||
|
||||
该插件暴露一个配置字段，由 schemastery 校验并记录在[配置目录](../../../../docs/config-catalog.md)中：`vmTimeoutMs`（默认 5000），挂载代码同步执行部分的毫秒上限。工具名、`cordis-dynamic` 分组名和 `dyn-` id 前缀是结构性词汇，保持固定。三个工具均按[工具实操手册](../../../../docs/cookbook/adding-a-tool.md)渲染为 `generic` 卡片（`cordis_inspect` 为 `read`，`cordis_mount` 为 `execute` 并将代码作为 `rawInput` 携带，`cordis_unmount` 为 `delete`），不覆盖 `presentResult`。
|
||||
该插件暴露一个配置字段，由 schemastery 校验并记录在[配置目录](../../../../docs/config-catalog.md)中：`vmTimeoutMs`（默认 5000），代码同步求值部分的毫秒上限。当前面向模型的名称是 `cordis_inspect`、`cordis_mount` 和 `cordis_unmount`；内部 `cordis-dynamic` 分组名和 `dyn-` id 前缀仍是结构性词汇。三个工具均按[工具实操手册](../../../../docs/cookbook/adding-a-tool.md)渲染为 `generic` 卡片：inspect 为 `read`，mount 为携带代码 `rawInput` 的 `execute`，unmount 为 `delete`。Web 对话行保留这些通用机制，同时为各工具设置操作标题 `Inspect`、`Mount temporary Plugin` 和 `Unmount temporary Plugin` 以及统一的 Cordis 强调色；mount 行仍使用共用的 JavaScript 展开视图和语法高亮。
|
||||
|
||||
「模型可见 ⟺ 已记录」成立，且无需新的会话事件类型：挂载或卸载仅通过其自身的 `tool/call` / `tool/result` 对可见（循环会记录它们），而挂载引起的工具集变化由循环在 schema 在步骤间发生变化时发出的完整变更 request header 记录。刻意不设 `cordis/mount` 溯源事件——它只会重复工具调用对已记录的内容。动态挂载是进程生命周期的，不是会话状态：恢复一个持久化的会话会重建对话，但不会重新挂载插件。
|
||||
「模型可见 ⟺ 已记录」成立，且无需新的会话事件类型：mount 与 unmount 通过已记录的 `tool/call` / `tool/result` 对可见，工具集变化由 schema 在 step 间变化时发出的完整 request header 记录。临时 Plugin 属于进程内存，而非 session 状态：恢复持久化 session 只会重建对话历史，绝不会重新创建它们。
|
||||
|
||||
## 曾考虑的替代方案
|
||||
|
||||
|
||||
@@ -1,6 +0,0 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-09-bash-backed-grep-glob-discovery.md: 9c25afb44885ca2519c5d74a3d721b34fe3561de
|
||||
2026-07-09-bash-backed-grep-glob-discovery.zh.md: d0dee8e49e0a500c87afbd59423fb65416c1dcc8
|
||||
@@ -1,170 +0,0 @@
|
||||
# Agent Note: Bash-backed grep and glob discovery tools
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-07-09-bash-backed-grep-glob-discovery.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
The harness needs model-facing `glob` and `grep` tools, but making them `ctx.fs` provider methods turns a local product convenience into a universal filesystem backend contract. Local workspace discovery is naturally a process-backed `rg` workflow; remote or virtual filesystem backends may expose their own search API, may not share a local `ripgrep` view, or may not support discovery at all. The v1 should not require every filesystem backend to implement search before the file read/write/edit seam has proven that need.
|
||||
|
||||
Search output also has two distinct budgets. The tool needs enough raw `rg` output to compute a stable logical result, but the model should receive only a bounded preview plus a recovery path when the formatted result is larger than the inline budget. The generic spill policy only sees the final tool result, so it cannot recover matches that a search tool already omitted. Search therefore needs tool-owned retention and best-effort formatted-result spill.
|
||||
|
||||
## Decision
|
||||
|
||||
`glob` and `grep` are conditional model-facing tools in `@deepseek-ai/dsh-tool-fs-search`, backed by the bash seam, not by new `ctx.fs` provider methods. At plugin load, the package checks `command -v rg >/dev/null 2>&1` through `ctx.bash.resolve(request)` followed by `ctx.bash.run(spec)`; if the command exits nonzero, the package logs a warning and registers neither tools nor prompt sections. A probe that cannot start, times out, aborts, is killed, or produces no exit code fails plugin load loudly because that is a broken bash executor rather than an absent optional binary. When registered, execution uses the same `ctx.bash.resolve(request)` followed by `ctx.bash.run(spec)` flow with fixed `rg` command templates assembled by the tool. The tool layer owns schemas, argument validation, shell quoting, result parsing, result formatting, retention, formatted-result spill handoff, and timeout declaration. The bash executor owns request defaulting/capping, subprocess execution, process-group termination, environment scrubbing, raw output capture, and backend substitution across local, sandboxed, or remote bash implementations.
|
||||
|
||||
The tools do not use `ctx.bash.start()` and do not create model-visible background tasks. They run as ordinary foreground tools from the agent loop's perspective: the tool call returns only after the `rg` command exits, times out, is aborted, or fails. `defineTool({ timeoutMs })` declares the cooperative tool-call budget, `@deepseek-ai/dsh-timeout-policy` enforces it through `exec.signal`, and the tool forwards that signal into the bash request before `resolve()` / `run()`. The bash backend's own timeout remains a second safety cap; whichever aborts first wins.
|
||||
|
||||
The tools align `path` with Claude Code's search tools while binding resolution to the bash workdir, not to `ctx.fs`. The tool derives the bash request workdir from `exec.agent?.session.header.cwd`, mirroring `dsh-tool-bash` and `dsh-tool-fs`; when no session cwd exists, it omits `request.workdir` so the bash implementation applies its configured cwd or process cwd through `resolve()`. For `grep`, `path` is an optional ripgrep target and may be a file or directory; omitted means the resolved bash workdir. For `glob`, `path` is an optional directory search root; omitted means the resolved bash workdir. Relative `path` values resolve against that workdir. Returned paths are displayed relative to the resolved bash workdir when possible and are intended to be follow-up-readable only in co-located deployments where the bash workdir and filesystem `read` root are the same workspace. v1 documents that deployment requirement but does not perform runtime cross-service validation. Remote or virtual filesystem search is deferred until there is a shared workspace/root contract or a provider-specific search backend.
|
||||
|
||||
The package does not inject `fs`. It injects `tools`, `systemPrompt`, and `bash`; it deliberately reads `spillStore` with `ctx.get('spillStore')` instead of static inject because formatted-result spill is optional. Existing `@deepseek-ai/dsh-tool-fs` deployments that only want `read` / `write` / `edit` do not need to load bash. Deployments that load search need `rg` available in the bash executor environment for the tools to enter the model-visible schema.
|
||||
|
||||
### Package shape
|
||||
|
||||
The v1 package stays small. Inside `@deepseek-ai/dsh-tool-fs-search`, the source layout is:
|
||||
|
||||
```text
|
||||
src/index.ts
|
||||
src/glob.ts
|
||||
src/grep.ts
|
||||
src/search-core.ts
|
||||
src/shell-quote.ts
|
||||
```
|
||||
|
||||
`glob.ts` and `grep.ts` own their parameter validation, command construction, result parsing, formatting, and registration. `shell-quote.ts` is one shared helper because shell quoting is the safety boundary both tools must use; `search-core.ts` is the other (an implementation-time amendment to the original four-file plan): the `SEARCH_*` error vocabulary, the bash-run + raw-output acquisition, the formatted-spill handoff, and workdir-relative display are byte-identical between the two tools, and duplicating that delicate plumbing per tool is exactly the missed extraction the symmetry convention flags. Command builders must not hand-roll quoting or concatenate unquoted model-controlled values into the shell command.
|
||||
|
||||
### Schemas and config
|
||||
|
||||
`glob` exposes the small discovery shape:
|
||||
|
||||
```ts
|
||||
interface GlobArgs {
|
||||
pattern: string
|
||||
path?: string
|
||||
}
|
||||
```
|
||||
|
||||
`grep` exposes the OpenCode-style minimal shape:
|
||||
|
||||
```ts
|
||||
interface GrepArgs {
|
||||
pattern: string
|
||||
path?: string
|
||||
include?: string
|
||||
}
|
||||
```
|
||||
|
||||
Routine budgets stay out of the model-facing schema. `@deepseek-ai/dsh-tool-fs-search` owns these defaulted, validated config fields:
|
||||
|
||||
| Field | Default | Role |
|
||||
|---|---:|---|
|
||||
| `globMaxResults` | `100` | Max paths retained inline; matches Claude Code's default `GlobTool` result limit. |
|
||||
| `grepMaxMatches` | `250` | Max flat matches retained inline; matches Claude Code's default `GrepTool` `head_limit`. |
|
||||
| `grepMaxLineBytes` | `2000` | Max bytes retained for one matched-line preview, applied with `TextRetainer({ kind: 'head', maxBytes: grepMaxLineBytes })`. |
|
||||
| `rawOutputMaxBytes` | `20000000` | Max complete raw `rg` stdout the tool will parse; matches Claude Code's ripgrep raw buffer. |
|
||||
| `timeoutMs` | `30000` | Tool-call timeout attached to both tool definitions and enforced by `@deepseek-ai/dsh-timeout-policy`. |
|
||||
|
||||
`globMaxResults` and `grepMaxMatches` use `ItemRetainer({ kind: 'head' })`. `grepMaxLineBytes` uses `TextRetainer({ kind: 'head', maxBytes: grepMaxLineBytes })` for each matched line so preview cuts preserve UTF-8 boundaries. This follows the [tool result retention library](../architecture/2026-07-06-tool-result-retention-library.md) mapping for discovery items: collect the complete result, retain head items inline, and keep path mapping, grouping, and per-line preview outside the retainer. `grep` does not expose `case_insensitive`, `head_limit`, `offset`, `count`, multiline, context lines, output modes, or file type filters in v1. A model that needs surrounding context reads the matched file with `read`; a model that needs later results follows the returned spill locator's retrieval hint.
|
||||
|
||||
The Claude Code values are reference points for the two-layer budget, not model-facing schema precedent. Its dedicated search tools buffer raw ripgrep output up to 20 MB for internal processing, use a 20-second ripgrep timeout on non-WSL platforms (60 seconds on WSL), then apply search-specific caps before the model sees a result: `GrepTool` defaults to `head_limit = 250` and persists formatted results above 20,000 characters, while `GlobTool` defaults to 100 paths and persists formatted results above 100,000 characters. This Agent Note mirrors the raw-buffer and inline-count defaults, chooses a 30-second default search timeout, and uses this harness's `ctx.spillStore.saveText()` path for formatted-result recovery.
|
||||
|
||||
The `path` field follows the same split as Claude Code: `grep.path` is a file-or-directory ripgrep target, while `glob.path` is a directory search root. v1 does not expose a separate cwd/workdir argument on these tools.
|
||||
|
||||
`include` is one positive glob filter, not a list and not an exclude syntax. Reject comma-separated or negated include patterns up front with a structured argument error. Every model-controlled value used in a shell command, including `pattern`, `path`, and `include`, must pass through the package-private shell quoting helper.
|
||||
|
||||
### Execution
|
||||
|
||||
`glob` builds a fixed `rg --files` command rooted at the resolved directory search root (`path` when supplied, else the bash workdir): `rg --files --glob <pattern> --sort=modified --no-ignore --hidden`, plus VCS metadata excludes for `.git`, `.svn`, `.hg`, `.bzr`, `.jj`, and `.sl`. This aligns with Claude Code on hidden/ignored-file discovery and modified-time ordering while keeping VCS internals out of broad searches. The tool parses one path per line, maps results back to paths relative to the bash workdir when possible, pushes each path into `ItemRetainer({ kind: 'head', maxItems: globMaxResults })`, and formats the full sorted path list for a spill artifact when the retained result is capped.
|
||||
|
||||
`grep` builds a fixed line-oriented `rg --json` command against the supplied file/directory target (`path` when supplied, else the bash workdir) so file path, line number, and line text are parsed without colon-splitting ambiguity. It consumes `match` records, treats malformed JSON or malformed match records as `SEARCH_FAILED`, maps result paths relative to the bash workdir when possible, applies per-line preview retention with `grepMaxLineBytes`, pushes each match into `ItemRetainer({ kind: 'head', maxItems: grepMaxMatches })`, then groups only the retained preview matches by file for inline output. The spill artifact stores the full formatted match list, not only the omitted tail, so the retrieval hint points at the same logical result the model saw.
|
||||
|
||||
Raw `rg` stdout is an internal transport detail. The tool requests `stdoutMaxBytes: rawOutputMaxBytes` through `ctx.bash.resolve()` and parses `stdout.text` only when the executor returns untruncated stdout within that cap. If stdout is larger than `rawOutputMaxBytes`, or the executor still returns `stdout.truncated`, the tool fails with a clear search error telling the model to narrow `pattern`, `path`, or `include`. The tool never exposes raw `rg` output or bash raw spill paths to the model.
|
||||
|
||||
Only stdout is a parse source. Stderr is diagnostic text for invalid patterns, runtime `rg` disappearance after registration, and search failures; if bash truncates stderr, the tool uses the retained stderr tail with a truncation note and does not read `stderr.spillPath`.
|
||||
|
||||
If `ctx.bash.run()` reports `aborted` because the tool timeout or caller cancellation fired, the tool returns a structured failure rather than pretending there were no matches. If bash reports its own timeout first, the tool likewise fails with a clear timeout message. Nonzero ripgrep exit semantics are tool-owned: exit 0 is success with matches, exit 1 is success with no matches, invalid pattern / runtime `rg` disappearance / inaccessible search workdir are failures.
|
||||
|
||||
Search failures use a package-owned `HarnessError` subclass with `SEARCH_*` codes, not `FsErrorCode`, because these tools are not `ctx.fs` provider operations. The v1 vocabulary is `SEARCH_INVALID_PATTERN`, `SEARCH_FAILED`, `SEARCH_RAW_OUTPUT_OVERFLOW`, and `SEARCH_ABORTED`. Model argument validation failures such as missing required fields, blank strings, or unsupported negated/list `include` values remain ordinary tool argument errors.
|
||||
|
||||
### Formatted result spill
|
||||
|
||||
`ctx.spillStore` is optional and used only for model-facing formatted results. This is the first tool-owned spill call pattern in the codebase, and it is intentional because search retention is item-level policy: `globMaxResults` caps paths and `grepMaxMatches` caps matches while the tool still holds the complete logical result. The generic `dsh-spill-policy` caps final text bytes on `tools/post-execute`; by then a search tool would already have omitted later paths or matches, so the policy cannot recover them.
|
||||
|
||||
When a search produces more logical results than the inline cap and `ctx.spillStore` is present, the tool saves the complete formatted result with `saveText()`. The spill owner is the calling agent's session header id (`exec.agent?.session.header.id`); without that owner, the search keeps the inline result and reports that the complete result could not be saved. The spill source is the tool execution identity: `{ toolName: exec.name, callId: exec.callId, label: 'result' }`. The suggested filenames are `grep-results.txt` and `glob-results.txt`; the spill backend still treats them as hints, never paths.
|
||||
|
||||
When spill storage is absent, the call has no session owner, or saving fails, the tool still returns the inline page and a footer explaining that the complete result could not be saved. Search success must not turn into an `isError` result solely because formatted-result spill storage is unavailable.
|
||||
|
||||
The bash raw output stream and the formatted search spill artifact are different artifacts. Raw `rg` stdout is parsed only in memory within the requested bash stdout cap; the formatted spill artifact is the stable model-facing recovery locator produced by `ctx.spillStore.saveText()`.
|
||||
|
||||
### Result shape
|
||||
|
||||
A capped `glob` result with successful formatted spill returns the inline page and a spill notice:
|
||||
|
||||
```text
|
||||
<first N paths>
|
||||
|
||||
(Showing N of M paths. Full sorted result stored at: /.../session-abc123/9f8e7d-glob-results.txt. Use read with offset/limit, or grep this path to search within it.)
|
||||
```
|
||||
|
||||
A capped `grep` result with successful formatted spill returns grouped preview matches and a spill notice:
|
||||
|
||||
```text
|
||||
Found N of M matches
|
||||
|
||||
<file>
|
||||
Line 12: ...
|
||||
|
||||
(Full grep result stored at: /.../session-abc123/9f8e7d-grep-results.txt. Use read with offset/limit, or grep this path to search within it.)
|
||||
```
|
||||
|
||||
If the complete logical result fits under the inline cap, no formatted spill artifact is created. If the complete logical result is too large but formatted spill is unavailable, the footer says that the result was capped and the complete result could not be saved. The `truncated` / omitted count is a budget fact, not an incomplete-search fact; timeout, invalid regex, runtime `rg` disappearance, inaccessible workdirs, raw-output overflow, binary skips, and parse failures stay in tool-domain error or incomplete fields.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
**Put `glob` / `grep` on `ctx.fs`.** Rejected for v1: it forces every filesystem backend to grow a search API and makes local ripgrep behavior part of the provider seam. Search is useful product behavior, but it is not a universal text-storage primitive like `readText` or `writeText`.
|
||||
|
||||
**Directly spawn ripgrep from `dsh-fs-local`.** Rejected for this Agent Note's v1: direct spawn gives the cleanest argv boundary, stdout/stderr control, and early-stop control, but it duplicates process execution concerns that the bash seam already owns: environment scrubbing, process-group kill, timeout propagation, sandbox/remote executor substitution, and bounded output capture. It remains a reasonable optimization if bash-backed search proves too shell-string-sensitive or if foreground streaming becomes necessary.
|
||||
|
||||
**Use `ctx.bash.start()` for streaming early stop.** Rejected: `start()` creates model-visible background task semantics: task ids, owner tokens, `bash_output`, `bash_kill`, completion notifications, and no built-in timeout. `grep` needs a foreground tool result, not a background bash workflow. If streaming search becomes necessary, the right abstraction is a foreground streaming process handle on the bash/process seam, not borrowing the public background-task API.
|
||||
|
||||
**Expose bash raw spill paths to the model.** Rejected: a bash raw spill path contains raw `rg` stdout (`rg --json` records for grep), not the stable formatted search result. Search parses raw stdout only as an internal transport; model recovery uses a formatted result saved through `ctx.spillStore.saveText()`.
|
||||
|
||||
**Add `spillStore.saveFile()` for bash output normalization first.** Rejected for this Agent Note's v1: `saveFile()` would help a future bash normalization pass move existing executor spill files into session-scoped spill storage, but search only needs bounded in-memory raw `rg` stdout before producing the model-facing artifact. `saveText()` is sufficient for the formatted search result.
|
||||
|
||||
**Rely on the generic `dsh-spill-policy`.** Rejected: generic post-execute spill sees only the final tool result. If `grep` / `glob` return the first page inline, the generic policy cannot recover omitted results. The search tools must save the complete formatted result themselves before returning the bounded model-facing text.
|
||||
|
||||
**Expose Claude Code's full `GrepTool` schema.** Rejected for v1: `output_mode`, context flags, multiline, `head_limit`, `offset`, `case_insensitive`, and type filters make the model-facing surface into a ripgrep wrapper. This harness keeps routine budgets and continuation mechanics in deployment policy and spill artifacts.
|
||||
|
||||
**Keep early-stop search and skip formatted spill artifacts.** Rejected for this proposal: early stop is more efficient but gives the model no path to inspect later results. The chosen v1 optimizes result recoverability and implementation simplicity, with `timeoutMs`, `rawOutputMaxBytes`, bash backend caps, and formatted spill artifacts as safety backstops.
|
||||
|
||||
**Expand the bash seam with a raw-output reader first.** Rejected: a portable `readRawOutput(ref, maxBytes)` API would add reference lifetime, permission, and backend storage semantics. A per-run `stdoutMaxBytes` request is the narrower seam: search either receives complete stdout within `rawOutputMaxBytes` or fails clearly.
|
||||
|
||||
**Always register and report missing `rg` only at execution time.** Rejected: a model-visible tool schema is a promise that the deployment can attempt that capability. If the bash executor cannot find ripgrep at load, the safer surface is no `glob` / `grep` tools or prompt guidance. Execution-time missing-`rg` classification remains as a defensive fallback for environments that change after registration.
|
||||
|
||||
## Testing
|
||||
|
||||
- Tests cover registration-time `rg` probing (probe success registers both tools and prompt sections, nonzero probe skips both tools and prompt sections with a warning, infrastructure probe failures reject plugin load), prove an aborted `exec.signal` reaches the bash backend (same-reference spec assertion plus the `SEARCH_ABORTED` result), and cover command construction/quoting (malicious patterns, paths with spaces, leading-dash values, quotes, newlines, glob metacharacters — unit assertions plus a real `bash -c` round-trip for every hostile value), `grep.path` as file and directory targets, `glob.path` as a directory search root, invalid pattern handling, no matches, malformed `rg --json` output, matched-line preview truncation, raw-output overflow, timeout/abort, formatted spill success/failure, the package-owned `SEARCH_*` error codes, and the no-background-task invariant.
|
||||
- The first-party tool-owned spill precedent is covered directly: spill backend present, spill backend absent, `saveText()` failure, and missing spill owner.
|
||||
- The package has real Loader-path coverage for the namespace plugin export shape (`name`, `inject`, `Config`, and `apply`, with no default export).
|
||||
- A real-executor integration suite (`dsh-bash-local` + a real `rg`) verifies the world: hostile patterns stay inert, per-session cwd resolution, VCS-metadata exclusion, modification-time ordering, and real ripgrep stderr classification. It self-skips where `rg` is not on the test process PATH (a CI accommodation mirroring the keyless e2e skip); the fake-executor suite carries registration and execution coverage for missing `rg`, plus the per-file 100% coverage gate.
|
||||
- Snapshot gap note for the transcript-visible spill notice: this landed with the gap note, not a snapshot. The snapshot tier replays the acp-agent tree, and adding the search plugin there changes the assembled system prompt — every expected output would need re-recording with a real key, which the implementing environment did not hold. The spill notice's exact transcript text is pinned by unit tests (`formatGlobOutput`/`formatGrepOutput` and the through-the-registry spill tests); wiring the plugin into the acp-agent tree plus a `test:snapshot:record` pass is the follow-up for the next key-holding session.
|
||||
|
||||
## Consequences
|
||||
|
||||
- `glob` and `grep` are conditional model-facing tools in `@deepseek-ai/dsh-tool-fs-search`, not `ctx.fs` provider methods and not part of the existing `@deepseek-ai/dsh-tool-fs` root plugin. They register only when the bash executor can find `rg`; the package injects `tools`, `systemPrompt`, and `bash`, does not inject `fs`, and keeps `ctx.spillStore` optional via `ctx.get('spillStore')`.
|
||||
- The schemas are exactly `glob(pattern, path?)` and `grep(pattern, path?, include?)`; search caps and timeout are defaulted, validated Config fields (`globMaxResults`, `grepMaxMatches`, `grepMaxLineBytes`, `rawOutputMaxBytes`, `timeoutMs`).
|
||||
- The tools execute through `ctx.bash.resolve(request)` → `ctx.bash.run(spec)`, forward `exec.signal`, never call `ctx.bash.start()`, and never expose a bash task id. The bash request workdir comes from `exec.agent?.session.header.cwd` when available; the resolved `spec.workdir` drives execution and relative-path display.
|
||||
- The tools request `stdoutMaxBytes: rawOutputMaxBytes` from the bash seam, parse only untruncated stdout within that cap, and treat over-cap or still-truncated raw output as a clear search failure; raw `rg` output is never exposed to the model.
|
||||
- Oversized complete formatted results are saved through `ctx.spillStore.saveText()` when available while inline results stay bounded; spill failure, a missing backend, or a missing owner preserves the inline result and reports the unsaved remainder — never an `isError`.
|
||||
- The package README, the generated config catalog, and exported JSDoc document the Config fields and `SEARCH_*` codes; the tui-agent example ships the conditional tool plugin (the acp-agent tree waits on the snapshot re-record above); the fs group README records the `rg` availability and co-located bash/filesystem deployment requirements.
|
||||
|
||||
## Risks
|
||||
|
||||
Full-run `grep` can be slower than an early-stop search on broad patterns. The v1 accepts that cost for simpler implementation and complete-result recovery, bounded by tool timeout, bash timeout, `rawOutputMaxBytes`, and output caps. If this proves too slow, the direct-ripgrep or foreground-streaming alternatives remain available.
|
||||
|
||||
Shell command construction is the sharpest safety edge. Because `ctx.bash` accepts a command string rather than an argv vector, the implementation must centralize shell quoting and test malicious patterns, paths with spaces, leading-dash patterns, quotes, newlines, and glob metacharacters.
|
||||
|
||||
The v1 assumes a co-located bash/filesystem deployment. If bash searches one workspace and the `read` tool resolves paths against another, returned paths may not be follow-up-readable. The package documents this requirement but does not verify it at runtime.
|
||||
|
||||
Spill locators are backend-owned. The current local backend returns local filesystem paths and works in deployments where `read`/`grep` can open those files; remote or workspace-confined deployments can use a backend whose locator and retrieval hint point at a supported retrieval mechanism.
|
||||
@@ -1,170 +0,0 @@
|
||||
# Agent Note: 由 Bash 支持的 grep 与 glob 发现工具
|
||||
|
||||
Status: implemented
|
||||
|
||||
[English](2026-07-09-bash-backed-grep-glob-discovery.md) | 中文
|
||||
|
||||
## 问题
|
||||
|
||||
harness 需要面向模型的 `glob` 和 `grep` 工具，但如果将它们实现为 `ctx.fs` 提供方的方法，就会把本地产品便利功能变成所有文件系统后端都必须实现的契约。本地工作区发现天然适合由进程支持的 `rg` 工作流；远程或虚拟文件系统后端可能公开自己的搜索 API，可能无法共享本地 `ripgrep` 视图，也可能完全不支持发现。文件读取／写入／编辑 seam 尚未证明此需求前，v1 不应要求每个文件系统后端都实现搜索。
|
||||
|
||||
搜索输出还受到两层不同的预算约束。工具需要足够多的原始 `rg` 输出，才能计算稳定的逻辑结果；模型则只能收到有界预览，并在格式化结果超出内联预算时获得恢复路径。通用落盘策略只能看到最终工具结果，因此无法恢复搜索工具已经省略的匹配项。搜索工具必须自行负责保留，并尽力落盘格式化结果。
|
||||
|
||||
## 决策
|
||||
|
||||
`glob` 和 `grep` 是 `@deepseek-ai/dsh-tool-fs-search` 中的条件式面向模型工具，由 bash seam 支持，不会成为新的 `ctx.fs` 提供方方法。加载插件时，该包（package）执行 `command -v rg >/dev/null 2>&1`：先通过 `ctx.bash.resolve(request)` 解析请求，再通过 `ctx.bash.run(spec)` 运行；如果命令以非零状态退出，该包会记录警告，并且既不注册工具，也不注册提示词章节。如果探针无法启动、超时、中止、被终止，或没有产生退出码，插件加载会明确失败，因为这意味着 bash 执行器损坏，而不是可选二进制文件缺失。注册后，执行流程同样依次调用 `ctx.bash.resolve(request)` 与 `ctx.bash.run(spec)`，并使用工具组装的固定 `rg` 命令模板。工具层负责 schema、参数验证、shell 引用、结果解析、结果格式化、保留、格式化结果落盘交接，以及超时声明。bash 执行器负责请求默认值解析与上限控制、子进程执行、进程组终止、环境清理、原始输出捕获，以及在本地、沙箱或远程 bash 实现之间替换后端。
|
||||
|
||||
这些工具不使用 `ctx.bash.start()`，也不创建模型可见的后台任务。从 agent loop（智能体循环）的视角看，它们是普通前台工具：只有当 `rg` 命令退出、超时、中止或失败后，工具调用才返回。`defineTool({ timeoutMs })` 声明协作式工具调用预算，`@deepseek-ai/dsh-timeout-policy` 通过 `exec.signal` 强制执行；工具会在 `resolve()`／`run()` 前将该信号转发给 bash 请求。bash 后端自身的超时仍作为第二层安全上限；先触发的中止生效。
|
||||
|
||||
这些工具使 `path` 与 Claude Code 的搜索工具保持一致，但将解析绑定到 bash workdir，而不是 `ctx.fs`。工具从 `exec.agent?.session.header.cwd` 派生 bash 请求 workdir，与 `dsh-tool-bash` 和 `dsh-tool-fs` 一致；如果会话没有 cwd，它会省略 `request.workdir`，由 bash 实现通过 `resolve()` 应用其配置的 cwd 或进程 cwd。对于 `grep`，`path` 是可选的 ripgrep 目标，可以是文件或目录；省略时使用已解析的 bash workdir。对于 `glob`，`path` 是可选的目录搜索根；省略时同样使用已解析的 bash workdir。相对 `path` 值基于该 workdir 解析。只要可行，返回路径就会显示为相对于已解析 bash workdir 的形式；只有在共置部署中，bash workdir 与文件系统 `read` 根指向同一个工作区时，这些路径才保证可以继续读取。v1 会记录这项部署要求，但不执行跨服务运行时验证。在形成共享工作区／根契约或提供方专用搜索后端之前，远程或虚拟文件系统搜索保持暂缓。
|
||||
|
||||
该包不注入 `fs`，而是注入 `tools`、`systemPrompt` 和 `bash`；它有意读取 `spillStore` 时使用 `ctx.get('spillStore')`，而不使用静态注入，因为格式化结果落盘是可选功能。现有 `@deepseek-ai/dsh-tool-fs` 部署若只需要 `read`／`write`／`edit`，则无需加载 bash。加载搜索功能的部署则必须让 bash 执行器环境可以使用 `rg`，这些工具才会进入模型可见 schema。
|
||||
|
||||
### 包结构
|
||||
|
||||
v1 包保持精简。`@deepseek-ai/dsh-tool-fs-search` 内部的源代码布局如下：
|
||||
|
||||
```text
|
||||
src/index.ts
|
||||
src/glob.ts
|
||||
src/grep.ts
|
||||
src/search-core.ts
|
||||
src/shell-quote.ts
|
||||
```
|
||||
|
||||
`glob.ts` 和 `grep.ts` 各自负责参数验证、命令构造、结果解析、格式化和注册。`shell-quote.ts` 是一个共享辅助模块，因为 shell 引用是两个工具都必须经过的安全边界；`search-core.ts` 是另一个共享模块（实现时对原四文件方案所作的修订）：`SEARCH_*` 错误词汇、bash 运行与原始输出获取、格式化结果落盘交接，以及 workdir 相对显示，在两个工具中完全相同。若在每个工具中重复这套精细管道，正是对称性约定所指出的漏提取问题。命令构造器禁止自行拼凑引用，也不能把未经引用、由模型控制的值直接连接到 shell 命令中。
|
||||
|
||||
### Schema 与配置
|
||||
|
||||
`glob` 公开精简的发现形状：
|
||||
|
||||
```ts
|
||||
interface GlobArgs {
|
||||
pattern: string
|
||||
path?: string
|
||||
}
|
||||
```
|
||||
|
||||
`grep` 公开 OpenCode 风格的最小形状：
|
||||
|
||||
```ts
|
||||
interface GrepArgs {
|
||||
pattern: string
|
||||
path?: string
|
||||
include?: string
|
||||
}
|
||||
```
|
||||
|
||||
常规预算不会进入面向模型的 schema。`@deepseek-ai/dsh-tool-fs-search` 拥有以下带默认值并经过验证的配置字段：
|
||||
|
||||
| 字段 | 默认值 | 作用 |
|
||||
|---|---:|---|
|
||||
| `globMaxResults` | `100` | 内联保留的最大路径数；与 Claude Code 的默认 `GlobTool` 结果上限一致。 |
|
||||
| `grepMaxMatches` | `250` | 内联保留的最大扁平匹配数；与 Claude Code 的默认 `GrepTool` `head_limit` 一致。 |
|
||||
| `grepMaxLineBytes` | `2000` | 每条匹配行预览保留的最大字节数，通过 `TextRetainer({ kind: 'head', maxBytes: grepMaxLineBytes })` 应用。 |
|
||||
| `rawOutputMaxBytes` | `20000000` | 工具会解析的完整原始 `rg` stdout 最大字节数；与 Claude Code 的 ripgrep 原始缓冲区一致。 |
|
||||
| `timeoutMs` | `30000` | 附加到两个工具定义并由 `@deepseek-ai/dsh-timeout-policy` 强制执行的工具调用超时。 |
|
||||
|
||||
`globMaxResults` 和 `grepMaxMatches` 使用 `ItemRetainer({ kind: 'head' })`。`grepMaxLineBytes` 针对每条匹配行使用 `TextRetainer({ kind: 'head', maxBytes: grepMaxLineBytes })`，使预览截断保留 UTF-8 边界。这遵循[工具结果保留库](../architecture/2026-07-06-tool-result-retention-library.md)对发现条目的映射：收集完整结果，在内联结果中保留头部条目，并将路径映射、分组和逐行预览放在保留器外部。v1 的 `grep` 不公开 `case_insensitive`、`head_limit`、`offset`、`count`、多行、上下文行、输出模式或文件类型过滤器。模型如需周边上下文，可使用 `read` 读取匹配文件；如需后续结果，则遵循返回的落盘定位符所给出的检索提示。
|
||||
|
||||
Claude Code 的数值只是两层预算的参考点，并非面向模型 schema 的先例。其专用搜索工具会缓冲最多 20 MB 的原始 ripgrep 输出用于内部处理；在非 WSL 平台上使用 20 秒 ripgrep 超时，在 WSL 上使用 60 秒，之后才在模型看到结果前应用搜索专用上限：`GrepTool` 默认 `head_limit = 250`，并持久化超过 20,000 个字符的格式化结果；`GlobTool` 默认最多 100 条路径，并持久化超过 100,000 个字符的格式化结果。此 Agent Note 采用同样的原始缓冲区和内联数量默认值，将默认搜索超时设为 30 秒，并通过本 harness 的 `ctx.spillStore.saveText()` 路径恢复格式化结果。
|
||||
|
||||
`path` 字段沿用与 Claude Code 相同的区分方式：`grep.path` 是文件或目录形式的 ripgrep 目标，`glob.path` 则是目录搜索根。v1 不为这些工具公开单独的 cwd／workdir 参数。
|
||||
|
||||
`include` 是一个正向 glob 过滤器，不是列表，也不是排除语法。系统会预先拒绝逗号分隔或取反的 include 模式，并返回结构化参数错误。shell 命令中使用的每个模型控制值，包括 `pattern`、`path` 和 `include`，都必须经过包私有的 shell 引用辅助模块。
|
||||
|
||||
### 执行
|
||||
|
||||
`glob` 构建固定的 `rg --files` 命令，并以解析后的目录搜索根为根（提供 `path` 时使用该值，否则使用 bash workdir）：`rg --files --glob <pattern> --sort=modified --no-ignore --hidden`，另加针对 `.git`、`.svn`、`.hg`、`.bzr`、`.jj` 和 `.sl` 的 VCS 元数据排除项。这样既与 Claude Code 的隐藏／忽略文件发现和修改时间排序保持一致，也避免宽泛搜索包含 VCS 内部文件。工具逐行解析路径，只要可行就将结果映射为相对于 bash workdir 的路径，将每条路径推入 `ItemRetainer({ kind: 'head', maxItems: globMaxResults })`；当保留结果达到上限时，它会格式化完整的已排序路径列表，作为落盘产物。
|
||||
|
||||
`grep` 构建固定的逐行 `rg --json` 命令，作用于所提供的文件／目录目标（提供 `path` 时使用该值，否则使用 bash workdir），从而无需按冒号拆分，就能解析文件路径、行号和行文本。它消费 `match` 记录，将格式错误的 JSON 或匹配记录视为 `SEARCH_FAILED`；只要可行，就将结果路径映射为相对于 bash workdir 的路径；通过 `grepMaxLineBytes` 应用逐行预览保留，将每个匹配推入 `ItemRetainer({ kind: 'head', maxItems: grepMaxMatches })`，然后只按文件分组内联输出中保留的预览匹配。落盘产物保存完整的格式化匹配列表，而不是只保存省略的尾部，因此检索提示指向模型已经看到的同一逻辑结果。
|
||||
|
||||
原始 `rg` stdout 是内部传输细节。工具请求 `stdoutMaxBytes: rawOutputMaxBytes`，并通过 `ctx.bash.resolve()` 解析；只有当执行器在该上限内返回未截断的 stdout 时，工具才解析 `stdout.text`。如果 stdout 超过 `rawOutputMaxBytes`，或者执行器仍返回 `stdout.truncated`，工具会以明确的搜索错误失败，要求模型缩小 `pattern`、`path` 或 `include`。工具绝不向模型公开原始 `rg` 输出或 bash 原始落盘路径。
|
||||
|
||||
只有 stdout 是解析源。对于无效模式、注册后运行时 `rg` 消失，以及搜索失败，stderr 作为诊断文本；如果 bash 截断 stderr，工具会使用保留的 stderr 尾部并附加截断说明，不会读取 `stderr.spillPath`。
|
||||
|
||||
如果 `ctx.bash.run()` 因工具超时或调用方取消触发而报告 `aborted`，工具会返回结构化失败，而不是假装没有匹配项。如果 bash 自身的超时先触发，工具同样会以明确的超时消息失败。非零 ripgrep 退出语义由工具负责：退出码 0 表示存在匹配并成功；退出码 1 表示没有匹配但成功；无效模式、运行时 `rg` 消失或无法访问搜索 workdir 则表示失败。
|
||||
|
||||
搜索失败使用包自有的 `HarnessError` 子类与 `SEARCH_*` 代码，而不使用 `FsErrorCode`，因为这些工具不是 `ctx.fs` 提供方操作。v1 的词汇包括 `SEARCH_INVALID_PATTERN`、`SEARCH_FAILED`、`SEARCH_RAW_OUTPUT_OVERFLOW` 和 `SEARCH_ABORTED`。缺少必填字段、空字符串或不支持的取反／列表式 `include` 值等模型参数验证失败，仍作为普通工具参数错误处理。
|
||||
|
||||
### 格式化结果落盘
|
||||
|
||||
`ctx.spillStore` 是可选服务，仅用于面向模型的格式化结果。这是代码库中首个工具自有落盘调用模式；此设计有意为之，因为搜索保留属于条目级策略：`globMaxResults` 限制路径数，`grepMaxMatches` 限制匹配数，而工具此时仍持有完整逻辑结果。通用 `dsh-spill-policy` 会在 `tools/post-execute` 阶段限制最终文本字节数；到那时搜索工具已经省略后续路径或匹配，策略无法恢复它们。
|
||||
|
||||
当搜索产生的逻辑结果数超过内联上限，且 `ctx.spillStore` 存在时，工具会通过 `saveText()` 保存完整的格式化结果。落盘所有者是调用 agent 的会话头 id（`exec.agent?.session.header.id`）；缺少该所有者时，搜索会保留内联结果，并报告完整结果无法保存。落盘来源是工具执行身份：`{ toolName: exec.name, callId: exec.callId, label: 'result' }`。建议文件名为 `grep-results.txt` 和 `glob-results.txt`；落盘后端仍将它们视为提示，而不是路径。
|
||||
|
||||
如果落盘存储不存在、调用没有会话所有者，或保存失败，工具仍返回内联页和页脚，说明完整结果无法保存。格式化结果落盘存储不可用本身绝不能把搜索成功变为 `isError` 结果。
|
||||
|
||||
bash 原始输出流与格式化搜索落盘产物是两个不同的产物。原始 `rg` stdout 只会在所请求的 bash stdout 上限内于内存中解析；格式化落盘产物则是 `ctx.spillStore.saveText()` 生成的稳定、面向模型的恢复定位符。
|
||||
|
||||
### 结果形状
|
||||
|
||||
带有成功格式化落盘的受限 `glob` 结果会返回内联页与落盘通知：
|
||||
|
||||
```text
|
||||
<first N paths>
|
||||
|
||||
(Showing N of M paths. Full sorted result stored at: /.../session-abc123/9f8e7d-glob-results.txt. Use read with offset/limit, or grep this path to search within it.)
|
||||
```
|
||||
|
||||
带有成功格式化落盘的受限 `grep` 结果会返回分组后的预览匹配与落盘通知：
|
||||
|
||||
```text
|
||||
Found N of M matches
|
||||
|
||||
<file>
|
||||
Line 12: ...
|
||||
|
||||
(Full grep result stored at: /.../session-abc123/9f8e7d-grep-results.txt. Use read with offset/limit, or grep this path to search within it.)
|
||||
```
|
||||
|
||||
如果完整逻辑结果未超过内联上限，系统不会创建格式化落盘产物。如果完整逻辑结果过大但无法格式化落盘，页脚会说明结果已受限，完整结果无法保存。`truncated`／省略计数是预算事实，并不表示搜索不完整；超时、无效正则表达式、运行时 `rg` 消失、无法访问 workdir、原始输出溢出、跳过二进制文件和解析失败，仍属于工具领域的错误或不完整字段。
|
||||
|
||||
## 考虑过的替代方案
|
||||
|
||||
**将 `glob`／`grep` 放在 `ctx.fs` 上。** v1 不采用：这会迫使每个文件系统后端增加搜索 API，并使本地 ripgrep 行为成为提供方 seam 的一部分。搜索是有用的产品行为，但不像 `readText` 或 `writeText` 那样属于通用文本存储原语。
|
||||
|
||||
**直接从 `dsh-fs-local` spawn ripgrep。** 此 Agent Note 的 v1 不采用：直接 spawn 提供最简洁的 argv 边界、stdout／stderr 控制与提前停止控制，但会重复 bash seam 已负责的进程执行事项，包括环境清理、进程组终止、超时传播、沙箱／远程执行器替换，以及有界输出捕获。如果 bash 支持的搜索被证明过于依赖 shell 字符串，或必须支持前台流式输出，该方案仍是合理优化。
|
||||
|
||||
**通过 `ctx.bash.start()` 实现流式提前停止。** 不采用：`start()` 会创建模型可见的后台任务语义，包括 task id、所有者 token、`bash_output`、`bash_kill`、完成通知，并且没有内置超时。`grep` 需要前台工具结果，而不是后台 bash 工作流。如果将来必须流式搜索，正确的抽象是在 bash／进程 seam 上增加前台流式进程句柄，而不是借用公开后台任务 API。
|
||||
|
||||
**向模型公开 bash 原始落盘路径。** 不采用：bash 原始落盘路径包含原始 `rg` stdout（对于 grep 即 `rg --json` 记录），并非稳定的格式化搜索结果。搜索只把原始 stdout 当作内部传输；模型恢复使用通过 `ctx.spillStore.saveText()` 保存的格式化结果。
|
||||
|
||||
**先为 bash 输出规范化增加 `spillStore.saveFile()`。** 此 Agent Note 的 v1 不采用：未来规范化 bash 时，`saveFile()` 可以帮助将现有执行器落盘文件移动到会话范围的落盘存储，但搜索只需在生成面向模型的产物前，在内存中获取有界的原始 `rg` stdout。`saveText()` 足以保存格式化搜索结果。
|
||||
|
||||
**依赖通用 `dsh-spill-policy`。** 不采用：通用 post-execute 落盘只能看到最终工具结果。如果 `grep`／`glob` 内联返回第一页，通用策略无法恢复省略的结果。搜索工具必须在返回有界的面向模型文本前，自行保存完整的格式化结果。
|
||||
|
||||
**公开 Claude Code 的完整 `GrepTool` schema。** v1 不采用：`output_mode`、上下文标志、多行、`head_limit`、`offset`、`case_insensitive` 和类型过滤器会使面向模型的接口变成 ripgrep 包装层。本 harness 将常规预算与续传机制保留在部署策略和落盘产物中。
|
||||
|
||||
**保留提前停止搜索，并省略格式化落盘产物。** 此提案不采用：提前停止效率更高，却不给模型检查后续结果的路径。所选 v1 优先保证结果可恢复性与实现简洁性，并以 `timeoutMs`、`rawOutputMaxBytes`、bash 后端上限和格式化落盘产物作为安全后备。
|
||||
|
||||
**先扩展 bash seam，增加原始输出读取器。** 不采用：可移植的 `readRawOutput(ref, maxBytes)` API 会增加引用生命周期、权限和后端存储语义。逐次运行的 `stdoutMaxBytes` 请求是更窄的 seam：搜索要么在 `rawOutputMaxBytes` 内收到完整 stdout，要么明确失败。
|
||||
|
||||
**始终注册，只有执行时才报告缺少 `rg`。** 不采用：模型可见工具 schema 是部署能够尝试该能力的承诺。如果 bash 执行器在加载时找不到 ripgrep，更安全的接口是完全没有 `glob`／`grep` 工具或提示词指引。对于注册后发生环境变化的情况，执行时的 `rg` 缺失分类仍作为防御性回退。
|
||||
|
||||
## 测试
|
||||
|
||||
- 测试覆盖注册时 `rg` 探测（探测成功会注册两个工具和提示词章节；非零探测会跳过工具与提示词章节并发出警告；基础设施探测失败会拒绝插件加载），证明中止的 `exec.signal` 会到达 bash 后端（通过同一引用的 spec 断言和 `SEARCH_ABORTED` 结果），并覆盖命令构造／引用（恶意模式、带空格路径、以短横线开头的值、引号、换行、glob 元字符：既有单元断言，也针对每个恶意值执行真实 `bash -c` 往返）、将 `grep.path` 用作文件与目录目标、将 `glob.path` 用作目录搜索根、无效模式处理、无匹配、格式错误的 `rg --json` 输出、匹配行预览截断、原始输出溢出、超时／中止、格式化落盘成功／失败、包自有 `SEARCH_*` 错误代码，以及无后台任务不变量。
|
||||
- 直接覆盖第一方工具自有落盘先例：落盘后端存在、落盘后端缺失、`saveText()` 失败，以及缺少落盘所有者。
|
||||
- 该包通过真实 Loader 路径覆盖命名空间插件的导出形状（`name`、`inject`、`Config` 和 `apply`，且没有默认导出）。
|
||||
- 真实执行器集成测试（`dsh-bash-local` + 真实 `rg`）验证外部世界：恶意模式保持惰性、逐会话 cwd 解析、VCS 元数据排除、按修改时间排序，以及真实 ripgrep stderr 分类。如果测试进程的 PATH 中没有 `rg`，该测试会自行跳过（这是与无密钥 e2e 跳过相似的 CI 兼容措施）；伪执行器测试覆盖注册和执行时缺少 `rg`，并由逐文件 100% 覆盖率门禁兜底。
|
||||
- transcript 可见落盘通知仍有快照缺口：此功能合入时记录了缺口说明，没有快照。快照层会回放 acp-agent 树；在其中加入搜索插件会改变组装的系统提示词，必须使用真实密钥重新录制每一份预期输出，而实现环境没有密钥。落盘通知的确切 transcript 文本由单元测试固定（`formatGlobOutput`／`formatGrepOutput` 以及通过注册表执行的落盘测试）；下一次拥有密钥的会话应把插件接入 acp-agent 树，并运行一次 `test:snapshot:record`。
|
||||
|
||||
## 后果
|
||||
|
||||
- `glob` 和 `grep` 是 `@deepseek-ai/dsh-tool-fs-search` 中的条件式面向模型工具，不是 `ctx.fs` 提供方方法，也不属于现有 `@deepseek-ai/dsh-tool-fs` 根插件。只有 bash 执行器能找到 `rg` 时才会注册；该包注入 `tools`、`systemPrompt` 和 `bash`，不注入 `fs`，并使 `ctx.spillStore` 保持可选，读取时使用 `ctx.get('spillStore')`。
|
||||
- Schema 严格为 `glob(pattern, path?)` 和 `grep(pattern, path?, include?)`；搜索上限与超时是带默认值并经过验证的 Config 字段（`globMaxResults`、`grepMaxMatches`、`grepMaxLineBytes`、`rawOutputMaxBytes`、`timeoutMs`）。
|
||||
- 工具通过 `ctx.bash.resolve(request)` → `ctx.bash.run(spec)` 执行，转发 `exec.signal`，绝不调用 `ctx.bash.start()`，也绝不公开 bash task id。如果存在 `exec.agent?.session.header.cwd`，bash 请求 workdir 来自该值；解析后的 `spec.workdir` 决定执行与相对路径显示。
|
||||
- 工具向 bash seam 请求 `stdoutMaxBytes: rawOutputMaxBytes`，只解析上限内未截断的 stdout，并将超限或仍被截断的原始输出视为明确的搜索失败；绝不向模型公开原始 `rg` 输出。
|
||||
- 只要可用，过大的完整格式化结果会通过 `ctx.spillStore.saveText()` 保存，而内联结果保持有界；落盘失败、后端缺失或所有者缺失时，系统保留内联结果并报告未保存的剩余内容，绝不会返回 `isError`。
|
||||
- 包 README、生成的配置目录与导出 JSDoc 会记录 Config 字段和 `SEARCH_*` 代码；tui-agent 示例会提供条件式工具插件（acp-agent 树等待完成上述快照重新录制）；fs 组 README 会记录 `rg` 可用性以及 bash／文件系统共置部署要求。
|
||||
|
||||
## 风险
|
||||
|
||||
在宽泛模式下，完整运行的 `grep` 可能比提前停止搜索更慢。v1 为了简化实现并恢复完整结果而接受这项成本，同时通过工具超时、bash 超时、`rawOutputMaxBytes` 和输出上限加以约束。如果实际运行过慢，仍可采用直接 ripgrep 或前台流式替代方案。
|
||||
|
||||
Shell 命令构造是最尖锐的安全边界。`ctx.bash` 接受命令字符串而不是 argv 向量，因此实现必须集中处理 shell 引用，并测试恶意模式、带空格路径、以短横线开头的模式、引号、换行和 glob 元字符。
|
||||
|
||||
v1 假设 bash 与文件系统共置部署。如果 bash 搜索一个工作区，而 `read` 工具基于另一个根解析路径，返回路径可能无法继续读取。该包会记录这项要求，但不在运行时验证。
|
||||
|
||||
落盘定位符由后端负责。当前本地后端返回本地文件系统路径，适用于 `read`／`grep` 能打开这些文件的部署；远程或工作区受限部署可以使用另一种后端，让其定位符和检索提示指向受支持的检索机制。
|
||||
@@ -2,5 +2,5 @@
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-10-agent-session-identity-and-log-location.md: a55bf276bff998a94f84ec1af078022e4881903c
|
||||
2026-07-10-agent-session-identity-and-log-location.zh.md: 84b8b22187ccd1078ff13e39f4a345efbecfaeac
|
||||
2026-07-10-agent-session-identity-and-log-location.md: 7b51ae41ac00c12a496940c891092580003646fa
|
||||
2026-07-10-agent-session-identity-and-log-location.zh.md: 2574e1327f424069cdff68ef9b8de20c490077f0
|
||||
|
||||
@@ -40,7 +40,7 @@ The registry rebuilds a trusted overlay for every foreground and background bash
|
||||
|
||||
Session persistence remains the fact owner: JSONL does not depend on tool-bash or register shell variables itself, and hooks continue to consume `locate()` directly. Tool-bash is the translation layer from the persistence fact into a shell convention. Other plugins that need shell-visible facts depend on the registry and register their own keys; they do not modify `process.env`.
|
||||
|
||||
The bash seam exports `DSH_ENV_PREFIX` as the single namespace source and derives `DshEnvironmentKey` from its `typeof`. Tool-bash derives built-in names and model guidance from that constant, while executors use it for filtering and channel validation. The seam carries the managed overlay separately as `BashExecRequest.dshEnv` / `BashExecSpec.dshEnv`. Ordinary `env` remains the general in-process plugin surface used by hooks, but cannot contain managed keys; symmetrically, `dshEnv` cannot contain ordinary keys. The local executor rejects either wrong channel before spawn, removes every inherited ambient managed key, applies its ordinary scrub/terminal environment/explicit `env`, and finally merges the trusted `dshEnv` snapshot. This guarantees that a missing value means absent now rather than inherited from an outer or previous harness. The model-facing tool still ignores model-supplied `env`/`stdin` arguments.
|
||||
The bash seam exports `DSH_ENV_PREFIX` as the single namespace source and derives `DshEnvironmentKey` from its `typeof`. Tool-bash derives built-in names and model guidance from that constant, while executors use it for ambient filtering. The seam carries the managed overlay separately as `BashExecRequest.dshEnv` / `BashExecSpec.dshEnv`: ordinary `env` remains the general in-process plugin surface used by hooks, while `dshEnv` is typed to managed keys. The local executor removes every inherited ambient managed key, applies its ordinary scrub/terminal environment/explicit `env`, and finally merges the trusted `dshEnv` snapshot, so an `env` entry can never displace a managed value. This guarantees that a missing value means absent now rather than inherited from an outer or previous harness. The model-facing tool still ignores model-supplied `env`/`stdin` arguments.
|
||||
|
||||
The bash tool description teaches only the durable convention: current harness environment facts are available through managed `$DSH_*` variables and may be inspected when needed. It does not enumerate persistence-specific keys or add a permanent system-prompt section. Tool schemas are already logged in request headers and tool output is logged as `tool/result`, so no new session event is required.
|
||||
|
||||
@@ -82,6 +82,6 @@ A keyless full-loop integration drives the real agent loop, JSONL persistence, t
|
||||
|
||||
## Consequences
|
||||
|
||||
Every model bash child receives current Harness home and shell identity, and agent calls additionally receive stable session identity. JSONL-backed calls get an optional target path; non-file persistence omits it honestly. The complete `DSH_*` namespace inside these children is managed by the harness: ambient values are removed, current trusted values are re-added, and ordinary callers cannot use `env` to bypass ownership checks.
|
||||
Every model bash child receives current Harness home and shell identity, and agent calls additionally receive stable session identity. JSONL-backed calls get an optional target path; non-file persistence omits it honestly. The managed `DSH_*` facts inside these children come from the harness: ambient values are removed, current trusted values are re-added last, and an ordinary caller's `env` entry cannot displace them.
|
||||
|
||||
The namespace is discoverable but not secret. Paths can reveal configured roots, lazy targets can be absent or stale, and a command can override variables inside its own shell syntax. Consumers treat them as correlation and environment facts, verify transcript metadata when attribution matters, and rely on sandbox/filesystem policy rather than variable secrecy for authorization.
|
||||
|
||||
@@ -40,7 +40,7 @@ interface SessionPersistence {
|
||||
|
||||
会话持久化仍然是事实所有者：JSONL 不依赖 tool-bash，也不会自行注册 shell 变量；钩子继续直接使用 `locate()`。tool-bash 是把持久化事实转换为 shell 约定的转换层。其他需要向 shell 公开事实的插件依赖该注册表，并注册各自的键；它们不修改 `process.env`。
|
||||
|
||||
bash seam 导出 `DSH_ENV_PREFIX` 作为唯一的命名空间来源，并派生 `DshEnvironmentKey`，其来源是该常量的 `typeof`。tool-bash 从该常量派生内置名称与模型指引，执行器则使用该常量进行过滤和通道校验。seam 通过 `BashExecRequest.dshEnv`／`BashExecSpec.dshEnv` 单独传递受管理的覆盖层。普通 `env` 仍是钩子所用的通用进程内插件接口，但不能包含受管理的键；对称地，`dshEnv` 不能包含普通键。本地执行器会在 spawn 前拒绝任一错误通道，移除环境中继承的全部受管理键，依次应用普通清理、终端环境和显式 `env`，最后合并受信任的 `dshEnv` 快照。这保证了值缺失表示它当前确实不存在，而不是从外层或先前的 harness 继承而来。面向模型的工具仍忽略模型提供的 `env`／`stdin` 参数。
|
||||
bash seam 导出 `DSH_ENV_PREFIX` 作为唯一的命名空间来源，并派生 `DshEnvironmentKey`，其来源是该常量的 `typeof`。tool-bash 从该常量派生内置名称与模型指引，执行器则使用该常量过滤环境中已有的值。seam 通过 `BashExecRequest.dshEnv`／`BashExecSpec.dshEnv` 单独传递受管理的覆盖层：普通 `env` 仍是钩子所用的通用进程内插件接口，`dshEnv` 则以类型约束为受管理键。本地执行器移除环境中继承的全部受管理键，依次应用普通清理、终端环境和显式 `env`，最后合并受信任的 `dshEnv` 快照，因此 `env` 条目永远无法顶掉受管理的值。这保证了值缺失表示它当前确实不存在，而不是从外层或先前的 harness 继承而来。面向模型的工具仍忽略模型提供的 `env`／`stdin` 参数。
|
||||
|
||||
bash 工具说明只讲解持久约定：当前 harness 环境事实通过受管理的 `$DSH_*` 变量提供，可以在需要时查看。它不会枚举持久化专用键，也不会添加永久的系统提示词章节。工具 schema 已记录在请求 header 中，工具输出则记录为 `tool/result`，因此无需新增会话事件。
|
||||
|
||||
@@ -82,6 +82,6 @@ bash 工具说明只讲解持久约定：当前 harness 环境事实通过受管
|
||||
|
||||
## 影响
|
||||
|
||||
每个面向模型的 bash 子进程都会收到当前 Harness home 和 shell 标识，关联 agent 的调用还会收到稳定的会话标识。使用 JSONL 后端的调用可以获得可选的目标路径；非文件持久化会如实省略该值。这些子进程中的完整 `DSH_*` 命名空间由 harness 管理：系统移除环境中已有的受管理值、重新加入当前受信任的值，并禁止普通调用方通过 `env` 绕过所有权检查。
|
||||
每个面向模型的 bash 子进程都会收到当前 Harness home 和 shell 标识，关联 agent 的调用还会收到稳定的会话标识。使用 JSONL 后端的调用可以获得可选的目标路径；非文件持久化会如实省略该值。这些子进程中受管理的 `DSH_*` 事实来自 harness：系统移除环境中已有的受管理值、在最后重新加入当前受信任的值，普通调用方的 `env` 条目无法顶掉它们。
|
||||
|
||||
该命名空间可被发现，但并非秘密。路径可能泄露配置的根目录，延迟创建的目标也可能不存在或处于陈旧状态，而且命令可以在自己的 shell 语法中覆盖变量。消费方应把这些值视为关联信息和环境事实，在归属关系重要时校验 transcript 元数据，并依靠沙箱／文件系统策略而不是变量保密性来完成授权。
|
||||
|
||||
@@ -1,42 +0,0 @@
|
||||
# Agent Note: Exact session query service
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-07-10-session-query-service.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
Session history exists in two places: current `SessionStore` objects and an optional persistence backend. Consumers that need exact inspection would otherwise duplicate live-versus-persisted precedence, persistence lifecycle handling, raw-event surface classification, relationship tracing, and defensive cloning. Durable state can lag the live log between checkpoints, so persistence alone is not a truthful current source.
|
||||
|
||||
Full-text search is related but materially larger. Putting provider coordination, synchronization, invalidation, ranking, and cursor state into the exact-read service would create a second state machine beside the concrete database owner.
|
||||
|
||||
## Decision
|
||||
|
||||
`@deepseek-ai/dsh-session-query` owns the single abstract `ctx.sessionQuery` service over one logical corpus. It concretely implements `listSessions()`, provider-independent `filterSessions(filters)`, `listEvents(sessionId)`, `filterEvents(sessionId, filters)`, bounded `readEvent(request)`, `traceSession(sessionId)`, and `traceEvent(request)`, while concrete backends implement its two full-text methods. The [unified service decision](../../archived/architecture/2026-07-23-unified-session-query-service.md) owns that topology, the [SQLite search decision](2026-07-10-sqlite-session-query-provider.md) owns search behavior, and the [tracing decision](2026-07-13-session-query-tracing.md) owns lineage and event-relationship semantics.
|
||||
|
||||
The service observes the optional `ctx.sessionPersistence` binding dynamically but retains no persisted cache or invalidation listener. Each cross-corpus list asks the active backend for authoritative metadata, then overlays a fresh live-store list. Matching ids become one `SessionRecord`: the live header wins and `live`/`persisted` independently report source availability. Immutable header disagreement is `SESSION_QUERY_SOURCE_CONFLICT`.
|
||||
|
||||
An exact target read first checks the live store and snapshots the live header and event log. This path never consults persistence, so a failing durable backend cannot make known live history unreadable. With no live target, the service lists current persistence metadata, proves the id exists, loads it, and rejects a list/load header mismatch. All returned headers and events cross one structured-clone boundary.
|
||||
|
||||
## Surface semantics
|
||||
|
||||
`dsh-session` exports `foldSurface(events)`, and `SurfaceManager` uses the same transition functions for its incremental cache. The fold returns detached current event sequences and each replacement's actual removed seqs. `listEvents()` and `traceEvent()` use that result to classify every raw event, so inspection cannot disagree with model-history derivation about positional replacement semantics.
|
||||
|
||||
`readEvent()` returns the complete target plus raw neighbors by contiguous seq. `before` and `after` default to zero and are independently bounded by `readWindowMax`, default 50. The result carries a cloned `SessionHeader`, not a source-availability record, because determining a live target's persisted flag would violate the guarantee that live exact reads do not depend on persistence health.
|
||||
|
||||
## Security boundary
|
||||
|
||||
The service is context-wide trusted infrastructure, not an authorization layer. A future model-facing history tool or human UI applies explicit caller/session scope. The service adds no model-facing tool and changes no transcript or snapshot surface.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
- **Put logical-corpus resolution directly in every consumer** — rejected because source precedence, conflicts, optional-service lifecycle, cloning, and surface classification are shared correctness rules.
|
||||
- **Query only persistence** — rejected because checkpoints can lag the current live log.
|
||||
- **Cache persisted metadata and listen for writes/removals** — rejected because exact reads can ask the authoritative sources directly, while cache invalidation adds lifecycle and concurrency state before scale requires it.
|
||||
- **Put provider registration into the exact-read service** — rejected because the SQLite package owns one reconciliation/transaction lifecycle; a registry would split that state without a second provider to justify it.
|
||||
|
||||
## Consequences
|
||||
|
||||
The inherited exact-read implementation has one source-resolution state variable: the currently mounted persistence service. It has no provider queues, fingerprints, extractor registries, observation generations, or derived index updates; a concrete backend owns its full-text state separately. Exact reads, semantic scans, and event traces remain usable in live-only deployments and deterministic when persistence is present.
|
||||
|
||||
Cross-corpus listing, lineage tracing, and persisted event operations perform backend I/O on each call. That is deliberate: correctness comes from current authoritative state, while scale-oriented full-text methods use the concrete backend's SQLite derived index.
|
||||
@@ -1,42 +0,0 @@
|
||||
# Agent Note: 精确会话查询服务
|
||||
|
||||
Status: implemented
|
||||
|
||||
[English](2026-07-10-session-query-service.md) | 中文
|
||||
|
||||
## 问题
|
||||
|
||||
会话历史存在于两处：当前的 `SessionStore` 对象与可选的持久化后端。需要精确检查的消费方若无统一服务，就不得不各自重复实现活跃/持久化优先级判定、持久化生命周期处理、原始事件的 surface 分类、关系追踪以及防御性克隆。在检查点之间，持久化状态可能落后于活跃日志，因此仅靠持久化并非当前状态的可靠来源。
|
||||
|
||||
全文搜索与此相关，但规模大得多。将提供方协调、同步、失效、排序和游标状态放入精确读取服务，会在具体数据库拥有方旁边再创建一个状态机。
|
||||
|
||||
## 决策
|
||||
|
||||
`@deepseek-ai/dsh-session-query` 拥有面向单一逻辑语料库的唯一抽象 `ctx.sessionQuery` 服务。它具体实现 `listSessions()`、提供方无关的 `filterSessions(filters)`、`listEvents(sessionId)`、`filterEvents(sessionId, filters)`、有界的 `readEvent(request)`、`traceSession(sessionId)` 和 `traceEvent(request)`，而具体后端实现其两个全文搜索方法。[统一服务决策](../../archived/architecture/2026-07-23-unified-session-query-service.md)拥有这一拓扑，[SQLite 搜索决策](2026-07-10-sqlite-session-query-provider.md)拥有搜索行为，[追踪决策](2026-07-13-session-query-tracing.md)拥有血缘与事件关系语义。
|
||||
|
||||
该服务动态观察可选的 `ctx.sessionPersistence` 绑定，但不保留持久化缓存或失效监听器。每次跨语料库列表操作向活跃后端请求权威元数据，然后叠加一份新鲜的活跃 store 列表。id 匹配的条目合并为一条 `SessionRecord`：活跃 header 优先，`live`/`persisted` 各自独立报告来源可用性。不可变 header 不一致时产生 `SESSION_QUERY_SOURCE_CONFLICT`。
|
||||
|
||||
精确目标读取首先检查活跃 store，快照活跃 header 与事件日志。此路径从不查询持久化，因此持久化后端故障不会导致已知的活跃历史不可读。若活跃 store 中无目标，服务列出当前持久化元数据、证明该 id 存在、加载它，并在列表/加载 header 不一致时拒绝。所有返回的 header 与事件都经过一次 structured-clone 边界。
|
||||
|
||||
## Surface 语义
|
||||
|
||||
`dsh-session` 导出 `foldSurface(events)`，`SurfaceManager` 使用相同的转换函数维护其增量缓存。fold 返回分离的当前事件 seq 以及每次替换实际移除的 seq。`listEvents()` 和 `traceEvent()` 利用该结果为每个原始事件分类，使检查结果不会在位置替换语义上与 model-history 推导产生分歧。
|
||||
|
||||
`readEvent()` 返回完整的目标加上按连续 seq 排列的原始相邻事件。`before` 和 `after` 默认为零，各自受 `readWindowMax`（默认 50）约束。结果携带克隆的 `SessionHeader` 而非来源可用性记录，因为判断活跃目标的 persisted 标志会违反「活跃精确读取不依赖持久化健康状态」这一保证。
|
||||
|
||||
## 安全边界
|
||||
|
||||
该服务是上下文级别的受信任基础设施，而非授权层。未来面向模型的历史工具或人类 UI 将施加显式的调用方/会话范围。该服务不添加面向模型的工具，也不改变 transcript（文本记录）或快照的 surface。
|
||||
|
||||
## 曾考虑的替代方案
|
||||
|
||||
- **将逻辑语料库解析直接放在每个消费方中**：否决。来源优先级、冲突处理、可选服务生命周期、克隆与 surface 分类是共享的正确性规则。
|
||||
- **仅查询持久化**：否决。检查点可能落后于当前活跃日志。
|
||||
- **缓存持久化元数据并监听写入/删除**：否决。精确读取可以直接询问权威来源，而缓存失效在规模尚未要求时就引入了生命周期与并发状态。
|
||||
- **将提供方注册放入精确读取服务**：否决。SQLite 包拥有一套对账/事务生命周期；若没有第二个提供方证明其必要性，注册表只会拆分该状态。
|
||||
|
||||
## 后果
|
||||
|
||||
继承的精确读取实现只有一个来源解析状态变量：当前挂载的持久化服务。它没有提供方队列、指纹、提取器注册表、观察代次或派生索引更新；具体后端单独拥有其全文搜索状态。精确读取、语义扫描和事件追踪在纯活跃部署中仍然可用，在持久化存在时具有确定性。
|
||||
|
||||
跨语料库列表、血缘追踪和持久化事件操作在每次调用时执行后端 I/O。这是有意为之：正确性来自当前权威状态，而面向规模的全文搜索方法使用具体后端的 SQLite 派生索引。
|
||||
@@ -2,5 +2,5 @@
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-07-16-persistent-pty-sessions.md
|
||||
2026-07-16-persistent-pty-sessions.md: 43c87bb159cfe1ab9f8d3a80c2adf25a57ae6e3b
|
||||
2026-07-16-persistent-pty-sessions.zh.md: 8afc2103447cc58b1fcbc1062b9564e8ed643477
|
||||
2026-07-16-persistent-pty-sessions.md: 55ab262b89526911e16a3fec5dd624fc42fb2222
|
||||
2026-07-16-persistent-pty-sessions.zh.md: 525d199298bc071c36444d689368317968186c9f
|
||||
|
||||
@@ -72,13 +72,15 @@ With `run_in_background: true`, `dsh-tool-pty` registers the in-flight send on `
|
||||
|
||||
### Local readiness detection
|
||||
|
||||
The local backend first recognizes a private OSC prompt marker emitted by its controlled bash startup, then requires printable prompt text after that marker before declaring prompt readiness and runs three bounded fallback tiers. Carrying that state across data callbacks covers macOS delivery where the OSC marker and `PS1` arrive separately; the marker alone can no longer publish an empty MOTD. The marker is removed before output reaches the model and avoids a fixed silence delay for ordinary shell commands on both platforms. Unpublished startup does not accept zero-output silence as readiness; timeout rejects the spawn. If caller cancellation wins during startup, the backend closes the private session and propagates the exact `AbortSignal.reason`; a foreground PGID that is not observable yet cannot replace cancellation with a lookup error. All timings are validated config fields: `pollIntervalMs`, `exactProbeAfterMs`, `idleSilenceMs`, and `timeoutMs`.
|
||||
The local backend first recognizes a private OSC prompt marker emitted by its controlled bash startup, then requires printable prompt text after that marker before declaring prompt readiness and runs three bounded fallback tiers. Carrying that state across data callbacks covers macOS delivery where the OSC marker and `PS1` arrive separately; the marker alone can no longer publish an empty MOTD. The marker is removed before output reaches the model and avoids a fixed silence delay for ordinary shell commands on both platforms. Unpublished startup does not accept zero-output silence as readiness; timeout rejects the spawn. If caller cancellation wins during startup, the backend closes the private session and propagates the exact `AbortSignal.reason`; a foreground PGID that is not observable yet cannot replace cancellation with a lookup error. All timings are validated config fields: `pollIntervalMs`, `exactProbeAfterMs`, `idleSilenceMs`, `handoffGraceMs`, and `timeoutMs`.
|
||||
|
||||
On Linux, the inspector reads the shell's terminal foreground PGID from `/proc/<shellPid>/stat`, enumerates every process and thread in that process group, and probes their current syscalls. A positive Tier 1 result requires an observed stdin wait: direct `read(0)`, a permitted read of a `select`/`pselect6` or `poll`/`ppoll` argument containing fd 0, or an epoll interest list containing fd 0. Unreadable process memory and unrecognized syscalls are misses, never positive guesses. Architecture tables contain only syscall numbers defined by the corresponding Linux UAPI; unsupported architectures skip Tier 1.
|
||||
|
||||
On macOS there is no exact syscall tier. Output silence returns `inferred_idle` for any foreground process group, including Python and `gdb`; `ps`-derived terminal PGID is used for signaling, not as proof that only the shell can be idle. Pure process-inspector logic is injectable and unit-tested on Linux, while a macOS CI job exercises the real PTY and process-table path.
|
||||
|
||||
Tier 2 returns `inferred_idle` after `idleSilenceMs` without output. A sleeping or network-blocked command can therefore look ready. Tier 3 returns `timeout` after `timeoutMs` so a foreground tool call cannot hold the agent indefinitely. The result preserves the distinction; callers may wait through `ctx.tasks`, signal the foreground group, or inspect from another session.
|
||||
Tier 2 returns `inferred_idle` after `idleSilenceMs` without output. A sleeping or network-blocked command can therefore look ready. When a prompt marker was already seen, Tier 2 waits a further `handoffGraceMs` so a bash foreground handoff that lands on the silence boundary still settles as the exact `stdin_read` attribution instead of the weaker inference; the grace is a deployment-owned config field validated to cover at least one `pollIntervalMs`, because a grace shorter than the poll period cannot contain a single readiness poll and so cannot change any outcome. It bounds only sends that saw a marker, so its cost is the interactive return latency of that one case rather than every send. Tier 3 returns `timeout` after `timeoutMs` so a foreground tool call cannot hold the agent indefinitely. The result preserves the distinction; callers may wait through `ctx.tasks`, signal the foreground group, or inspect from another session.
|
||||
|
||||
Once a send settles under any tier, `PtySendOperation.append` stops accepting output, so later child output no longer reaches that settled operation; it still reaches the scrollback, and any send that is active when it arrives. A test that waits for a marker on the operation it started must therefore set `idleSilenceMs` and `timeoutMs` above the child's own startup latency; interpreter startup on a loaded macOS runner otherwise ends the send before the marker is printed.
|
||||
|
||||
`node-pty` data notifications feed one terminal parser. Parser carry state handles control sequences and a trailing carriage return split across callbacks, so a divided CRLF produces one newline rather than a pagination-changing blank line. The implementation normalizes line-oriented output, but it does not promise correct interaction with a full-screen application.
|
||||
|
||||
@@ -114,6 +116,7 @@ plugins:
|
||||
pollIntervalMs: 50
|
||||
exactProbeAfterMs: 150
|
||||
idleSilenceMs: 3000
|
||||
handoffGraceMs: 500
|
||||
timeoutMs: 30000
|
||||
disposeGraceMs: 3000
|
||||
'@deepseek-ai/dsh-tool-pty':
|
||||
@@ -152,7 +155,7 @@ The package ships concise tool guidance explaining persistent state, owner isola
|
||||
|
||||
## Verification
|
||||
|
||||
- Per-file coverage pins owner fencing, concurrent reservations, unpublished-spawn cancellation and awaited teardown, sandbox-mode change rejection, retriable lifecycle cleanup, readiness tiers, sanitizer carry state, complete UTF-8 bounds, task integration, schemas, and exact render intents.
|
||||
- Per-file coverage pins owner fencing, concurrent reservations, unpublished-spawn cancellation and awaited teardown, sandbox-mode change rejection, retriable lifecycle cleanup, readiness tiers, the configured handoff grace holding the idle fallback past one poll and its rejection below `pollIntervalMs`, sanitizer carry state, complete UTF-8 bounds, task integration, schemas, and exact render intents.
|
||||
- Linux process fixtures cover non-leader and non-main-thread stdin waits, zombie quiescence, unreadable process state, supported syscall tables, unsupported architectures, and false-positive rejection; macOS inspector logic is injected into the same unit suite.
|
||||
- Real `node-pty` tests exercise shell state, shared sandbox policy, environment scrubbing, raw-mode foreground `SIGINT` after deliberately delayed child readiness under scenario-owned timing bounds, a TERM-ignoring descendant, and immediate post-disposal quiescence on supported hosts.
|
||||
- A Loader-driven `cordis.yml` test mounts the real three-package composition. ACP and headless snapshots pin the six schemas, bounded results, and errors through opt-in overlays; TUI snapshots pin terminal and generic card presentation.
|
||||
@@ -165,6 +168,8 @@ The package ships concise tool guidance explaining persistent state, owner isola
|
||||
|
||||
**Idle below Linux Tier 1 is heuristic.** Output silence cannot distinguish a prompt from sleep or network I/O. The typed result preserves uncertainty, and bounded timeout plus task waiting and signaling keep control with the model.
|
||||
|
||||
**The exact-versus-inferred boundary is a latency trade, not a solvable race.** Attribution depends on whether the kernel publishes the foreground handoff before or after the silence bound elapses, so any fixed grace is a scheduling bet. `handoffGraceMs` puts that bet in deployment configuration: raising it buys exact `stdin_read` attribution on a slow or loaded host at the cost of interactive return latency after a prompt marker, and lowering it does the reverse. Tests that must not depend on the winner assert the observable behavior — the next send runs — rather than the attribution.
|
||||
|
||||
**Persistent state can drift from the model's belief.** The model may forget its cwd or active REPL. Session summaries and retained output help recovery, but no prompt can make state persistence deterministic.
|
||||
|
||||
**A daemonized descendant can leave the captured tree.** A process that reparents before teardown is no longer discoverable from the `node-pty` root. The implementation accepts that cleanup gap instead of risking SID-wide signals to unrelated processes.
|
||||
|
||||
@@ -72,13 +72,15 @@ UI 渲染契约精确且不携带位置信息。`terminal_send` 只为前台发
|
||||
|
||||
### 本地就绪检测
|
||||
|
||||
本地后端先识别受控 bash 启动时发出的私有 OSC prompt marker，并且只有在该 marker 后出现可打印的 prompt 文本时才据此声明 prompt 就绪；除此之外，它还运行 3 个有界 fallback 层级。在 data callback 之间保留这项状态，可以适配 macOS 分开交付 OSC marker 与 `PS1` 的情况；单独的 marker 不会发布空 MOTD。marker 在输出到达模型前被移除，使两个平台上的普通 shell 命令都无需固定等待静默阈值。尚未发布的 startup 不会把零输出静默视为就绪；timeout 会拒绝 spawn。若调用方取消在 startup 期间胜出，后端会关闭私有会话并原样抛出 `AbortSignal.reason`；尚不可观察的前台 PGID 不会再用查找错误覆盖取消原因。所有时间参数都是经校验的配置字段：`pollIntervalMs`、`exactProbeAfterMs`、`idleSilenceMs` 和 `timeoutMs`。
|
||||
本地后端先识别受控 bash 启动时发出的私有 OSC prompt marker，并且只有在该 marker 后出现可打印的 prompt 文本时才据此声明 prompt 就绪；除此之外，它还运行 3 个有界 fallback 层级。在 data callback 之间保留这项状态，可以适配 macOS 分开交付 OSC marker 与 `PS1` 的情况；单独的 marker 不会发布空 MOTD。marker 在输出到达模型前被移除，使两个平台上的普通 shell 命令都无需固定等待静默阈值。尚未发布的 startup 不会把零输出静默视为就绪；timeout 会拒绝 spawn。若调用方取消在 startup 期间胜出，后端会关闭私有会话并原样抛出 `AbortSignal.reason`；尚不可观察的前台 PGID 不会再用查找错误覆盖取消原因。所有时间参数都是经校验的配置字段：`pollIntervalMs`、`exactProbeAfterMs`、`idleSilenceMs`、`handoffGraceMs` 和 `timeoutMs`。
|
||||
|
||||
在 Linux 上，检查器从 `/proc/<shellPid>/stat` 读取 shell 的终端前台 PGID，枚举该进程组中的每个进程与线程，并检查它们当前的 syscall。Tier 1 只有观察到 stdin 等待才返回正结果：直接 `read(0)`、获准读取且含 fd 0 的 `select`/`pselect6` 或 `poll`/`ppoll` 参数，或者含 fd 0 的 epoll interest list。无法读取的进程内存和未识别的 syscall 都是 miss，绝不作为正向猜测。架构表只包含对应 Linux UAPI 定义的 syscall number；不支持的架构跳过 Tier 1。
|
||||
|
||||
macOS 没有精确 syscall 层。任何前台进程组输出静默都会返回 `inferred_idle`，包括 Python 和 `gdb`；从 `ps` 推导的终端 PGID 只用于发送信号，不作为「只有 shell 才能 idle」的证明。纯进程检查逻辑可注入并在 Linux 上完成 unit 覆盖率，同时由 macOS CI job 驱动真实 PTY 和进程表路径。
|
||||
|
||||
Tier 2 在持续 `idleSilenceMs` 没有输出后返回 `inferred_idle`，因此 sleep 或网络阻塞的命令可能看似 ready。Tier 3 在 `timeoutMs` 后返回 `timeout`，避免前台工具调用无限占住 agent。结果保留这些区别；调用方可以通过 `ctx.tasks` 等待、向前台组发信号，或从另一个会话排查。
|
||||
Tier 2 在持续 `idleSilenceMs` 没有输出后返回 `inferred_idle`，因此 sleep 或网络阻塞的命令可能看似 ready。如果此前已经见过 prompt marker，Tier 2 会再等待 `handoffGraceMs`，使恰好落在静默边界上的 bash 前台交接仍然以精确的 `stdin_read` 归因结束，而不是退到较弱的推断；该宽限是由部署方拥有的配置字段，并被校验为至少覆盖一个 `pollIntervalMs`——短于轮询周期的宽限装不下一次就绪轮询，因此不可能改变任何结果。它只约束见过 marker 的 send，代价是这一种情况的交互返回延迟，而不是每一次 send。Tier 3 在 `timeoutMs` 后返回 `timeout`，避免前台工具调用无限占住 agent。结果保留这些区别；调用方可以通过 `ctx.tasks` 等待、向前台组发信号，或从另一个会话排查。
|
||||
|
||||
一次 send 在任一层级 settle 之后，`PtySendOperation.append` 就不再接受输出，此后子进程的输出不会再进入那个已 settle 的 operation；它仍然会进入 scrollback，以及此时恰好处于活跃状态的任何 send。因此，等待自己所启动的 operation 上出现标记的测试，必须把 `idleSilenceMs` 与 `timeoutMs` 设得高于子进程自身的启动耗时；否则在负载较高的 macOS runner 上，解释器启动会在标记打印之前就结束这次 send。
|
||||
|
||||
`node-pty` data 通知进入同一个终端 parser。parser 的 carry state 会处理跨 callback 的控制序列和位于 callback 末尾的回车；因此，即使 CRLF 被拆开，也只会生成一个换行，而不会产生改变分页的空行。实现会规范化行式输出，但不承诺正确操作全屏应用。
|
||||
|
||||
@@ -114,6 +116,7 @@ plugins:
|
||||
pollIntervalMs: 50
|
||||
exactProbeAfterMs: 150
|
||||
idleSilenceMs: 3000
|
||||
handoffGraceMs: 500
|
||||
timeoutMs: 30000
|
||||
disposeGraceMs: 3000
|
||||
'@deepseek-ai/dsh-tool-pty':
|
||||
@@ -152,7 +155,7 @@ plugins:
|
||||
|
||||
## 验证
|
||||
|
||||
- 每文件覆盖率固定 owner 隔离、并发预留、未发布 spawn 的取消与等待式 teardown、沙箱模式变更拒绝、可重试的生命周期清理、就绪层级、sanitizer carry state、完整 UTF-8 结果上限、task 集成、schema 和精确 render intent。
|
||||
- 每文件覆盖率固定 owner 隔离、并发预留、未发布 spawn 的取消与等待式 teardown、沙箱模式变更拒绝、可重试的生命周期清理、就绪层级、配置化交接宽限把 idle fallback 顶过一次轮询以及低于 `pollIntervalMs` 时的拒绝、sanitizer carry state、完整 UTF-8 结果上限、task 集成、schema 和精确 render intent。
|
||||
- Linux 进程 fixture 覆盖非 leader 与非主线程的 stdin 等待、僵尸进程静止性、不可读进程状态、受支持的 syscall 表、不支持的架构和误报拒绝；同一单元测试套件通过注入覆盖 macOS 检查器逻辑。
|
||||
- 真实 `node-pty` 测试在受支持宿主上覆盖 shell 状态、共享沙箱策略、环境清洗、在由场景掌控的时间界限内先有意延迟子进程就绪，再对 raw mode 前台进程发送 `SIGINT`、忽略 `SIGTERM` 的子进程，以及 dispose 返回后立即完全停稳。
|
||||
- Loader 驱动的 `cordis.yml` 测试挂载真实三包组合。ACP 与 headless 快照通过 opt-in overlay 固定 6 个 schema、有界结果和错误；TUI 快照固定 terminal 与 generic 卡片展示。
|
||||
@@ -165,6 +168,8 @@ plugins:
|
||||
|
||||
**Linux Tier 1 之外的 idle 都是启发式结果。**输出静默无法区分 prompt、sleep 和网络 I/O。类型化结果保留不确定性，有界 timeout、task 等待与信号让模型仍能掌握控制权。
|
||||
|
||||
**精确归因与推断归因的边界是延迟取舍，不是可消除的竞态。**归因取决于内核在静默上限到达之前还是之后发布前台交接，因此任何固定宽限都是一次调度上的赌注。`handoffGraceMs` 把这个赌注交给部署配置：调大它可以在慢速或高负载主机上换到精确的 `stdin_read` 归因，代价是见过 prompt marker 之后的交互返回延迟；调小则相反。不应依赖胜负结果的测试断言可观察行为——下一次 send 能正常执行——而不是断言归因路径。
|
||||
|
||||
**持久状态可能偏离模型认知。**模型可能忘记 cwd 或活跃 REPL。会话摘要和保留输出有助恢复，但任何 prompt 都无法让状态持久化变成确定行为。
|
||||
|
||||
**daemonized 子进程可能离开捕获树。**在 teardown 前 reparent 的进程无法再从 `node-pty` 根进程发现。实现接受这个清理缺口，不冒险按 SID 向无关进程发送信号。
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-17-dedicated-full-screen-tui-front-door.md: 8d7c7b00c8d9b15ea3f2419ed44ca88209e60dad
|
||||
2026-07-17-dedicated-full-screen-tui-front-door.zh.md: 49e9541c9f98c9f3beba11945ff452fc38bd9ede
|
||||
# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-07-17-dedicated-full-screen-tui-front-door.md
|
||||
2026-07-17-dedicated-full-screen-tui-front-door.md: a3f3d6b51e85ad20218ce5aebf526bd96946be55
|
||||
2026-07-17-dedicated-full-screen-tui-front-door.zh.md: 6a0c2f12815e9418a2b5bcb237d3db3616ccd133
|
||||
|
||||
@@ -20,9 +20,9 @@ The selected front door receives the exact generated or resumed `SessionId` used
|
||||
|
||||
### Session projection and interaction
|
||||
|
||||
The TUI rebuilds the transcript from the active `session.surface` and reprojects it whenever an event carries a `surfaceOp`, so resumed and compacted history matches the model-visible conversation. It renders Markdown text and reasoning, token totals, the latest `todo/write` plan, and tool cards produced through each tool definition's `presentCall` and `presentResult` methods. Long card bodies retain a configurable head/tail preview with the hidden-line count; one terminal control expands or collapses every card. Pending chunks and tool calls update the same components that completed events settle.
|
||||
The TUI rebuilds the transcript from the active `session.surface` and reprojects it whenever an event carries a `surfaceOp`, so resumed and compacted history matches the model-visible conversation. It renders Markdown text and reasoning, including fenced code with hidden Markdown markers, a dim optional language label, and a code-colored body, token totals, the latest `todo/write` plan, and tool cards produced through each tool definition's `presentCall` and `presentResult` methods. Long card bodies retain a configurable head/tail preview with the hidden-line count; one terminal control expands or collapses every card. Pending chunks and tool calls update the same components that completed events settle.
|
||||
|
||||
Editor input calls `agent.send()` while idle and `agent.steer()` while a turn is running. Cancellation, reasoning visibility, tool-card expansion, redraw, transcript clearing, and exit are terminal-only controls. The idle footer derives context occupancy from `tokenMeter` and shows the selected model and explicit reasoning effort; during a run, elapsed activity and the Escape interrupt hint replace that summary. `/status` remains available in either state and appends a detailed terminal-only snapshot: session identity and timestamps, selected model, reasoning effort/default state and reasoning visibility, lifecycle counts folded from the event log, the same deduplicated usage buckets and KV-cache rate as the footer, and context use from `tokenMeter` plus the selected model's advertised capacity. The plugin registers the shared `userInteraction` provider and presents queued questions in a wide bottom-left keyboard panel with batch progress, numbered options, and aligned descriptions; agent behavior and answer logging remain owned by their existing services.
|
||||
Editor input calls `agent.send()` while idle and `agent.steer()` while a turn is running. Cancellation, reasoning visibility, tool-card expansion, redraw, transcript clearing, and exit are terminal-only controls. `/exit` and `/quit` share the same exit path: they cancel an active turn, wait for idle, and then restore and close the terminal. The idle footer derives context occupancy from `tokenMeter` and shows the selected model and explicit reasoning effort; during a run, elapsed activity and the Escape interrupt hint replace that summary. `/status` remains available in either state and appends a detailed terminal-only snapshot: session identity and timestamps, selected model, reasoning effort/default state and reasoning visibility, lifecycle counts folded from the event log, the same deduplicated usage buckets and KV-cache rate as the footer, and context use from `tokenMeter` plus the selected model's advertised capacity. The plugin registers the shared `userInteraction` provider and presents queued questions in a wide bottom-left keyboard panel with batch progress, numbered options, and aligned descriptions; the panel's controls hint lists only actions meaningful for the current option count, omitting navigation when exactly one option is shown; agent behavior and answer logging remain owned by their existing services.
|
||||
|
||||
The `/model` command presents the advisory `ctx.llm` catalog as a keyboard selector and changes only this TUI session's target; argument forms remain available for direct selection. Each model row owns the adapter-advertised reasoning-effort order and default: Shift+Tab cycles that row's efforts, includes provider-default behavior when the adapter advertises no default, and leaves models without selectable metadata unchanged. Agent-scoped prompt-assembly and request waterfalls snapshot one provider/model/reasoning-effort target per step, so `{{provider}}` / `{{model}}` interpolation and request routing cannot split when a command arrives during assembly. The latest logged request header restores a used target; a selection that never reaches a request remains process-local.
|
||||
|
||||
|
||||
@@ -20,9 +20,9 @@ DeepSeek Harness 将 [`@deepseek-ai/dsh-tui`](../../../../packages/ui/tui/README
|
||||
|
||||
### 会话投影与交互
|
||||
|
||||
TUI 从活跃的 `session.surface` 重建 transcript（文本记录），并在事件携带 `surfaceOp` 时重新投影，因此恢复或压缩后的历史与模型可见会话保持一致。TUI 渲染 Markdown 文本与推理、token 用量、最新 `todo/write` 计划，以及各工具定义通过 `presentCall` 和 `presentResult` 方法生成的工具卡片。较长的工具卡片正文会保留可配置的头尾预览，并显示隐藏行数；一个终端控制可以展开或收起全部卡片。进行中的分片与工具调用会更新同一组组件，随后由完成事件收束状态。
|
||||
TUI 从活跃的 `session.surface` 重建 transcript（文本记录），并在事件携带 `surfaceOp` 时重新投影，因此恢复或压缩后的历史与模型可见会话保持一致。TUI 渲染 Markdown 文本与推理（其中围栏代码块隐藏 Markdown 标记、保留一个暗色的可选语言标签，并使用代码配色的正文）、token 用量、最新 `todo/write` 计划，以及各工具定义通过 `presentCall` 和 `presentResult` 方法生成的工具卡片。较长的工具卡片正文会保留可配置的头尾预览，并显示隐藏行数；一个终端控制可以展开或收起全部卡片。进行中的分片与工具调用会更新同一组组件，随后由完成事件收束状态。
|
||||
|
||||
agent 空闲时，编辑器输入调用 `agent.send()`；轮次运行中则调用 `agent.steer()`。取消、推理显隐、工具卡片展开、重绘、清空 transcript 和退出都只是终端控制。空闲态页脚根据 `tokenMeter` 得出上下文占用率，并显示所选模型和显式选定的推理强度；agent 运行期间，该摘要会替换为带已用时长的活动指示和 Escape 中断提示。`/status` 在这两种状态下均可用，并会追加一份仅在终端显示的详细快照，其中包括会话标识与时间戳、所选模型、推理强度（或默认状态）及推理显隐状态、从事件日志归并得出的生命周期计数、与页脚一致的去重用量分项和 KV 缓存命中率，以及 `tokenMeter` 给出的上下文用量和所选模型公布的容量。插件注册共享的 `userInteraction` 提供方，在左下角宽幅键盘操作面板中呈现排队的问题，面板显示批次进度、带编号的选项和对齐的描述；agent 行为和答案日志仍由既有服务负责。
|
||||
agent 空闲时，编辑器输入调用 `agent.send()`；轮次运行中则调用 `agent.steer()`。取消、推理显隐、工具卡片展开、重绘、清空 transcript 和退出都只是终端控制。`/exit` 和 `/quit` 共用同一条退出路径：先取消进行中的轮次，等待 agent 空闲，然后恢复并关闭终端。空闲态页脚根据 `tokenMeter` 得出上下文占用率，并显示所选模型和显式选定的推理强度；agent 运行期间，该摘要会替换为带已用时长的活动指示和 Escape 中断提示。`/status` 在这两种状态下均可用，并会追加一份仅在终端显示的详细快照，其中包括会话标识与时间戳、所选模型、推理强度（或默认状态）及推理显隐状态、从事件日志归并得出的生命周期计数、与页脚一致的去重用量分项和 KV 缓存命中率，以及 `tokenMeter` 给出的上下文用量和所选模型公布的容量。插件注册共享的 `userInteraction` 提供方，在左下角宽幅键盘操作面板中呈现排队的问题，面板显示批次进度、带编号的选项和对齐的描述；面板的操作提示只列出在当前选项数量下有意义的操作，仅有一个选项时不显示导航项；agent 行为和答案日志仍由既有服务负责。
|
||||
|
||||
`/model` 命令将建议性的 `ctx.llm` 目录呈现为键盘选择器，并且只更改当前 TUI 会话的目标；带参数的形式仍可直接选择目标。每个模型行都持有适配器公布的推理强度顺序和默认值：按 Shift+Tab 可循环切换该行的推理强度；如果适配器没有公布默认值，循环中还会包含提供方默认行为；没有可选元数据的模型则保持不变。agent 作用域内的 prompt 组装和请求两条 waterfall（瀑布式事件）会为每个步骤快照一次同一个提供方/模型/推理强度目标，因此即使命令在组装期间到达，`{{provider}}` / `{{model}}` 插值与请求路由也不会分裂。系统通过日志中最新的请求头恢复已经使用过的目标；未被请求使用的选择只保留在当前进程中。
|
||||
|
||||
|
||||
@@ -2,5 +2,5 @@
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-19-model-facing-goal-tools.md: 2c53a7658e213dee4fecc93709244f97b821aca0
|
||||
2026-07-19-model-facing-goal-tools.zh.md: aa7b5ea14b7afff88819f0efa35c5f2d5e2e933e
|
||||
2026-07-19-model-facing-goal-tools.md: bc4305af80bb13ceeff1888d489dcd8a00132f94
|
||||
2026-07-19-model-facing-goal-tools.zh.md: b07f62aa526902c4b2e9c081777a76ca53783d31
|
||||
|
||||
@@ -22,13 +22,13 @@ The prompt tells the model that it may infer goal intent from a direct human req
|
||||
|
||||
All three tools use exclusive execution so a model-ordered batch observes prior mutations and their new revisions. Results are compact JSON. UI presentation is a pure function of arguments and uses generic read or mutation cards; mutation cards select meaningful action values before the goal id, so accepted fillers cannot blank their input. Activation is reported only as live observation and is never written into replay state.
|
||||
|
||||
An autonomous goal round that successfully reports completion or blocking contributes the existing terminal `agent/turn-stop` decision for that physical turn, preventing an unnecessary follow-up request. Direct-human mutations do not contribute a terminal stop: the assistant can acknowledge the change, and concurrent human steering remains available to ordinary continuation folding.
|
||||
An autonomous goal round that successfully reports completion or blocking marks its tool result as concluding the physical turn, preventing an unnecessary follow-up request. Direct-human mutations do not conclude the turn: the assistant can acknowledge the change, and concurrent human steering remains available to ordinary stopping checks.
|
||||
|
||||
### Execution authority
|
||||
|
||||
Every call requires an `exec.agent` that is the exact running object in `AgentRegistry`, is the current inherited driver initiator, and has an open turn. These are execution-time checks and cannot be bypassed by prompt injection or hand-authored tool arguments.
|
||||
|
||||
Create, edit, pause, and resume additionally require an accepted user message or user steering event in the current turn of a runtime-root agent. Root ownership is derived from the live agent graph rather than durable fork ancestry: a resumed fork can receive direct human authority, while a live child remains a subagent and cannot mutate these states. User source is a host attestation: `Agent.send()` and `steer()` default an omitted source to `{ kind: 'user' }`, so non-human producers must label their own content. The runtime proves provenance, not whether the human's wording semantically warrants creation or resumption; that interpretation remains with the model.
|
||||
Create, edit, pause, and resume additionally require an accepted user message or user steering event in the current turn of a runtime-root agent. Root ownership is derived from the live agent graph rather than durable fork ancestry: a resumed fork can receive direct human authority, while a live child remains a subagent and cannot mutate these states. User source is a host attestation: every `Agent.send()` or `steer()` input requires an explicit source, so the host labels direct human content `{ kind: 'user' }` and non-human producers label their own provenance. The runtime proves provenance, not whether the human's wording semantically warrants creation or resumption; that interpretation remains with the model.
|
||||
|
||||
Complete and blocked accept either direct-human authority or the exact current goal round. Goal-round authority requires a goal-sourced `user/message` whose goal id, revision, and round all equal the folded current goal. It grants only the two terminal reports. Direct human authority may stop a goal immediately.
|
||||
|
||||
|
||||
@@ -22,13 +22,13 @@ Status: implemented
|
||||
|
||||
三个工具都采用独占执行，使模型排序的批次可以观察此前变更及其新修订号。结果为紧凑 JSON。UI 展示是参数的纯函数，使用通用读取或变更卡片；变更卡片选择输入时，先取有实际意义的操作值，再取目标 id，因此允许的占位值不会使卡片输入留空。激活态仅作为实时观察返回，绝不会写入回放状态。
|
||||
|
||||
自主目标回合成功报告完成或阻塞后，插件会为该物理轮次贡献现有的终止型 `agent/turn-stop` 决策，避免再发起一次不必要的模型请求。直接人类发起的变更不会贡献终止决策：智能体可以确认该变更，并且并发的人类 steering（转向）仍可参与普通的继续执行折叠。
|
||||
自主目标回合成功报告完成或阻塞后，其工具结果会被标记为结束该物理轮次，避免再发起一次不必要的模型请求。直接人类发起的变更不会结束轮次：agent 可以确认该变更，并且并发的人类 steering（中途引导）仍可参与普通的停止检查。
|
||||
|
||||
### 执行权限
|
||||
|
||||
每次调用都要求存在 `exec.agent`，且它必须是 `AgentRegistry` 中完全相同的运行中对象、当前继承的驱动发起者，并处于开放轮次内。这些检查在执行时进行，不能通过提示词注入或手写工具参数绕过。
|
||||
|
||||
创建、编辑、暂停与恢复还要求运行时根智能体的当前轮次已经接纳一条用户消息或用户 steering（转向）事件。根所有权来自实时智能体图，而非持久的 fork 祖先关系：恢复后的派生会话可以接收新的直接人类权限，实时子智能体则仍然是子智能体，不能改变这些状态。用户来源是宿主的证明：`Agent.send()` 和 `steer()` 会把省略的来源默认为 `{ kind: 'user' }`，因此非人类生产者必须标注自己的内容。运行时证明来源，而不判断人类措辞在语义上是否足以创建或恢复目标；该解释仍由模型完成。
|
||||
创建、编辑、暂停与恢复还要求运行时根 agent 的当前轮次已经接纳一条用户消息或用户 steering 事件。根所有权来自实时 agent 图，而非持久的 fork 祖先关系：恢复后的派生会话可以接收新的直接人类权限，实时子级则仍是 subagent，不能改变这些状态。用户来源是宿主的证明：每个 `Agent.send()` 或 `steer()` 输入都必须显式提供来源，因此宿主把直接人类内容标为 `{ kind: 'user' }`，非人类生产者则标注自己的来源信息。运行时证明来源，而不判断人类措辞在语义上是否足以创建或恢复目标；该解释仍由模型完成。
|
||||
|
||||
完成与阻塞既接受直接人类权限，也接受准确的当前目标回合。目标回合权限要求存在一条来源为目标的 `user/message`，其中目标 id、修订号和回合都与折叠后的当前目标相等。它只授予这两种终止报告权限。直接人类权限可以立即停止目标。
|
||||
|
||||
|
||||
@@ -2,5 +2,5 @@
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-19-same-session-goal-round-driver.md: d23af9a9b05d60d2dccad095455524844f1185b9
|
||||
2026-07-19-same-session-goal-round-driver.zh.md: f4f0cd6fd575d14427025bdbd8d10bc90e25f780
|
||||
2026-07-19-same-session-goal-round-driver.md: 0e6be9fe3109336d47867ab52c585dc267309fb4
|
||||
2026-07-19-same-session-goal-round-driver.zh.md: cfd9d1aa8cbc3c17cd046df4f57a8f79a6877f5c
|
||||
|
||||
@@ -24,7 +24,7 @@ When an agent is idle, has no competing queued work, and its current goal is `ac
|
||||
|
||||
The `agent/prompt-submit` waterfall is the admission fence. A positive goal source is allowed only when it exactly matches the driver's pending identity and content, the live goal still has that id and revision, activation remains armed, and the round is still the next number. The plugin checks once before delegating and again after downstream hooks return. This second check prevents an async hook from editing or pausing the goal while still admitting the old prompt.
|
||||
|
||||
Only the resulting `user/message` is an admitted round and advances the goal fold. A stale reservation becomes a durable `prompt/blocked` plus zero-step rejected turn, but the driver marks it stale and does not charge the round. A downstream policy rejection that is not caused by staleness blocks the goal rather than retrying around policy.
|
||||
Only the resulting `user/message` is an admitted round and advances the goal fold. A stale reservation is discarded before a turn opens; the driver marks it stale and does not charge the round. A downstream policy rejection that is not caused by staleness blocks the goal rather than retrying around policy.
|
||||
|
||||
### Human work and revision races
|
||||
|
||||
@@ -43,7 +43,6 @@ The driver classifies one closed goal-owned turn as follows:
|
||||
| `error` with code `RATE_LIMIT` or `QUOTA` | block with code `usage-limited` |
|
||||
| other `error` | block with code `turn-error` |
|
||||
| `max-tokens` | block with code `max-tokens` |
|
||||
| non-stale `rejected` | block with code `prompt-rejected` |
|
||||
| failed durability checkpoint | disarm without changing durable phase |
|
||||
| `disposed` or `interrupted` | disarm |
|
||||
| plugin-added unknown result | block for inspection |
|
||||
|
||||
@@ -24,7 +24,7 @@ Status: implemented
|
||||
|
||||
`agent/prompt-submit` 瀑布是接纳栅栏。正数目标来源只有在完全匹配驱动器待处理的身份和内容、实时目标仍具有相同 id 与修订号、激活态仍为 armed，并且该回合仍是下一个编号时才会获准。插件在委托下游监听器前检查一次，在下游返回后再检查一次。第二次检查防止异步钩子编辑或暂停目标后，旧提示词仍被接纳。
|
||||
|
||||
只有最终产生的 `user/message` 才是已接纳目标回合，并推进目标折叠。过期预留会生成持久的 `prompt/blocked` 和零步骤 rejected 轮次，但驱动器会把它标记为过期，不消耗回合数。若下游策略拒绝并非由过期导致，目标会进入 blocked，而不会绕过该策略自动重试。
|
||||
只有最终产生的 `user/message` 才是已接纳目标回合，并推进目标折叠。陈旧预留会在轮次打开前被丢弃；驱动器会把它标记为陈旧，不消耗回合数。若下游策略拒绝并非由陈旧状态导致，目标会进入 blocked，而不会绕过该策略自动重试。
|
||||
|
||||
### 人类工作与修订竞争
|
||||
|
||||
@@ -43,7 +43,6 @@ Status: implemented
|
||||
| 代码为 `RATE_LIMIT` 或 `QUOTA` 的 `error` | 以 `usage-limited` 代码阻塞 |
|
||||
| 其他 `error` | 以 `turn-error` 代码阻塞 |
|
||||
| `max-tokens` | 以 `max-tokens` 代码阻塞 |
|
||||
| 非过期的 `rejected` | 以 `prompt-rejected` 代码阻塞 |
|
||||
| 持久检查点失败 | 解除激活，但不改变持久阶段 |
|
||||
| `disposed` 或 `interrupted` | 解除激活 |
|
||||
| 插件新增的未知结果 | 阻塞并等待检查 |
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-20-code-mode-typed-tool-returns.md: 3d8642d66baf521f22dfb1ea0ef3e64683f918d4
|
||||
2026-07-20-code-mode-typed-tool-returns.zh.md: fea1be3e236c0ccba729e449ab9714ed497d900a
|
||||
# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-07-20-code-mode-typed-tool-returns.md
|
||||
2026-07-20-code-mode-typed-tool-returns.md: 2081d8161f0ee14493a09762b18ec7d9d07ea3c4
|
||||
2026-07-20-code-mode-typed-tool-returns.zh.md: 0fa5eec7a96ebba6796dda6221b83e91f14ebf7d
|
||||
|
||||
@@ -69,7 +69,7 @@ Compute time, wall time, worker heap, cancellation, and fresh-worker isolation r
|
||||
|
||||
Background producers return a typed canonical handle such as `{ kind: 'background', taskId }` while retaining their established Native sentence. A pre-aborted background call remains a failure because successful output promises an id and no task was created. After `ctx.tasks.start()` publishes the id, task-owned cancellation governs the work: settlement or later cancellation of the enclosing `run_code` call does not kill it. A later program can pass the returned id to `task_output`, and `task_kill`, owner disposal, or service teardown owns cancellation. Foreground execution remains coupled to the call signal. The task lifetime contract is owned by the [background task runtime note](../architecture/2026-06-20-generic-long-running-tool-runtime.md).
|
||||
|
||||
Dynamic Cordis mounting follows the same rule: `cordis_mount` returns `{ id, pluginName, state, provides, waitingFor }`, so a program can read `mounted.id`, inspect active or pending state, and pass that id to `cordis_unmount` without parsing the stable Native sentence.
|
||||
Temporary Cordis Plugins follow the same rule: `cordis_mount` returns `{ id, pluginName, state, provides, waitingFor }`, so a program can read `mounted.id`, inspect active or pending state, and pass that id to `cordis_unmount` without parsing the stable Native sentence.
|
||||
|
||||
### Persistence, metadata, and spill
|
||||
|
||||
|
||||
@@ -69,7 +69,7 @@ Code Mode 通过运行时请求中的 `{ name: "ToolCallError", memberNameProper
|
||||
|
||||
后台 producer 返回类型化的规范句柄，例如 `{ kind: 'background', taskId }`，同时保留既有的 Native 语句。已预先中止的后台调用仍是失败，因为成功输出承诺返回 id，而此时并未创建任务。`ctx.tasks.start()` 发布 id 后，工作由任务自有的取消机制控制：外围 `run_code` 调用完成，或随后被取消，都不会终止该任务。后续程序可以把返回的 id 传给 `task_output`；取消则由 `task_kill`、owner dispose 或服务 teardown 负责。前台执行仍与本次调用的信号耦合。任务生命周期契约由[后台任务运行时 Agent Note](../architecture/2026-06-20-generic-long-running-tool-runtime.md)定义。
|
||||
|
||||
动态 Cordis 挂载遵循同一规则：`cordis_mount` 返回 `{ id, pluginName, state, provides, waitingFor }`，因此程序可以直接读取 `mounted.id`，检查 active 或 pending 状态，并把该 id 传给 `cordis_unmount`，无需解析稳定的 Native 语句。
|
||||
临时 Cordis Plugin 遵循同一规则：`cordis_mount` 返回 `{ id, pluginName, state, provides, waitingFor }`，因此程序可以直接读取 `mounted.id`，检查 active 或 pending 状态，并把该 id 传给 `cordis_unmount`，无需解析稳定的 Native 语句。
|
||||
|
||||
### 持久化、元数据与输出落盘
|
||||
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-21-cross-session-references.md: fc084b36e7920a72efff0f363278d24eaebc4c69
|
||||
2026-07-21-cross-session-references.zh.md: fe4a876b5265fa7ad298adf3b829bcec70e878e8
|
||||
# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-07-21-cross-session-references.md
|
||||
2026-07-21-cross-session-references.md: 18dab5fb85f2201258e2f15c3c069d2668ae80d3
|
||||
2026-07-21-cross-session-references.zh.md: 44e33ab1ed762b08b6be30d66b82396a5b519c02
|
||||
|
||||
@@ -10,7 +10,7 @@ TUI users need to bring relevant work from another conversation into one new mes
|
||||
|
||||
## Decision
|
||||
|
||||
`@deepseek-ai/dsh-session-reference` is one context consumer service at `ctx.sessionReferences`. Hosts normalize their protocol into `SessionReferenceInput[]`, call `prepare()` before enqueue, and pass the returned contexts through the generic `SendOptions.contexts` boundary. Core agent packages know only that one queued message may carry frozen `HookContext[]`; they do not parse session URIs or read another log.
|
||||
`@deepseek-ai/dsh-session-reference` is one context consumer service at `ctx.sessionReferences`. Hosts normalize their protocol into `SessionReferenceInput[]` and call `prepare()` before delivery. The service returns detached readable content plus an optional sourced `UserMessageData` snapshot; core agent packages do not parse session URIs or read another log.
|
||||
|
||||
`dsh-session:<base64url(JSON.stringify(sessionId))>` is the canonical host-independent identifier. JSON string encoding precedes base64url so quotes, slashes, backslashes, Unicode, newlines, and every other JavaScript string value round-trip without delimiter ambiguity. TUI renders that URI inside `@[label](uri)`; text-only clients may use the same inline mention. Explicit Markdown mentions reject malformed URIs. Bare text becomes a reference only for a non-empty base64url-shaped payload, whose decode must still be canonical; empty or punctuation-only uses remain ordinary discussion text.
|
||||
|
||||
@@ -22,17 +22,17 @@ Preparation deduplicates in first-appearance order, rejects the target id, enfor
|
||||
|
||||
Projection retains direct-user messages and steering, completed assistant text, and checkpoint user messages carrying the canonical source exported by `dsh-compact`. That marker is part of the compaction capability contract rather than a backend package name. When a source prompt already contains baked prefix context, projection reads only its model-hidden display content, so referencing that target later does not recursively propagate an earlier snapshot. Projection excludes shadowed pre-compaction nodes, tools and results, reasoning, injected context, other plugin user messages, log-only records, and incomplete assistant chunks. Repeated compaction therefore exposes only the latest folded checkpoint lineage still on the current surface plus its retained tail; there is no raw/current switch and no shadow recovery.
|
||||
|
||||
One aggregated context is serialized as JSON beneath a fixed untrusted-background warning. The warning tells the model not to follow instructions, permission claims, or tool requests from referenced sessions unless the current user repeats them. Tag-safe serialization emits every data `<` as the lossless JSON escape `\u003c`; source strings therefore cannot spell the surrounding XML-like tags. The `## My request:` text is a routing cue rather than the trust boundary: referenced data may spell those words inside a JSON string, but it cannot forge the closing `</referenced-sessions>` tag or escape the data region. The same serializer drives each source's independent byte accounting. The context declares `prompt-prefix` placement, so AgentLoop persists one `user/message` or `steering/message` containing the snapshot, `## My request:` delimiter, and effective direct prompt. Its model-hidden envelope retains the direct display content and source/retention metadata. Target replay therefore satisfies the model-visible/log-reconstructable invariant without a new event type or a separate user-role context message.
|
||||
One aggregated context is serialized as JSON beneath a fixed untrusted-background warning. The warning tells the model not to follow instructions, permission claims, or tool requests from referenced sessions unless the current user repeats them. Tag-safe serialization emits every data `<` as the lossless JSON escape `\u003c`; source strings therefore cannot spell the surrounding XML-like tags or escape the data region. The same serializer drives each source's independent byte accounting. AgentLoop persists the snapshot as a sourced `user/message` immediately before the direct `user/message` or `steering/message`; target replay therefore satisfies the model-visible/log-reconstructable invariant without a new event type, placement mode, or prompt envelope.
|
||||
|
||||
## Message ownership
|
||||
|
||||
`send()` and `steer()` snapshot content, resolved source, and contexts together as one deeply frozen lossless-JSON inbox record. Synthetic `inject()` accepts source and model-hidden metadata but not attached contexts, which belong to inbox messages. A claimed ordinary message exposes its attached contexts as the default `agent/prompt-submit` additional contexts; a block writes neither user message nor contexts. The waterfall's returned allow is authoritative, so a listener wrapping `next()` preserves downstream content and contexts unless it intentionally replaces them. After admission, absent or `separate` placement writes an independent `context/message`, while `prompt-prefix` placement bakes context and the effective request into one prompt event. Drained steering bypasses `agent/prompt-submit` but applies the same placement split. Late steering retains the same record when converted to queued input, while cancellation, disposal, and terminal discard drop message and contexts together. `agent/queued` reports the frozen contexts so the observation event describes the complete retained item.
|
||||
TUI owns the snapshot/direct-message transaction without extending the generic inbox record. Outside the next-step acceptance window, it installs a one-shot outer `agent/prompt-submit` listener before `followup()`; an allowed decision receives the snapshot as `additionalContexts`, while a blocked or discarded prompt releases the listener and writes neither message. During prompt admission or an open turn, TUI calls `inject(snapshot)` then `steer(prompt)`, and AgentLoop stages both for the same safe boundary. If admission fails before that boundary, both remain staged for retry or a later admitted prompt; cancellation or disposal may discard them. The [separate-context decision](../architecture/2026-07-24-separate-context-injection-from-turn-execution.md) owns this generic delivery boundary.
|
||||
|
||||
This preserves host driving semantics: TUI decides `send()` versus `steer()` from the agent state after preparation, so only its queued path dispatches UserPromptSubmit hooks. Reference preparation is not a new steering protocol and does not create a turn by itself.
|
||||
Reference preparation is not a new steering protocol and does not create a turn by itself. A `followup()` outside the next-step acceptance window dispatches prompt admission; steering inside the window bypasses it while retaining snapshot order through the shared outbox.
|
||||
|
||||
## Host adapters
|
||||
|
||||
TUI combines session candidates with the existing `@` file provider. Each candidate displays the latest folded session title and falls back to the session id; lookup follows the editor's cancellation signal, and session id, cwd, and mention labels escape external terminal controls while the canonical URI retains the original id. TUI prepares only submissions containing structured mentions, disables duplicate submit while awaiting snapshots, restores failed input, renders the prompt envelope's display content as the user message, and renders its session-reference metadata as a compact source list instead of exposing the complete JSON in the terminal.
|
||||
TUI combines session candidates with the existing `@` file provider. Each candidate displays the latest folded session title and falls back to the session id; lookup follows the editor's cancellation signal, and session id, cwd, and mention labels escape external terminal controls while the canonical URI retains the original id. TUI prepares only submissions containing structured mentions, disables duplicate submit while awaiting snapshots, restores failed input, renders the readable direct content as the user message, and renders session-reference source metadata as a compact source list instead of exposing the complete JSON in the terminal.
|
||||
|
||||
The [automation-only ACP transport](../simplification/2026-07-23-acp-automation-only-protocol.md) deliberately does not mount session-query or session-reference services.
|
||||
|
||||
@@ -45,15 +45,15 @@ Each of at most three references is independently capped at 65,536 UTF-8 bytes b
|
||||
- **Wait for SQLite FTS5** — rejected because snapshot correctness requires exact id reads and canonical surface folding, not content search. FTS improves discovery only.
|
||||
- **Put mention syntax in `Agent.send()`** — rejected because it would make the core protocol parse one host's presentation syntax and prevent typed non-text hosts from sharing the semantic layer.
|
||||
- **Implement references separately in each host** — rejected because projection, security warning, retention, and persistence would drift across hosts.
|
||||
- **Place a separate user-role context message beside the prompt** — rejected because two adjacent user messages weaken the prompt's deictic binding: in `@foo what does this session discuss?`, the model may resolve “this session” as the current conversation instead of the referenced snapshot.
|
||||
- **Bake the prefix host-side before `send()`** — rejected because `agent/prompt-submit` must inspect and rewrite only the direct prompt. The effective prompt and attached contexts meet only after admission in AgentLoop, which can apply an `allow.content` rewrite consistently to both combined model content and `envelope.displayContent`; earlier host assembly would expose snapshot bytes to the hook or let those two views diverge.
|
||||
- **Attach context to `SendOptions` and the inbox record** — rejected because generic delivery would own a domain transaction through admission, steering, cancellation, and observation. A domain-specific admission wrapper and the existing next-step outbox preserve the required pairing without enlarging every message.
|
||||
- **Bake the prefix host-side before `send()`** — rejected because `agent/prompt-submit` must inspect and rewrite only the direct prompt. Keeping the snapshot as a separate sourced message preserves that boundary and lets TUI hide background bytes from the direct user bubble.
|
||||
- **Replay the raw source log or restore shadowed events** — rejected because compact defines the current model surface and may intentionally retire sensitive or expensive history.
|
||||
- **Resume or fork the source** — rejected because the feature supplies read-only background for one target message, not identity or lifecycle continuity.
|
||||
- **Inject at request time by rereading the source** — rejected because the reference would become nondeterministic, cancellation races could alter its bytes, and target replay would depend on external mutable state.
|
||||
|
||||
## Verification
|
||||
|
||||
Unit and integration coverage pins URI round-trips and text-boundary punctuation, explicit malformed references, title-aware candidate ranking, terminal-control escaping, projection exclusions, non-recursive prompt-envelope projection, backend-independent compact checkpoints, tag-safe framing, deduplication, self-reference, count limits, all-or-nothing reads, prompt cancellation against a non-settling storage read, independent per-source byte retention, frozen message ownership, prompt blocking, send/steer placement, title isolation, missing capability, and compact TUI replay. A keyless TUI snapshot runs the real agent loop: the source surface replaces old user/assistant history with a compact checkpoint, the target submits a mention, and the captured model request contains one user message ordered as snapshot, request delimiter, and current prompt, without either shadowed string.
|
||||
Unit and integration coverage pins URI round-trips and text-boundary punctuation, explicit malformed references, title-aware candidate ranking, terminal-control escaping, projection exclusions, non-recursive snapshot projection, backend-independent compact checkpoints, tag-safe framing, deduplication, self-reference, count limits, all-or-nothing reads, prompt cancellation against a non-settling storage read, independent per-source byte retention, prompt blocking, admission-time staging, send/steer placement, title isolation, missing capability, and compact TUI replay. A keyless TUI snapshot runs the real agent loop: the source surface replaces old user/assistant history with a compact checkpoint, the target submits a mention, and the captured model request contains a sourced snapshot message followed by the readable current prompt, without either shadowed string.
|
||||
|
||||
## Consequences
|
||||
|
||||
|
||||
@@ -10,7 +10,7 @@ TUI 用户需要把另一场对话中的相关工作带入一条新消息，但
|
||||
|
||||
## 决策
|
||||
|
||||
`@deepseek-ai/dsh-session-reference` 是注册在 `ctx.sessionReferences` 上的单一上下文消费服务。宿主先把各自的协议规范化为 `SessionReferenceInput[]`，在入队前调用 `prepare()`，再通过通用的 `SendOptions.contexts` 边界传递返回的上下文。核心 agent 包只知道一条排队消息可以携带已冻结的 `HookContext[]`；它们既不解析会话 URI，也不读取其他日志。
|
||||
`@deepseek-ai/dsh-session-reference` 是注册在 `ctx.sessionReferences` 上的单一上下文消费服务。宿主先把各自的协议规范化为 `SessionReferenceInput[]`，并在交付前调用 `prepare()`。该服务返回分离的可读内容和一份可选的、带来源信息的 `UserMessageData` 快照；核心 agent 包既不解析会话 URI，也不读取其他日志。
|
||||
|
||||
`dsh-session:<base64url(JSON.stringify(sessionId))>` 是与宿主无关的规范标识符。系统先执行 JSON 字符串编码，再执行 base64url 编码，因此引号、正斜杠、反斜杠、Unicode、换行符以及其他任意 JavaScript 字符串值都能无损往返，不会因分隔符产生歧义。TUI 把该 URI 渲染到 `@[label](uri)` 中；纯文本客户端可以使用同一种行内提及标记。显式 Markdown 提及标记会拒绝格式错误的 URI。裸文本只有在负载非空且形状符合 base64url 时才会成为引用，而且解码过程仍须通过规范性校验；空负载或只含标点符号的用法仍按普通讨论文本处理。
|
||||
|
||||
@@ -22,17 +22,17 @@ TUI 用户需要把另一场对话中的相关工作带入一条新消息，但
|
||||
|
||||
投影会保留直接用户消息与 steering（中途引导）、已完成的 assistant 文本，以及携带由 `dsh-compact` 导出的规范来源标记的检查点用户消息。该标记属于压缩功能契约的一部分，而非某个后端包名称。当源提示词已包含合并写入的前缀上下文时，投影只读取其模型不可见的显示内容，因此后续引用该目标不会递归传播先前的快照。投影会排除压缩前已被遮蔽的节点、工具及其结果、推理（reasoning）、注入的上下文、其他插件用户消息、仅用于日志的记录，以及尚未完成的 assistant 分片。因此，重复压缩只会暴露当前表层仍保留的最新折叠检查点谱系及其尾部消息；系统不提供 raw/current 开关，也不恢复被遮蔽的内容。
|
||||
|
||||
系统把一个聚合上下文序列化为 JSON，并置于固定的不可信背景警告之后。该警告要求模型不要遵循被引用会话中的指令、权限声明或工具请求，除非当前用户再次提出这些内容。标签安全序列化会把数据中的每个 `<` 无损转义为 JSON `\u003c`；因此源字符串无法拼出外围类似 XML 的标签。`## My request:` 文本只是路由提示，不是信任边界：被引用数据可以在 JSON 字符串中包含这些词，但无法伪造闭合的 `</referenced-sessions>` 标签，也无法逃逸数据区域。同一个序列化器会独立核算每个源的字节数。该上下文声明 `prompt-prefix` 放置方式，因此 AgentLoop 会持久化一条 `user/message` 或 `steering/message`，其中包含快照、`## My request:` 分隔符和最终生效的直接提示词。其模型不可见封套保留直接显示内容以及来源与保留元数据。因此，目标回放无需新增事件类型或单独的用户角色上下文消息，也能满足「模型可见／日志可重建」不变量。
|
||||
系统把一个聚合上下文序列化为 JSON，并置于固定的不可信背景警告之后。该警告要求模型不要遵循被引用会话中的指令、权限声明或工具请求，除非当前用户再次提出这些内容。标签安全序列化会把数据中的每个 `<` 无损转义为 JSON `\u003c`；因此源字符串无法拼出外围类似 XML 的标签，也无法逃逸数据区域。同一个序列化器会独立核算每个源的字节数。AgentLoop 会把快照持久化为一条带来源信息的 `user/message`，紧接在直接 `user/message` 或 `steering/message` 之前。因此，目标回放无需新增事件类型、放置模式或提示词封套，也能满足「模型可见／日志可重建」不变量。
|
||||
|
||||
## 消息所有权
|
||||
|
||||
`send()` 与 `steer()` 会把内容、解析后的来源和上下文一起快照为一条深度冻结、无损 JSON 的收件箱记录。合成的 `inject()` 接受来源和模型不可见的元数据，但不接受附加上下文，因为上下文属于收件箱消息。普通消息被认领后，其附带的上下文会作为 `agent/prompt-submit` 的默认附加上下文公开；提示词被阻止时，系统既不写入用户消息，也不写入上下文。waterfall（瀑布式事件）返回的 allow 结果具有最终权威性，因此监听器包装 `next()` 时会保留下游内容与上下文，除非它有意替换这些值。消息被接纳后，未指定放置方式或指定为 `separate` 时会写入独立的 `context/message`；指定为 `prompt-prefix` 时则会把上下文与最终生效的请求合并写入同一个提示词事件。排空 steering 消息时会绕过 `agent/prompt-submit`，但采用相同的放置方式分流。延迟到达的 steering 转换为排队输入时保留同一条记录；取消、dispose（资源释放）和到达终止态后的丢弃则会同时丢弃消息与上下文。`agent/queued` 会报告已冻结的上下文，使观察事件能够描述完整的保留项。
|
||||
TUI 负责快照／直接消息事务，不扩展通用收件箱记录。在 next-step 接受窗口之外，它会在调用 `followup()` 前安装一次性的外层 `agent/prompt-submit` 监听器；获准决策会把快照作为 `additionalContexts` 接收，而被阻止或丢弃的提示词会释放监听器，并且不写入任何消息。提示词准入期间或轮次打开时，TUI 会依次调用 `inject(snapshot)` 和 `steer(prompt)`，AgentLoop 则将两者暂存到同一个安全边界。如果准入在抵达该边界前失败，两者都会保留暂存状态，供重试或后续获准提示词使用；取消或资源释放可能丢弃它们。这一通用交付边界由[上下文分离决策](../architecture/2026-07-24-separate-context-injection-from-turn-execution.md)规定。
|
||||
|
||||
这保留了宿主的驱动语义：TUI 在准备完成后根据 agent 状态决定调用 `send()` 还是 `steer()`，因此只有它的排队路径才会分派 UserPromptSubmit 钩子。引用准备过程不是新的 steering 协议，本身也不会创建轮次。
|
||||
引用准备过程不是新的 steering 协议，本身也不会创建轮次。在 next-step 接受窗口之外调用 `followup()` 会分派提示词准入；窗口内的 steering 会绕过它，同时通过共享 outbox 保持快照顺序。
|
||||
|
||||
## 宿主适配器
|
||||
|
||||
TUI 把会话候选与现有 `@` 文件提供方组合在一起。每个候选项显示最新折叠后的会话标题，没有标题时回退到 session id。候选查询遵循编辑器的取消信号；session id、cwd 和提及标签中的外部终端控制字符会被转义，但规范 URI 仍保留原始 id。TUI 只准备包含结构化提及标记的提交；等待快照时禁用重复提交；失败时恢复输入；它把提示词封套的显示内容渲染为用户消息，并把其中的会话引用元数据渲染为精简的来源列表，不在终端中暴露完整 JSON。
|
||||
TUI 把会话候选与现有 `@` 文件提供方组合在一起。每个候选项显示最新折叠后的会话标题，没有标题时回退到 session id。候选查询遵循编辑器的取消信号；session id、cwd 和提及标签中的外部终端控制字符会被转义，但规范 URI 仍保留原始 id。TUI 只准备包含结构化提及标记的提交；等待快照时禁用重复提交；失败时恢复输入；它把可读的直接内容渲染为用户消息，并把会话引用来源元数据渲染为精简的来源列表，不在终端中暴露完整 JSON。
|
||||
|
||||
[仅面向自动化的 ACP（Agent Client Protocol）传输层](../simplification/2026-07-23-acp-automation-only-protocol.md)有意不挂载会话查询或会话引用服务。
|
||||
|
||||
@@ -45,15 +45,15 @@ TUI 把会话候选与现有 `@` 文件提供方组合在一起。每个候选
|
||||
- **等待 SQLite FTS5**：不予采纳，因为快照正确性依赖按准确 id 读取和规范表层折叠，而不是内容搜索。FTS 只改进候选发现。
|
||||
- **把提及标记语法放入 `Agent.send()`**：不予采纳，因为这会迫使核心协议解析某个宿主的展示语法，并阻止带类型的非文本宿主复用同一语义层。
|
||||
- **在每个宿主中分别实现引用**：不予采纳，因为投影、安全警告、保留策略和持久化会在不同宿主之间逐渐偏离。
|
||||
- **在提示词旁放置单独的用户角色上下文消息**：不予采纳，因为相邻的两条用户消息会削弱提示词的指示语绑定：在 `@foo what does this session discuss?` 中，模型可能把「this session」解析为当前对话，而不是被引用的快照。
|
||||
- **在调用 `send()` 前由宿主合并前缀**：不予采纳，因为 `agent/prompt-submit` 必须只检查和改写直接提示词。最终生效的提示词与附加上下文只有在 AgentLoop 接纳后才汇合；此时 AgentLoop 可以把 `allow.content` 改写一致应用于合并后的模型内容和 `envelope.displayContent`。若由宿主更早组装，就会向该钩子暴露快照字节，或使这两个视图发生偏离。
|
||||
- **把上下文附加到 `SendOptions` 和收件箱记录**：不予采纳，因为通用投递将不得不负责贯穿准入、steering、取消和观察的领域事务。领域专用的准入包装层和现有 next-step outbox 可以保持所需配对，而无需扩大每条消息。
|
||||
- **在调用 `send()` 前由宿主合并前缀**：不予采纳，因为 `agent/prompt-submit` 必须只检查和改写直接提示词。将快照保留为独立的带来源消息，可以维持该边界，并让 TUI 从直接用户气泡中隐藏背景字节。
|
||||
- **回放原始源日志或恢复被遮蔽的事件**：不予采纳，因为压缩定义了当前模型表层，并且可能有意淘汰敏感或开销高昂的历史内容。
|
||||
- **恢复或 fork 源会话**：不予采纳，因为本功能只为一条目标消息提供只读背景，不提供身份或生命周期连续性。
|
||||
- **在请求时重新读取源会话并注入**：不予采纳，因为这会让引用变得不确定，取消竞态可能改变其字节内容，目标回放也会依赖可变的外部状态。
|
||||
|
||||
## 验证
|
||||
|
||||
单元与集成测试覆盖 URI 无损往返与文本边界标点、显式格式错误的引用、会考虑标题的候选排序、终端控制字符转义、投影排除规则、提示词封套的非递归投影、与后端无关的压缩检查点、标签安全封套、去重、自引用、数量限制、读取的全有或全无、存储读取不结束时取消提示词、逐源独立字节保留、冻结的消息所有权、提示词阻止、send／steer 放置方式、标题隔离、功能缺失和精简的 TUI 回放。无密钥 TUI 快照会运行真实的 agent loop（智能体循环）：源表层用一个压缩检查点替换旧的用户／assistant 历史，目标会话提交一个提及标记，捕获到的模型请求只包含一条用户消息，其中依次为快照、请求分隔符和当前提示词，并且不包含任一被遮蔽的字符串。
|
||||
单元与集成测试覆盖 URI 无损往返与文本边界标点、显式格式错误的引用、会考虑标题的候选排序、终端控制字符转义、投影排除规则、快照的非递归投影、与后端无关的压缩检查点、标签安全封套、去重、自引用、数量限制、读取的全有或全无、存储读取不结束时取消提示词、逐源独立字节保留、提示词阻止、准入期间的暂存、send／steer 放置方式、标题隔离、功能缺失和精简的 TUI 回放。无密钥 TUI 快照会运行真实的 agent loop（智能体循环）：源表层用一个压缩检查点替换旧的用户／assistant 历史，目标会话提交一个提及标记，捕获到的模型请求包含一条带来源的快照消息，后面跟随可读的当前提示词，并且不包含任一被遮蔽的字符串。
|
||||
|
||||
## 后果
|
||||
|
||||
|
||||
@@ -2,5 +2,5 @@
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-21-tui-skill-slash-command.md: d7532a05fce5605491ce42c87a2a523eb4c19acc
|
||||
2026-07-21-tui-skill-slash-command.zh.md: 16930020bd404f7bc9476169cd1d063aa57b5c94
|
||||
2026-07-21-tui-skill-slash-command.md: 8370ab61f552a6a60177b6da0b598dd142d21960
|
||||
2026-07-21-tui-skill-slash-command.zh.md: 66edec6ecd5c2a974b56e08bd9e924729302cc4a
|
||||
|
||||
@@ -14,7 +14,7 @@ The [`@deepseek-ai/dsh-tui`](../../../../packages/ui/tui/README.md) front door o
|
||||
|
||||
The TUI reads the skill service through `ctx.get('skills')`, not a declared injection, because skills mount conditionally: a deployment without the registry keeps a working front door, and `/skill:` there reports that skills are unavailable rather than failing to mount. `createTuiChat` is synchronous while `ctx.skills.list()` is async, so autocomplete seeds the static slash commands immediately and rebuilds the provider with `skill:<name>` entries once the catalog resolves; a resolution that arrives after disposal is dropped, and a rejected lookup keeps the base commands.
|
||||
|
||||
Autocomplete lists only model-invocable skills — it is built from `list()`, which omits `disableModelInvocation` skills — while manual submission resolves through `get()`, which the skill registry documents as the trusted-caller path that returns disabled skills too. So a person can load any skill by typing its exact name, but the completion menu never advertises a skill the model is meant not to see. An unknown name, an empty name after the prefix, and a lookup failure each surface as a transcript notice without sending anything.
|
||||
Autocomplete lists only model-invocable skills — it is built from `list()`, which omits `disableModelInvocation` skills — while manual submission resolves through `get()`, which the skill registry documents as the trusted-caller path that returns disabled skills too. So a person can load any skill by typing its exact name, but the completion menu never advertises a skill the model is meant not to see. Each completion entry is labeled with its winning source's scope — `(project)` for the `project-` sources, `(user)` for every other source — in the slash-command argument-hint slot, which the menu shows but selection never inserts, so trailing instructions still follow the completed name. An unknown name, an empty name after the prefix, and a lookup failure each surface as a transcript notice without sending anything.
|
||||
|
||||
`renderSkillInvocation` and the resource-base line are the TUI's own, deliberately not reused from `dsh-tool-skill`'s `skill` tool result. The tool wraps a body in `<skill_content>`/`<skill_resources>`/`<skill_instructions>` for a *tool result*; a manual invocation is a *user turn*, and coupling the two renderers would force one model-facing shape to serve both surfaces. The cost is two renderers that both format a skill body; the benefit is that each surface's model-facing text evolves independently, and each is pinned where it is produced.
|
||||
|
||||
|
||||
@@ -14,7 +14,7 @@ Status: implemented
|
||||
|
||||
TUI 通过 `ctx.get('skills')` 读取 skill 服务，而非声明式注入，因为 skill 是条件挂载的：没有注册表的部署仍保有可用的前门，此时 `/skill:` 会报告 skill 不可用，而不是挂载失败。`createTuiChat` 是同步的，而 `ctx.skills.list()` 是异步的，所以自动补全先立即种入静态斜杠命令，待目录解析完成后再用 `skill:<name>` 条目重建 provider（提供方）；在 dispose（资源释放）之后才到达的解析结果会被丢弃，而被拒绝的查找会保留基础命令。
|
||||
|
||||
自动补全只列出模型可调用的 skill——它基于 `list()` 构建，而 `list()` 会略去 `disableModelInvocation` 的 skill——手动提交则通过 `get()` 解析，skill 注册表将其记录为返回被禁用 skill 的可信调用方路径。因此用户可以通过键入 skill 的确切名称加载任意 skill，但补全菜单绝不会宣传一个本不该让模型看见的 skill。未知名称、前缀之后为空的名称、以及查找失败，都会各自呈现为 transcript（文本记录）中的一条通知，且不发送任何内容。
|
||||
自动补全只列出模型可调用的 skill——它基于 `list()` 构建，而 `list()` 会略去 `disableModelInvocation` 的 skill——手动提交则通过 `get()` 解析，skill 注册表将其记录为返回被禁用 skill 的可信调用方路径。因此用户可以通过键入 skill 的确切名称加载任意 skill，但补全菜单绝不会宣传一个本不该让模型看见的 skill。每个补全条目都以其胜出来源的作用域为标签——`project-` 来源标为 `(project)`，其他一切来源标为 `(user)`——标签置于斜杠命令的参数提示位，菜单会显示它，但选中时绝不会插入，因此尾随指令仍然跟在补全后的名称之后。未知名称、前缀之后为空的名称、以及查找失败，都会各自呈现为 transcript（文本记录）中的一条通知，且不发送任何内容。
|
||||
|
||||
`renderSkillInvocation` 及资源基址行是 TUI 自有的，刻意不复用 `dsh-tool-skill` 的 `skill` 工具结果。该工具把正文包进 `<skill_content>`/`<skill_resources>`/`<skill_instructions>` 是为了一个*工具结果*；而手动调用是一个*用户轮次*，把两个渲染器耦合起来会迫使一种面向模型的形态同时服务两个界面。代价是两个都在格式化 skill 正文的渲染器；收益是各界面面向模型的文本可以独立演进，且各自在其产出处被固定。
|
||||
|
||||
|
||||
@@ -0,0 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-07-23-session-telemetry-otel-revival.md
|
||||
2026-07-23-session-telemetry-otel-revival.md: a58598d8a956d47cb0cf6aa3e659f38314bc4b17
|
||||
2026-07-23-session-telemetry-otel-revival.zh.md: cc09717e349d5ae2ab5157bf46de30b1823c775f
|
||||
@@ -0,0 +1,37 @@
|
||||
# Agent Note: Session telemetry seam with mandatory redaction and the OTel backend
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-07-23-session-telemetry-otel-revival.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
Every deployment that wants harness sessions in an observability stack must hand-roll a session-log consumer: subscription, lifecycle handoff, and — hardest — redaction, since the raw log carries file contents and command output that may embed credentials. A telemetry seam and OTel backend shipped once on the `session-telemetry-otlp-rfc` branch (PR #222/#231) but never reached master: the proposal exported raw session events verbatim, which legal review declined. The capture-side design (backend contract, coordinator, handoff cursor, chunk projection) was sound and reviewed; the export-side stance was the blocker.
|
||||
|
||||
## Decision
|
||||
|
||||
`packages/telemetry/` revives the two reviewed packages under the SDK stance — the harness provides the capability, the deployment configures where records go and owns what leaves in them:
|
||||
|
||||
- **`@deepseek-ai/dsh-session-telemetry`** — the seam. `TelemetryBackend` (`emit`/`flush?`/`shutdown`), the service-registered `Telemetry` form, and `TelemetryCoordinator` owning capture: adoption with cursor read-back, the per-append firehose (project → `structuredClone` → redact → `emit`, zero I/O), the fixed first-chunk-per-(turn, step) projection, the `agent/error` relay, and dispose-time `shutdown` records.
|
||||
- **The `telemetry/record` waterfall** — the delta over the branch version and the seam's redaction extension point. Every record passes it before reaching any backend; the seam ships NO rules of its own — the innermost `next()` is a pass-through, deployments mount their rules as listeners (stacking by transforming `next()`'s return value), and a throwing rule withholds the record fail-closed. Redaction applies to the exported copy only; the canonical log is never rewritten.
|
||||
- **`@deepseek-ai/dsh-session-telemetry-otel`** — the reference backend: OTel JS SDK log pipeline (`LoggerProvider` → `BatchLogRecordProcessor` → OTLP/HTTP exporter), configured verbatim through `exporter`/`processor` passthroughs. `exporter.url` is required and validated at load; unmounted or unconfigured, nothing leaves the process.
|
||||
|
||||
The boundary axiom holds: the harness's aspect ends at `emit()`. Batching, retry, queueing, and loss policy are the reporting SDK's, configured through passthroughs — delivery is best-effort (at-most-once across a crash), which the READMEs state plainly.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
**Implement the runtime-telemetry RFC's outbox (durable spool, per-sink cursors, at-least-once, a `readCommitted` persistence-seam method).** Deferred, not rejected: the SDK stance makes delivery semantics the reporting SDK's territory, and the OTel SDK's own batch pipeline is the honest default. The outbox is a pure additive layer (the `emit()` contract does not move); revive it when a deployment states a crash-loss requirement telemetry must satisfy.
|
||||
|
||||
**No in-process redaction point, delegating to receiver-side collector processors.** Rejected — receiver-side redaction ships the secret first and scrubs it second. The waterfall puts an auditable, stackable scrubbing point before bytes leave the process; where the branch version (what PR #222 shipped) had no redaction point at all, every record now passes one.
|
||||
|
||||
**A built-in conservative rule set as the waterfall's innermost `next()`.** Rejected: as an SDK we cannot know which patterns are secrets in a given deployment, a shipped list invites false confidence ("redaction is on") while catching only known shapes, and false positives would corrupt exported bodies for consumers who never asked. The seam owns the mechanism; the deployment owns the policy — the innermost `next()` is a pass-through, and rules mount as listeners.
|
||||
|
||||
**Map onto OTel spans (GenAI semantic conventions) instead of logs.** Rejected for this revival: the branch implementation's log mapping is reviewed and shipped-shaped; the span model is lossy for forkable, interruptible sessions and belongs to a future consumer with real span queries to serve.
|
||||
|
||||
**Full-log replay when no handoff cursor survived (re-export constructor seeds).** Shipped in the first revival round, then narrowed: adoption now replays from the session's construction boundary (`Session.firstLiveSeq`, the constructor-seed length — a fact the session already validated but did not expose; `header.seedLength` cannot serve, it is the durable fork-lineage value and a resumed session's constructor seed is its full stored log). A resumed session's history already shipped from the previous process under the same id, and a fork's inherited prefix already shipped in the parent's stream — re-exporting either re-billed every resume for its full history and doubled query-time counts on OTLP backends with no native ingest dedupe. Receivers stitch fork lineage via `session.parent_id` + `session.seed_length`. What the narrowing gives up, consistently with the at-most-once stance: a resume no longer backfills records the previous process failed to deliver (telemetry unmounted then, or queued at crash) — the full replay's only real benefit, bought at the common case's expense. A deployment that states a backfill requirement needs the deferred outbox above, not replay. The boundary also swallows the synthetic turn closers `SessionPersistence.load()` writes when repairing a crash-interrupted log (they sit below `firstLiveSeq` despite never existing in the previous process) — deliberate, not incidental: exporting a synthetic closer cannot complete the remote turn whose real tail records died in the crashed process's queue, it can only make an incomplete turn look closed. The wire stream stays faithful to what the crashed process actually shipped; receivers read a never-closed turn on a resumed stream as "the previous process died inside it" (the OTel README states the rule), and a later clean `shutdown` marker attests only to the resumed process's exit. Threading the pre-repair boundary through load/prepare so repairs export as live events would couple three packages to un-ship that signal.
|
||||
|
||||
**Forwarding the seam's turn-boundary `flush()` hint to the OTel provider's `forceFlush()`.** Shipped in the first revival round, then removed after three review rounds each found a new silent-loss path in the same wrapper state: a dispose racing an in-flight flush (the SDK's concurrent-flush guard makes shutdown's internal drain skip), overlapping hints displacing the retained promise, and the provider's fixed 30-second flush timeout rejecting while the processor still drains. Every path exists only because the forwarding made this backend the process's second flusher against undocumented SDK internals from the upstream experimental tree; with no `flush()` implemented, the batch processor is the only flusher, its `scheduledDelayMillis` (already deployment-tunable through the `processor` passthrough) governs export cadence, and `shutdown()`'s drain is complete by construction. Reinstate only if a deployment states a turn-boundary latency requirement `scheduledDelayMillis` cannot meet — and then by calling the retained `BatchLogRecordProcessor`'s own `forceFlush()`, never the provider's timeout-wrapped one.
|
||||
|
||||
## Consequences
|
||||
|
||||
A deployment adds one `cordis.yml` entry with an OTLP endpoint and gets its session stream in any OTel-compatible stack; removing the entry is the opt-out, with no residual state. A rule-free deployment exports records exactly as captured — including any credentials embedded in file contents or command output — so a deployment crossing a trust boundary must mount `telemetry/record` listeners, and both READMEs state this plainly. Where rules are mounted, exported bodies can differ from canonical log bytes, so receivers must not treat telemetry as a byte-exact replica; the log remains the source of truth. Crash durability is explicitly out of scope until the outbox decision above is revisited.
|
||||
@@ -0,0 +1,37 @@
|
||||
# Agent Note: Session telemetry seam with mandatory redaction and the OTel backend
|
||||
|
||||
Status: implemented
|
||||
|
||||
[English](2026-07-23-session-telemetry-otel-revival.md) | 中文
|
||||
|
||||
## Problem
|
||||
|
||||
每个想把 harness 会话接入可观测性体系的部署方都得手写一套会话日志消费端：订阅、生命周期交接、以及最难的脱敏——原始日志携带文件内容与命令输出，可能内嵌凭据。遥测 seam 和 OTel backend 曾在 `session-telemetry-otlp-rfc` 分支（PR #222/#231）上完成过一版，但从未进入 master：该提案将原始会话事件原样导出，法务评审未予通过。捕获侧设计（backend 契约、coordinator、handoff 游标、chunk 投影）本身合理且经过评审；导出侧的立场才是阻塞点。
|
||||
|
||||
## Decision
|
||||
|
||||
`packages/telemetry/` 以 SDK 立场复活这两个经过评审的包——harness 提供能力，部署方配置上报去向并对导出内容负责：
|
||||
|
||||
- **`@deepseek-ai/dsh-session-telemetry`** —— seam 本体。`TelemetryBackend`（`emit`/`flush?`/`shutdown`）、服务注册形态的 `Telemetry`、以及拥有捕获侧的 `TelemetryCoordinator`：带游标回读的收养、逐 append 的 firehose（投影 → `structuredClone` → 脱敏 → `emit`，零 I/O）、固定的每 (turn, step) 首 chunk 投影、`agent/error` 转发、以及 dispose 时的 `shutdown` 记录。
|
||||
- **`telemetry/record` waterfall** —— 相对分支版本的增量，也是该 seam 的脱敏扩展点。每条记录抵达任何 backend 前必经此处；seam 自身不带任何规则——最内层 `next()` 原样透传，部署方以监听器挂载自己的规则（通过变换 `next()` 的返回值堆叠），抛异常的规则将该记录 fail-closed 扣下。脱敏只作用于导出副本；canonical log 永不改写。
|
||||
- **`@deepseek-ai/dsh-session-telemetry-otel`** —— 参考 backend：OTel JS SDK 日志管线（`LoggerProvider` → `BatchLogRecordProcessor` → OTLP/HTTP exporter），经 `exporter`/`processor` passthrough 原样配置。`exporter.url` 必填且加载时校验；未挂载或未配置时，任何数据都不会离开进程。
|
||||
|
||||
边界公理保持不变：harness 的职责止于 `emit()`。批处理、重试、排队与丢失策略属于 reporting SDK，经 passthrough 配置——投递是尽力而为（崩溃时至多一次），README 对此如实陈述。
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
**实现 runtime-telemetry RFC 的 outbox（落盘 spool、每 sink 游标、at-least-once、persistence seam 的 `readCommitted` 方法）。** 推迟而非否决：SDK 立场使投递语义归属 reporting SDK，OTel SDK 自身的批处理管线是诚实的默认。outbox 是纯增量层（`emit()` 契约不动）；待某个部署提出遥测必须满足的崩溃丢失要求时再复活。
|
||||
|
||||
**不设进程内脱敏点，交给接收端 collector processor。** 否决——接收端脱敏是先把秘密发出去再擦除。waterfall 在字节离开进程前提供一个可审计、可堆叠的擦除点；分支版本（PR #222 交付的形态）完全没有脱敏点，如今每条记录都必经其一。
|
||||
|
||||
**在 waterfall 最内层 `next()` 内置一套保守规则集。** 否决：作为 SDK 我们无法预知某个部署里什么模式算秘密，内置列表只覆盖已知形状却会带来"脱敏已开启"的虚假信心，且误报会替从未要求过的消费者破坏导出 body。seam 拥有机制，部署方拥有策略——最内层 `next()` 原样透传，规则以监听器挂载。
|
||||
|
||||
**映射到 OTel span（GenAI 语义约定）而非日志。** 本次复活否决：分支实现的日志映射已经过评审、形态可交付；span 模型对可 fork、可中断的会话有损，留给将来真正有 span 查询需求的消费者。
|
||||
|
||||
**handoff 游标未存活时全量回放日志（重新导出构造函数种子）。** 首轮复活曾交付此方案，其后收窄：收养现在从会话的构造边界起回放（`Session.firstLiveSeq`，即构造函数种子长度，这一事实会话早已校验过却未曾暴露；`header.seedLength` 不能胜任：它是持久保存的 fork 谱系（lineage）值，而恢复会话的构造函数种子是其完整的已存储日志）。恢复会话的历史已由上一个进程以同一 id 发出，fork 继承的前缀也已在父会话的流中发出；再次导出任何一者，都会让每次恢复为其完整历史重复付费，并在没有原生摄取去重的 OTLP 后端上使查询时的计数翻倍。接收端基于 `session.parent_id` + `session.seed_length` 拼接 fork 谱系。此次收窄放弃的内容与至多一次立场一致：恢复不再回填上一个进程未能投递的记录（彼时遥测未挂载，或崩溃时仍在队列中）——这本是全量回放唯一的真实收益，代价却由常见情形承担。提出回填要求的部署需要的是上文已推迟的 outbox，而不是回放。该边界同样吞掉 `SessionPersistence.load()` 修复被崩溃打断的日志时写入的合成轮次关闭事件（它们落在 `firstLiveSeq` 之前，尽管在上一个进程中从未存在过）。这是有意为之，而非附带效果：远端轮次的真实尾部记录已随崩溃进程的队列一同消亡，导出合成关闭事件无法补全该轮次，只会让一个未完成的轮次看起来已经关闭。导出的流忠实于崩溃进程实际发出的内容；接收端会把恢复后的流中一个从未关闭的轮次读作「上一个进程死在了该轮次之内」（OTel README 陈述了这条规则），其后干净的 `shutdown` 标记也只证明恢复后进程自身的退出。若为让修复以实时事件的身份导出而将修复前边界贯穿 load/prepare 传递，将使三个包相互耦合，只为抹除这一信号。
|
||||
|
||||
**将 seam 的轮次边界 `flush()` 提示转发到 OTel provider 的 `forceFlush()`。** 首轮复活曾交付此转发，其后移除：三轮评审在同一份包装层状态中各发现一条新的静默丢失路径——dispose 与进行中的 flush 之间的竞态（SDK 的并发 flush 防护会令 shutdown 的内部排空被跳过）、相互重叠的提示顶掉留存的 promise、以及 provider 固定的 30 秒 flush 超时在批处理器仍在排空时便 reject。这些路径存在的唯一原因，是该转发让这个后端成为进程内第二个执行 flush 的组件，面对的还是上游实验性（experimental）源码树中未见诸文档的 SDK 内部行为；不实现 `flush()` 时，批处理器就是唯一执行 flush 的组件，其 `scheduledDelayMillis`（已可由部署方经 `processor` passthrough 调优）决定导出节奏，`shutdown()` 的排空从构造上就是完整的。仅当某个部署提出 `scheduledDelayMillis` 无法满足的轮次边界延迟要求时才恢复此转发——且届时应调用留存的 `BatchLogRecordProcessor` 自身的 `forceFlush()`，绝不调用 provider 那个带超时包装的版本。
|
||||
|
||||
## Consequences
|
||||
|
||||
部署方在 `cordis.yml` 加一个带 OTLP endpoint 的条目即可把会话流接入任何 OTel 兼容体系；删除条目即退出，无残留状态。未挂载规则的部署导出的记录与捕获时完全一致——包括文件内容与命令输出中内嵌的任何凭据——因此跨信任边界的部署必须挂载 `telemetry/record` 监听器，两个 README 对此如实陈述。挂载规则后，导出的 body 可能与 canonical log 字节不同，接收端不得把遥测当作字节精确副本；日志仍是唯一事实源。崩溃持久性在上述 outbox 决定重启前明确不在范围内。
|
||||
@@ -0,0 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-23-tui-footer-session-identity.md: aa17ead4194c52464de0caad86d8611eae94786c
|
||||
2026-07-23-tui-footer-session-identity.zh.md: 686c11294ffd02304dc87cd790252a347fe35011
|
||||
@@ -0,0 +1,27 @@
|
||||
# Agent Note: Keep the TUI session identity visible
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-07-23-tui-footer-session-identity.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
The startup banner identifies the active session, but it scrolls out of view during a conversation. Operators working with several resumable sessions then lack a persistent way to confirm which session receives their input.
|
||||
|
||||
## Decision
|
||||
|
||||
The TUI footer begins with the active session id, before the model, working directory, token counts, cache rate, and context use. It shows tool-card state only while cards are expanded; the default collapsed state adds no label. The session id uses the same control-character escaping as other terminal labels and participates in the footer's existing left-to-right clipping behavior.
|
||||
|
||||
The footer reads the id from the mounted agent's session, so fresh and resumed sessions use the same authoritative identity without separate UI state.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
- **Keep the identity only in the startup banner** — rejected because the banner leaves the viewport in longer conversations.
|
||||
- **Show the session id only in `/status`** — rejected because an on-demand diagnostic does not let an operator confirm identity before sending input.
|
||||
- **Put the session id in the right footer segment** — rejected because narrow terminals clip that segment first; session identity is more important than context and expanded tool-card state.
|
||||
|
||||
## Consequences
|
||||
|
||||
The current session remains identifiable while the editor is active. On narrow terminals, the longer left segment leaves less room for context and the expanded tool-card label, while the existing clipping policy preserves session identity, model, and as much operational context as fits.
|
||||
|
||||
Package coverage pins the footer ordering and escaping path, and the runnable TUI terminal snapshots pin the assembled layout.
|
||||
@@ -0,0 +1,27 @@
|
||||
# Agent Note: 保持 TUI 会话标识可见
|
||||
|
||||
[English](2026-07-23-tui-footer-session-identity.md) | 中文
|
||||
|
||||
Status: implemented
|
||||
|
||||
## Problem
|
||||
|
||||
启动横幅会标识当前会话，但在对话过程中会滚出视野。操作多个可恢复会话时，用户因而无法持续确认输入将发送到哪个会话。
|
||||
|
||||
## Decision
|
||||
|
||||
TUI 页脚以当前会话 id 开头，之后依次显示模型、工作目录、token 用量、缓存命中率和上下文用量。工具卡片状态仅在卡片展开时显示；默认的折叠状态不添加任何标签。会话 id 与其他终端标签采用相同的控制字符转义，并遵循页脚现有的从左到右裁剪行为。
|
||||
|
||||
页脚从已挂载 agent 的会话读取 id，因此新建和恢复的会话都使用同一权威标识，无需单独维护 UI 状态。
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
- **仅在启动横幅中保留标识** — 未采用，因为对话较长时横幅会离开视野。
|
||||
- **仅在 `/status` 中显示会话 id** — 未采用，因为按需诊断无法让用户在发送输入前确认会话标识。
|
||||
- **将会话 id 放入页脚右侧区域** — 未采用，因为窄终端会优先裁剪该区域；会话标识比上下文和展开的工具卡片状态更重要。
|
||||
|
||||
## Consequences
|
||||
|
||||
编辑器处于活动状态时，当前会话始终可识别。在窄终端中，更长的左侧区域会减少上下文和展开的工具卡片标签的显示空间；现有裁剪策略会保留会话标识、模型，以及空间允许的其他运行信息。
|
||||
|
||||
包级覆盖固定页脚顺序和转义路径，可运行 TUI 的终端快照固定组装后的布局。
|
||||
@@ -2,5 +2,5 @@
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-06-25-ask-user-question.md: 51b9a67bfc0fdc88d84a9fa51662d02b127f7925
|
||||
2026-06-25-ask-user-question.zh.md: 4bd2c2b8664ba085919eb9d4b8c5dee92d4bf72e
|
||||
2026-07-23-tui-status-prompt-tools.md: 42524d021d0f2786371762b447ad5d195dc828bd
|
||||
2026-07-23-tui-status-prompt-tools.zh.md: 5a33e19e9780749a721395a0b07f43790103013c
|
||||
@@ -0,0 +1,29 @@
|
||||
# Agent Note: TUI status inspects model request inputs
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-07-23-tui-status-prompt-tools.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
Session counters describe activity but do not reveal the instructions and capabilities that the next model request receives. Diagnosing scoped prompt contributions and tool restrictions otherwise requires leaving the TUI or inferring configuration from files.
|
||||
|
||||
## Decision
|
||||
|
||||
`/status` assembles the current agent's system prompt through `ctx.systemPrompt` and renders it with the same renderer used by the agent loop. After the bordered diagnostics card, separate unbordered `System prompt` and `Registered tools` sections show the rendered prompt and the assembly's ordered tool names, which are the schemas exposed to the model for that agent and presentation mode.
|
||||
|
||||
Assembly uses the command's cancellation signal and current agent scope, so scoped sections, variables, tool restrictions, and assembly listeners match a request made at that point. Prompt and tool values are escaped through the TUI's terminal-control sanitizer before rendering. Empty prompt text and an empty tool list render as `(empty)` and `(none)`.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
**Read prompt sections and the tool registry independently.** Rejected: that bypasses prompt assembly waterfalls, tool ordering, presentation modes, and per-agent restrictions, so the diagnostics could disagree with the next request.
|
||||
|
||||
**Show complete tool schemas.** Rejected: names answer which capabilities are registered without making the status card dominated by parameter JSON; schema details remain available in the generated tool catalog and source definitions.
|
||||
|
||||
## Consequences
|
||||
|
||||
The command can run prompt providers and assembly listeners, just like request preparation, and reports their failures through the existing command-error notice. The snapshot is point-in-time: a later registration, restriction, mode change, or dynamic provider can alter the next request.
|
||||
|
||||
## Testing
|
||||
|
||||
Unit coverage pins scoped assembly output, ordered tool names, empty labels, and terminal-control escaping. The keyless TUI smoke and terminal snapshot exercise `/status` through the assembled application.
|
||||
@@ -0,0 +1,29 @@
|
||||
# Agent Note: TUI 状态检查模型请求输入
|
||||
|
||||
Status: implemented
|
||||
|
||||
[English](2026-07-23-tui-status-prompt-tools.md) | 中文
|
||||
|
||||
## 问题
|
||||
|
||||
会话计数器可以描述活动情况，却无法显示下一次模型请求将收到的指令和能力。若要诊断按作用域贡献的提示词与工具限制，用户只能离开 TUI，或根据配置文件进行推断。
|
||||
|
||||
## 决策
|
||||
|
||||
`/status` 通过 `ctx.systemPrompt` 为当前 agent（智能体）组装系统提示词，并使用与 agent loop（智能体循环）相同的渲染器完成渲染。在带边框的诊断卡片之后，独立且无边框的 `System prompt` 和 `Registered tools` 区域分别显示渲染后的提示词与 assembly 中按顺序排列的工具名称；这些名称对应当前 agent 与呈现模式向模型公开的 schema。
|
||||
|
||||
组装使用命令的取消信号和当前 agent 作用域，因此按作用域注册的 section、变量、工具限制及 assembly listener 与此时发起的请求保持一致。提示词和工具值在呈现前经过 TUI 的终端控制字符净化。空提示词与空工具列表分别显示为 `(empty)` 和 `(none)`。
|
||||
|
||||
## 曾考虑的替代方案
|
||||
|
||||
**分别读取提示词 section 和工具注册表。** 已否决：该做法会绕过提示词组装 waterfall（瀑布式事件）、工具排序、呈现模式和按 agent 限制，因此诊断结果可能与下一次请求不一致。
|
||||
|
||||
**显示完整工具 schema。** 已否决：工具名称足以回答注册了哪些能力，同时避免参数 JSON 占据大部分状态卡片；schema 详情仍可在生成的工具目录和源代码定义中查看。
|
||||
|
||||
## 后果
|
||||
|
||||
该命令可能像请求准备一样运行提示词提供方与 assembly listener，并通过现有命令错误提示报告失败。结果是一个时点快照：后续注册、限制、模式变更或动态提供方都可能改变下一次请求。
|
||||
|
||||
## 测试
|
||||
|
||||
单元测试固定按作用域组装的输出、工具名称顺序、空值标签和终端控制字符转义。无密钥 TUI 冒烟测试与终端快照通过完整组装的应用执行 `/status`。
|
||||
@@ -0,0 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-24-configurable-tui-prompt-theme.md: 4008f23a3f545e9b4484f0fa3f8490ad9b2c5541
|
||||
2026-07-24-configurable-tui-prompt-theme.zh.md: daf15b54d7e04d3860eacad47b754b963cde36fa
|
||||
@@ -0,0 +1,39 @@
|
||||
# Agent Note: TUI prompt themes compose mutable plugin values
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-07-24-configurable-tui-prompt-theme.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
The terminal prompt row and editor prefix were assembled inside the TUI from a fixed set of workspace, model, usage, cache, context, and timing fields. Deployments could change colors globally but could not choose field order, replace the input prefix, add plugin state, or build a Powerline prompt.
|
||||
|
||||
## Decision
|
||||
|
||||
The TUI theme groups `color`, `truecolor`, `leftPrompt`, `rightPrompt`, `inputPrompt`, and the static running-state `inputPlaceholder`. The three prompt strings interpolate `${name}` references; unknown or unavailable values disappear with adjacent horizontal separator whitespace. The left and right templates share one row, retain the right side on overlap, and use ANSI-aware visible widths. The input template controls the first-line editor prefix and continuation indentation.
|
||||
|
||||
`ctx.tuiPrompt` is a context-global registry supplied by `@deepseek-ai/dsh-tui/prompt`. `register(name, initialValue)` returns a handle with `set(value)` and `dispose()`. Values are stored strings rather than callbacks: updates are explicit, unchanged strings are ignored, and a registration, mutation, or disposal schedules one coalesced notification. The renderer reads current values with `get(name)` and subscribes with `subscribe(listener)` to learn when to redraw. That subscription is a direct in-service callback, not a Cordis event, so a value changing on its own schedule still repaints without a bus entry other consumers would never use. Both `subscribe` and each registration are owned by the caller's Cordis effect, so they are removed when the subscriber's or contributor's fiber disposes. Each `subscribe` call is a distinct subscription keyed by record identity, so two fibers may pass the same callback and disposing one leaves the other live. The coalesced notification contains every observer — a synchronous throw, a rejected returned promise, and even an error hostile to string rendering (logs go through the non-throwing `errorChain`) — so one broken observer cannot starve the rest, and it re-checks each subscription's liveness during delivery so a listener that synchronously unsubscribes another in the same burst silences it immediately. Registration follows Cordis effect ownership, rejects duplicate names, and removes the value on plugin disposal.
|
||||
|
||||
Registered fragments are trusted ANSI-capable presentation output. Template literals and ordinary external content remain sanitized, but a prompt-value plugin may emit terminal controls. Composite values own coordinated background transitions and separators, so one `${powerline}` value can render a complete Powerline segment without coupling adjacent atomic providers.
|
||||
|
||||
The built-in `cwd`, `git/worktree`, `token_meter/cache_hit_rate`, `model`, `context`, `timing`, styled `symbol` label, and `indicator` caret values use the same registry. Session and agent events update their handles, while the running timer updates `timing` and the animated `indicator` each tick. The shipped input template is `${symbol} ${indicator}`, preserving the existing `dsh > ` prefix.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
**Evaluate synchronous provider callbacks on every render.** Rejected: render-time plugin code adds an avoidable failure boundary; stored strings keep the render pass free of plugin evaluation.
|
||||
|
||||
**Publish the change notification as a Cordis event.** Rejected: the notification has exactly one consumer (the TUI renderer for the current session), so a global typed event adds a bus entry, scoped-dispatch surface, and cross-plugin fan-out no one else observes. A direct `subscribe` callback contained inside the service carries the same coalesced redraw with less surface.
|
||||
|
||||
**Expose semantic style roles instead of ANSI.** Rejected: semantic roles cannot express arbitrary Powerline background transitions without expanding the shared style protocol for each presentation technique.
|
||||
|
||||
**Put prompt fields at the top level of TUI config.** Rejected: templates and color selection jointly define terminal presentation and belong under one `theme` object.
|
||||
|
||||
## Consequences
|
||||
|
||||
Prompt contributors depend on the TUI-specific registry and are loaded after the service but before the TUI consumer. The namespace is global to the Cordis context, matching the TUI's current single-session transcript ownership. Arbitrary ANSI is intentionally trusted: unsupported cursor-affecting sequences can disrupt layout, and alignment is reliable only for sequences understood by pi-tui's visible-width utilities.
|
||||
|
||||
Changing `inputPrompt` through a registered value preserves editor text, cursor, history, completion, and focus because pi-tui supports replacing equal-width first and continuation prefixes in place. The static `inputPlaceholder` is sanitized and appears only while the agent runs and the editor is empty.
|
||||
|
||||
## Testing
|
||||
|
||||
Registry tests pin validation, duplicate rejection, updates, unavailable values, coalesced-notification containment, unsubscribe, disposal, interpolation, trailing-literal retention, whitespace cleanup, and ANSI preservation. TUI tests pin nested theme defaults, custom templates, out-of-band value redraw, mutable redraw, Powerline-capable fragments, dynamic input-prefix width, and the static running placeholder. The assembled TUI demo test pins service load order and config forwarding.
|
||||
@@ -0,0 +1,39 @@
|
||||
# Agent Note: TUI 提示符主题组合可变的插件值
|
||||
|
||||
Status: implemented
|
||||
|
||||
[English](2026-07-24-configurable-tui-prompt-theme.md) | 中文
|
||||
|
||||
## 问题
|
||||
|
||||
终端提示符行与编辑器前缀原先在 TUI 内部由一组固定字段拼装而成，涵盖工作区、模型、用量、缓存、上下文与计时。部署方可以全局更改颜色，却无法调整字段顺序、替换输入前缀、加入插件状态，也无法构建 Powerline 风格的提示符。
|
||||
|
||||
## 决策
|
||||
|
||||
TUI 主题把 `color`、`truecolor`、`leftPrompt`、`rightPrompt`、`inputPrompt` 以及运行状态下的静态 `inputPlaceholder` 归为一组。三个提示符字符串通过插值引用 `${name}`；未知或不可用的值连同相邻的横向分隔空白一起消失。左右模板共用一行，重叠时保留右侧，宽度计算使用可识别 ANSI 的可见宽度。输入模板控制编辑器首行前缀与续行缩进。
|
||||
|
||||
`ctx.tuiPrompt` 是由 `@deepseek-ai/dsh-tui/prompt` 提供的上下文全局注册表。`register(name, initialValue)` 返回带 `set(value)` 与 `dispose()` 的句柄。存储的值是字符串而非回调：更新必须显式发起，未变化的字符串会被忽略，而一次注册、变更或 dispose 会安排一次合并后的通知。渲染器用 `get(name)` 读取当前值，并用 `subscribe(listener)` 订阅何时重绘。该订阅是服务内部的直接回调，而非 Cordis 事件，因此一个自行变化的值仍能重绘，而不需要一个其他消费方永远不会观察的总线条目。`subscribe` 与每个注册都由调用方的 Cordis effect 拥有，因此在订阅方或贡献方的 fiber dispose 时一并移除。每次 `subscribe` 都是一个按记录身份区分的独立订阅，因此两个 fiber 可以传入同一个回调，而 dispose 其中一个不会影响另一个。合并通知会容错每个观察者——同步抛出、返回被拒 promise，甚至一个对字符串渲染也会抛异常的错误（日志走不抛异常的 `errorChain`）——因此一个损坏的观察者不会饿死其余观察者；并且在派发过程中会重新校验每个订阅的存活性，因此同一批次中同步取消了另一个订阅的监听器会立即使其静默。注册遵循 Cordis 的 effect 所有权模型，拒绝重复名称，并在插件 dispose（资源释放）时移除对应的值。
|
||||
|
||||
注册的片段被视为可信的、允许携带 ANSI 的呈现输出。模板中的字面文本与普通外部内容仍会被清洗，但提供提示符值的插件可以输出终端控制序列。复合值自行负责协调背景色过渡与分隔符，因此一个 `${powerline}` 值就能渲染完整的 Powerline 段，而无需与相邻的原子提供方耦合。
|
||||
|
||||
内置的 `cwd`、`git/worktree`、`token_meter/cache_hit_rate`、`model`、`context`、`timing`、带样式的 `symbol` 标签与 `indicator` 光标符值使用同一个注册表。会话与 agent（智能体）事件更新各自的句柄，运行计时器每一拍更新 `timing` 与带动画的 `indicator`。随附的输入模板为 `${symbol} ${indicator}`，保留了原有的 `dsh > ` 前缀。
|
||||
|
||||
## 曾考虑的替代方案
|
||||
|
||||
**每次渲染时求值同步的提供方回调。** 不予采纳：在渲染期执行插件代码会引入一个本可避免的故障边界；存储字符串能让渲染过程不涉及插件求值。
|
||||
|
||||
**把变更通知发布为 Cordis 事件。** 已否决：该通知只有一个消费方（当前会话的 TUI 渲染器），因此全局类型事件会增加一个总线条目、scope 分发面以及无人观察的跨插件扇出。服务内部包裹的直接 `subscribe` 回调以更小的面积承载同样的合并重绘。
|
||||
|
||||
**暴露语义化的样式角色而非 ANSI。** 不予采纳：语义角色无法表达任意的 Powerline 背景色过渡，除非为每种呈现技巧扩展共享的样式协议。
|
||||
|
||||
**把提示符字段放在 TUI 配置顶层。** 不予采纳：模板与颜色选择共同定义终端呈现，应归属于同一个 `theme` 对象之下。
|
||||
|
||||
## 后果
|
||||
|
||||
提示符值的贡献插件依赖 TUI 专属的注册表，加载顺序位于该服务之后、TUI 消费方之前。命名空间对整个 Cordis 上下文全局生效，与 TUI 当前的单会话 transcript（文本记录）所有权一致。允许任意 ANSI 是有意的信任决策：不受支持的、影响光标的序列可能破坏布局，只有 pi-tui 可见宽度工具能理解的序列才能保证对齐可靠。
|
||||
|
||||
通过注册值更改 `inputPrompt` 时，编辑器文本、光标、历史、自动补全与焦点均得以保留，因为 pi-tui 支持原地替换等宽的首行与续行前缀。静态的 `inputPlaceholder` 会被清洗，且仅在 agent 运行且编辑器为空时显示。
|
||||
|
||||
## 测试
|
||||
|
||||
注册表测试固定校验、重名拒绝、更新、不可用值、合并通知的容错、取消订阅、dispose、插值、尾随字面保留、空白清理与 ANSI 保留等行为。TUI 测试固定嵌套主题默认值、自定义模板、带外值重绘、可变重绘、支持 Powerline 的片段、动态输入前缀宽度以及运行状态下的静态占位文本。组装后的 TUI 演示测试固定服务加载顺序与配置转发。
|
||||
@@ -2,5 +2,5 @@
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-08-repeat-tool-guard.md: 67ec29c6c9fa38bf1d5935c469f3f71b1119dc3a
|
||||
2026-07-08-repeat-tool-guard.zh.md: 01037f29810c781c33beee414e50412a0c9b0f89
|
||||
2026-07-24-provider-retry-policies.md: 1831ce6b96178d11e7c9927ceccbe07ea578cd2c
|
||||
2026-07-24-provider-retry-policies.zh.md: 788f1f1963861e1b46ff0d8798e53e01bd9892b4
|
||||
@@ -0,0 +1,65 @@
|
||||
# Agent Note: Per-provider request retry policies
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-07-24-provider-retry-policies.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
One process may route model requests to providers with different reliability and cost constraints. A single transient classifier and finite retry budget cannot express a deployment that wants bounded recovery for most providers but requires one provider to keep retrying every model-request failure until the request succeeds or the caller cancels it.
|
||||
|
||||
Provider policy must follow the request that actually failed, including a route selected by `agent/request`, rather than the agent's initial options. Unbounded policy also cannot store JavaScript `Infinity` in the durable session event, and neither provider error text nor discarded partial output may enter the next model request.
|
||||
|
||||
## Decision
|
||||
|
||||
Each concrete adapter accepts an optional `retryPolicy` inside its provider configuration. The adapter validates and resolves the policy, and `ctx.llm` captures it when that exact provider route registers. When a call enters its final adapter boundary, `ctx.llm` binds the serving registration's immutable policy to that call; the agent loop passes it to closed-step recovery even if the route is disposed or replaced while the request is in flight. `@deepseek-ai/dsh-llm-retry` combines that call-local policy with the failed step's durable provider identity. A call that never reaches a final adapter has no serving policy and delegates. A provider without `retryPolicy` uses the normal defaults.
|
||||
|
||||
```yaml
|
||||
providers:
|
||||
- provider: deepseek
|
||||
retryPolicy:
|
||||
mode: normal
|
||||
maxRetries: 2
|
||||
retryableCodes: [EMPTY_RESPONSE, RATE_LIMIT, SERVER, TIMEOUT, TRANSPORT]
|
||||
backoff:
|
||||
initialDelayMs: 500
|
||||
maxDelayMs: 10000
|
||||
jitterRatio: 0.1
|
||||
- provider: internal
|
||||
retryPolicy:
|
||||
mode: always
|
||||
backoff:
|
||||
initialDelayMs: 1000
|
||||
maxDelayMs: 30000
|
||||
jitterRatio: 0.2
|
||||
```
|
||||
|
||||
The listener reads the provider from the durable `request/header` in force when the failed step closed, excluding later recovery mutations, but never re-resolves policy from the mutable provider registry. It derives a canonical key from every field of the resolved serving policy, sorting `retryableCodes` because eligibility uses set membership, and continues retry history only for the same provider and key. Replacing a route with different limits, code membership, or backoff therefore starts a new count and initial delay even when the mode is unchanged. Normal mode retains the bounded transient behavior: it retries configured codes up to `maxRetries` and otherwise delegates.
|
||||
|
||||
Always mode asks downstream recovery first so a specialized policy such as context-overflow compaction can make progress. A downstream retry wins. A downstream failure decision or thrown recovery error falls back to an unbounded retry of the same provider request; the thrown error is logged. The retry listener owns and drains delegated recovery before cancellation or plugin disposal can finish, then applies the abort instead of a late downstream decision. Success, turn cancellation, and plugin disposal are the only termination paths.
|
||||
|
||||
Both modes use exponential local delays from `initialDelayMs` to `maxDelayMs`. `jitterRatio` multiplies each target by a uniform sample in `[1 - jitterRatio, 1 + jitterRatio]`, then applies the cap. A positive provider `Retry-After` within the cap remains exact and unjittered. An over-cap provider delay makes normal mode delegate; always mode retains its guarantee by using the configured local backoff.
|
||||
|
||||
Each scheduled retry appends a non-surface `llm/retry` event with the failed provider, policy mode, canonical resolved-policy key, provider-policy retry number, delay, and failure facts. Normal events carry finite `maxRetries`; always events omit it, and UIs render the limit as `∞`. The event and failed `assistant/chunk` records do not contribute surface messages, so the next request contains the same derived context as the failed request unless another recovery policy deliberately changes the surface.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
**One global `always` switch** — rejected because it cannot isolate the unbounded cost and latency risk to the provider that needs it and can silently apply after runtime rerouting.
|
||||
|
||||
**A separate exact-provider list on `dsh-llm-retry`** — rejected because it duplicates provider route names outside their owning adapter configuration and lets provider registration drift from recovery policy.
|
||||
|
||||
**A very large finite retry count** — rejected because it eventually violates the requested keep-retrying contract and serializes an arbitrary operational limit as if it were meaningful.
|
||||
|
||||
**Provider-SDK retries** — rejected because hidden attempts multiply agent-level budgets, cannot use the closed-step durability boundary, and may splice or discard streamed output without a reconstructable retry record.
|
||||
|
||||
**Put the error into model context** — rejected because a transport or provider diagnostic is operational state, not conversation content. It can expose sensitive provider details and changes the retried request instead of repeating the failed request.
|
||||
|
||||
## Verification
|
||||
|
||||
Adapter tests validate nested policies at provider load, prove registration captures configured and default policies, and retain the serving policy across in-flight route replacement. Unit tests select policies from the failed request's serving registration, separate provider and changed-policy histories, exercise always mode beyond the normal budget, pin jitter and delay caps, prove downstream recovery ordering, prove cancellation and disposal drain delegated recovery before reaching quiescence, and prove both abort active backoff waits. Request-level coverage compares the complete messages of failed and retried attempts and rejects both provider error text and discarded partial output. A keyless headless `stream-json` snapshot runs failure, retry, and success through the assembled app, pins the complete `llm/retry` record, and rejects any model-message change between attempts. JSONL and SQLite tests round-trip an always event without `Infinity`; invariant tests bind provider identity to the request header, validate failure and mode-specific timer bounds, and bind retry numbers to provider-policy keys; TUI tests render finite and infinite limits.
|
||||
|
||||
## Consequences
|
||||
|
||||
Normal mode remains a finite default, while an explicit always policy can spend unbounded requests and time on permanent authentication, quota, invalid-request, protocol, or context failures. Operators must pair always mode with a cancellable caller and provider-specific cost controls. Retry state stays observable and durable without becoming model-visible, and serving-registration capture prevents adapter lifecycle changes from retroactively changing an in-flight request's recovery contract.
|
||||
|
||||
This decision extends the closed-step recovery, single visible adapter attempt, structured failure, and durable status design in [bounded recovery for transient LLM request failures](../architecture/2026-06-21-bounded-llm-request-recovery.md).
|
||||
@@ -0,0 +1,65 @@
|
||||
# Agent Note: 逐提供方请求重试策略
|
||||
|
||||
Status: implemented
|
||||
|
||||
[English](2026-07-24-provider-retry-policies.md) | 中文
|
||||
|
||||
## 问题
|
||||
|
||||
同一进程可能把模型请求路由到可靠性和成本约束各不相同的提供方。单一的瞬态错误分类器与有限重试预算无法表达这种部署需求：大多数提供方只需有界恢复，但其中一个提供方必须持续重试每次模型请求失败，直到请求成功或调用方取消。
|
||||
|
||||
提供方策略必须跟随实际失败的请求，包括 `agent/request` 选择的路由，而不能跟随 agent（智能体）的初始选项。无界策略也不能把 JavaScript `Infinity` 存入持久会话事件；提供方错误文本与丢弃的部分输出都不得进入下一次模型请求。
|
||||
|
||||
## 决策
|
||||
|
||||
每个具体适配器都在其提供方配置中接受可选的 `retryPolicy`。适配器负责校验并解析策略，`ctx.llm` 则在该精确提供方路由注册时捕获策略。当调用进入最终适配器边界时，`ctx.llm` 会把实际提供服务的注册项所持不可变策略绑定到该调用；即使路由在请求进行期间被 dispose 或替换，agent loop 仍会把该策略传给已关闭步骤恢复。`@deepseek-ai/dsh-llm-retry` 会把绑定到该调用的策略与失败步骤的持久提供方标识结合起来。未到达最终适配器的调用没有实际提供服务的策略，因而会委托后续处理。未配置 `retryPolicy` 的提供方使用 normal 默认值。
|
||||
|
||||
```yaml
|
||||
providers:
|
||||
- provider: deepseek
|
||||
retryPolicy:
|
||||
mode: normal
|
||||
maxRetries: 2
|
||||
retryableCodes: [EMPTY_RESPONSE, RATE_LIMIT, SERVER, TIMEOUT, TRANSPORT]
|
||||
backoff:
|
||||
initialDelayMs: 500
|
||||
maxDelayMs: 10000
|
||||
jitterRatio: 0.1
|
||||
- provider: internal
|
||||
retryPolicy:
|
||||
mode: always
|
||||
backoff:
|
||||
initialDelayMs: 1000
|
||||
maxDelayMs: 30000
|
||||
jitterRatio: 0.2
|
||||
```
|
||||
|
||||
监听器从失败步骤关闭时生效的持久 `request/header` 读取提供方，后续恢复产生的改动不参与选择，但绝不会从可变的提供方注册表重新解析策略。它会根据已解析实际服务策略的所有字段生成规范键；由于错误资格按集合成员判断，生成时会对 `retryableCodes` 排序。重试历史只会对同一提供方和同一规范键延续。因此，即使模式未变，只要路由替换后的次数上限、错误代码成员或退避不同，重试计数与初始延迟都会重新开始。normal 模式保留有界瞬态错误处理行为：它重试配置的错误代码，次数不超过 `maxRetries`；其他情况委托后续处理。
|
||||
|
||||
always 模式先请求下游恢复，使上下文溢出压缩（compaction）之类的专用策略有机会取得进展。下游若决定重试，则以该决定为准。下游若决定失败或恢复过程抛出错误，则回退为无界重试同一提供方请求；抛出的错误会写入日志。重试监听器会持有并排空已委托的恢复，轮次取消或插件 dispose（资源释放）只能在其结束后完成；随后监听器会应用取消，而不会采用迟到的下游决定。成功、轮次取消和插件 dispose 是仅有的终止路径。
|
||||
|
||||
两种模式的本地延迟都按指数增长，从 `initialDelayMs` 增至 `maxDelayMs`。`jitterRatio` 用 `[1 - jitterRatio, 1 + jitterRatio]` 区间内的均匀随机样本乘以每次目标值，再应用上限。提供方给出的正数 `Retry-After` 若未超过上限，则保持精确且不加抖动。若提供方延迟超过上限，normal 模式会委托后续处理；always 模式则改用配置的本地退避，以维持无限重试保证。
|
||||
|
||||
每次安排重试都会追加一条不进入表层的 `llm/retry` 事件，其中包含失败的提供方、策略模式、已解析策略的规范键、提供方策略内的重试编号、延迟和失败事实。normal 事件包含有限的 `maxRetries`；always 事件省略该字段，UI 将上限渲染为 `∞`。该事件与失败的 `assistant/chunk` 记录都不会生成表层消息，因此除非其他恢复策略有意改变表层，否则下一次请求包含的派生上下文与失败请求相同。
|
||||
|
||||
## 曾考虑的替代方案
|
||||
|
||||
**单一全局 `always` 开关**：不予采纳，因为它无法把无界成本与延迟风险限制在确有需要的提供方，还可能在运行时重新路由后悄然生效。
|
||||
|
||||
**在 `dsh-llm-retry` 上维护单独的精确提供方列表**：不予采纳，因为它会在所属适配器配置之外重复提供方路由名称，并让提供方注册与恢复策略发生偏差。
|
||||
|
||||
**设置很大的有限重试次数**：不予采纳，因为它最终仍会违反持续重试的契约，并把任意选取的运维上限序列化成看似有意义的数值。
|
||||
|
||||
**使用提供方 SDK 重试**：不予采纳，因为隐藏尝试会叠加 agent 层预算，无法利用已关闭 step 的持久性边界，还可能在没有可重建重试记录的情况下拼接或丢弃流式输出。
|
||||
|
||||
**把错误放入模型上下文**：不予采纳，因为传输或提供方诊断信息属于运维状态，而非对话内容。它可能暴露敏感的提供方细节，并会改变重试请求，无法重复原本失败的请求。
|
||||
|
||||
## 验证
|
||||
|
||||
适配器测试会在提供方加载时校验嵌套策略，证明注册流程会捕获已配置策略和默认策略，并证明请求进行期间替换路由后仍会保留实际提供服务的策略。单元测试根据失败请求实际使用的注册项选择策略、分离不同提供方和策略变更后的重试历史、验证 always 模式可越过 normal 预算、固定抖动和延迟上限、证明下游恢复顺序、证明取消与 dispose 会先排空已委托的恢复再达到完全停稳，并证明二者都会停止正在进行的退避等待。请求级覆盖会比较失败尝试与重试尝试的完整消息，并排除提供方错误文本和丢弃的部分输出。一个无密钥 headless `stream-json` 快照会通过组装后的应用执行失败、重试与成功流程，固定完整的 `llm/retry` 记录，并拒绝各次尝试之间出现任何模型消息变化。JSONL 与 SQLite 测试会往返读写不含 `Infinity` 的 always 事件；不变式测试会将提供方标识绑定到请求头、校验失败事实和各模式的计时器边界，并将重试编号绑定到提供方策略键；TUI 测试会渲染有限和无限上限。
|
||||
|
||||
## 后果
|
||||
|
||||
normal 模式仍是有限的默认策略；显式的 always 策略可能在永久性的身份验证、配额、无效请求、协议或上下文错误上耗费无限次请求和无限时间。运维方必须为 always 模式配备可取消的调用方和针对提供方的成本控制。重试状态保持可观察且持久，但不会对模型可见；捕获实际提供服务的注册项，也能防止适配器生命周期变化反过来改变进行中请求的恢复契约。
|
||||
|
||||
本决策扩展了[瞬态 LLM（大语言模型）请求失败的有界恢复](../architecture/2026-06-21-bounded-llm-request-recovery.md)中确定的已关闭 step 恢复、单次可见适配器尝试、结构化失败与持久状态设计。
|
||||
@@ -2,5 +2,5 @@
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-07-session-prefix.md: 75113952fc5f3df8da1580d42ed2a385b6135fe8
|
||||
2026-07-07-session-prefix.zh.md: e38bf09298296203b275d6d66a62ef17be7c045d
|
||||
2026-07-24-readable-xml-tool-output.md: 4f7327a7c6f5e2f04e36576da0fb739c34955e8a
|
||||
2026-07-24-readable-xml-tool-output.zh.md: 3c56d256b489863210b44449111f03a5752a889a
|
||||
@@ -0,0 +1,27 @@
|
||||
# Agent Note: Readable XML tool output
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-07-24-readable-xml-tool-output.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
Model-facing context and tool result text can expose transport-oriented XML wrappers instead of the information people need. Context producers do not declare presentation intent, and replayed tool calls whose definition is unavailable still need a conservative fallback that does not reinterpret ordinary prose or partial markup.
|
||||
|
||||
## Decision
|
||||
|
||||
The read tool declares a generic completed-result presentation that removes its `<path>`, `<type>`, and `<content>` wrapper while preserving the numbered content and footer. This tool-owned projection applies consistently to every UI that consumes tool presentation intent.
|
||||
|
||||
The TUI parses a context message or unavailable-tool result as XML only when the complete text is one supported XML document. It renders element names and attributes as an indented tree, preserves the context source label, applies the collapsed line budget independently to each tool result's top-level child lines and child count, and keeps raw text for malformed XML, mixed text, declarations, processing instructions, doctypes, and comments. A known tool's raw XML remains literal unless that tool declares its own result presenter. This XML fallback is TUI-only.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
**Strip XML-like tags with regular expressions.** Rejected because nested elements, attributes, entities, and malformed input require a real parser; partial conversion would make ambiguous output harder to inspect.
|
||||
|
||||
**Parse every generic result.** Rejected because known tools own their presentation contract, and silently reinterpreting their literal XML would override that decision.
|
||||
|
||||
**Show only raw XML.** Rejected because wrappers optimized for model consumption add terminal noise, particularly for filesystem reads and deeply nested structured results.
|
||||
|
||||
## Consequences
|
||||
|
||||
Filesystem reads are shorter in TUI cards without changing canonical model-facing content. Complete XML context messages, including workspace instruction reminders, become readable trees; unknown complete XML results become navigable trees and retain per-child context when collapsed. The TUI adds a strict SAX parser dependency and deliberately declines XML features (undefined entities, DOCTYPE, comments, processing instructions) that could hide or transform input beyond the conservative tree view. Predefined entities and character references do expand, so parsed text and attribute values are re-escaped for terminal output after parsing: a character reference can produce a control character that escaping the raw source never saw. Other UIs show raw generic content.
|
||||
@@ -0,0 +1,27 @@
|
||||
# Agent Note: 可读的 XML 工具输出
|
||||
|
||||
Status: implemented
|
||||
|
||||
[English](2026-07-24-readable-xml-tool-output.md) | 中文
|
||||
|
||||
## 问题
|
||||
|
||||
面向模型的上下文和工具结果文本可能呈现面向传输的 XML 包装，而不是人们真正需要的信息。上下文生产方不声明呈现意图，而对于回放时拿不到工具定义的调用，仍需要一个保守的回退方案，并且该方案不得重新解释普通文字或不完整的标记。
|
||||
|
||||
## 决策
|
||||
|
||||
read 工具声明一个通用的完成结果呈现：去除自身的 `<path>`、`<type>` 和 `<content>` 包装，同时保留带行号的内容和尾部信息。这一由工具自身持有的投影一致地作用于所有消费工具呈现意图的 UI。
|
||||
|
||||
只有当完整文本恰为一个受支持的 XML 文档时，TUI 才把上下文消息或工具定义不可用的工具结果按 XML 解析。TUI 将元素名和属性渲染为缩进树，保留上下文的来源标签；对于每个工具结果，分别按折叠行数预算限制各顶层子元素的行数和顶层子元素数量；对于格式错误的 XML、混合文本、XML 声明、处理指令、doctype 和注释，则保留原始文本。除非已知工具声明了自己的结果呈现器，否则其原始 XML 仍按字面显示。这一 XML 回退机制仅限 TUI。
|
||||
|
||||
## 曾考虑的替代方案
|
||||
|
||||
**用正则表达式剥除类 XML 标签。** 已否决：嵌套元素、属性、实体和格式错误的输入都需要真正的解析器；部分转换会让本就有歧义的输出更难检查。
|
||||
|
||||
**解析所有通用结果。** 已否决：已知工具拥有自己的呈现契约，静默重新解释它们的字面 XML 会推翻这一决定。
|
||||
|
||||
**只显示原始 XML。** 已否决：为模型消费而优化的包装会给终端增加噪音，对文件系统读取和嵌套很深的结构化结果尤其如此。
|
||||
|
||||
## 后果
|
||||
|
||||
文件系统读取在 TUI 卡片中变得更短，而规范的面向模型内容保持不变。完整的 XML 上下文消息（包括工作区指令提醒）变成可读的树；未知的完整 XML 结果变成可导航的树，折叠时也保留每个子元素的上下文。TUI 新增一个严格 SAX 解析器依赖，并有意不支持那些可能在保守树视图之外隐藏或变换输入的 XML 特性（未定义实体、DOCTYPE、注释、处理指令）。预定义实体和字符引用会被展开，因此解析出的文本和属性值在解析后会为终端输出重新转义：字符引用可能产生对原始源文本转义时从未见过的控制字符。其他 UI 展示原始的通用内容。
|
||||
@@ -0,0 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-07-24-tui-banner-model-deduplication.md
|
||||
2026-07-24-tui-banner-model-deduplication.md: afd370a8762d8e6c17c61d50c95d68998a063df5
|
||||
2026-07-24-tui-banner-model-deduplication.zh.md: 86a3cdb0e714642253162f1fe062e19bdc40bbe4
|
||||
@@ -0,0 +1,33 @@
|
||||
# Agent Note: The startup banner omits the model
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-07-24-tui-banner-model-deduplication.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
The startup banner repeated the selected model directly above the prompt context, which already keeps the model visible while the TUI is idle. The duplicate added no information and made the banner detail line harder to scan.
|
||||
|
||||
## Decision
|
||||
|
||||
- The borderless startup banner shows the product title, optional `welcome` or session-title subtitle, and session id.
|
||||
- The banner omits the model name. The prompt context remains the persistent model display and updates after `/model` selection.
|
||||
- The sweep animation and configured-welcome behavior are unchanged.
|
||||
|
||||
This supersedes only the model-in-banner portion of the [borderless banner decision](../../archived/feature/2026-07-21-tui-borderless-banner.md).
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
**Remove the entire detail line.** Rejected: the session id remains useful for identifying and resuming the active session, and it is not duplicated in the prompt context.
|
||||
|
||||
**Remove the model from the prompt context instead.** Rejected: the prompt context stays visible after the startup banner scrolls away and reflects later model selections.
|
||||
|
||||
## Consequences
|
||||
|
||||
- Startup uses the banner detail row only for the session id.
|
||||
- The model appears once in the initial idle view, in the prompt context.
|
||||
- Banner snapshots and runnable TUI replay snapshots contain a shorter detail row.
|
||||
|
||||
## Testing
|
||||
|
||||
`packages/ui/tui/tests/tui.spec.ts` asserts that completed banners retain the session id without the former `<model> • <session-id>` text. Package-local and runnable-example TUI snapshots pin the resulting rows.
|
||||
@@ -0,0 +1,33 @@
|
||||
# Agent Note：启动横幅不再显示模型
|
||||
|
||||
Status: implemented
|
||||
|
||||
[English](2026-07-24-tui-banner-model-deduplication.md) | 中文
|
||||
|
||||
## 问题
|
||||
|
||||
启动横幅在提示区上下文（prompt context）的正上方重复显示所选模型，而提示区上下文本身已在 TUI 空闲时持续展示模型。这一重复不提供任何信息，还让横幅详情行更难扫读。
|
||||
|
||||
## 决策
|
||||
|
||||
- 无边框启动横幅显示产品标题、可选的 `welcome` 或会话标题副标题，以及会话 id。
|
||||
- 横幅不再显示模型名。提示区上下文仍是常驻的模型展示位，并在 `/model` 选择后随之更新。
|
||||
- 扫入动画和配置了欢迎语时的行为保持不变。
|
||||
|
||||
本 note 仅取代[无边框横幅决策](../../archived/feature/2026-07-21-tui-borderless-banner.md)中模型进横幅的那部分。
|
||||
|
||||
## 考虑过的替代方案
|
||||
|
||||
**移除整条详情行。** 否决：会话 id 对识别和恢复当前会话仍然有用，而且它在提示区上下文中没有重复。
|
||||
|
||||
**改为把模型从提示区上下文移除。** 否决：提示区上下文在启动横幅滚出视野后仍保持可见，并会反映之后的模型选择。
|
||||
|
||||
## 后果
|
||||
|
||||
- 启动时横幅详情行只承载会话 id。
|
||||
- 在初始空闲视图中模型只出现一次，位于提示区上下文。
|
||||
- 横幅快照和可运行的 TUI 回放快照包含更短的详情行。
|
||||
|
||||
## 测试
|
||||
|
||||
`packages/ui/tui/tests/tui.spec.ts` 断言完成后的横幅保留会话 id，且不含先前的 `<model> • <session-id>` 文本。包内快照与可运行示例的 TUI 快照固定了最终的各行内容。
|
||||
@@ -0,0 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-24-tui-message-header-timing.md: 94a4d04c75f9b0ad76e2460738a07ba82ac3bb9f
|
||||
2026-07-24-tui-message-header-timing.zh.md: 4713555290bbc47bb3af56cd3b4d0c493c81e6f1
|
||||
@@ -0,0 +1,25 @@
|
||||
# Agent Note: TUI message header timing
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-07-24-tui-message-header-timing.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
Turn timing beside the editor disappears from the transcript when the user scrolls and cannot appear until the editor status renders. A whole-turn aggregate also obscures the latency of later model requests after tool calls.
|
||||
|
||||
## Decision
|
||||
|
||||
Every model step creates an assistant header at `step/start`, before the first streamed chunk. The header displays `Model wait` immediately and refreshes at 100 ms resolution, then adds exclusive `Thinking`, `Response`, and `Tools` buckets as session events move the step between phases.
|
||||
|
||||
`step/end` freezes the header and adds the local completion timestamp. Transcript replay derives the same timing from durable event timestamps. Empty and tool-only steps retain a header, while failed live output and its header retract together when retry handling rebuilds the active session surface.
|
||||
|
||||
The prompt context retains only queued-steering state. Timing belongs to the model step that produced it rather than to the editor or the whole turn.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
Keeping timing beside the editor preserves a stable layout but hides per-step latency in scrollback and resume. Adding a second status line duplicates the same metric in two places. Labeling the first bucket `TTFT` is compact but requires protocol terminology; `Model wait` states the user-visible meaning without claiming that the first chunk is always text.
|
||||
|
||||
## Consequences
|
||||
|
||||
Users receive visible feedback before model output and can compare each request after tools or retries. Updating at 100 ms resolution causes more terminal renders while a model step is active. Internal timing state keeps the established `ttft` name because it identifies the measured bucket precisely; only rendered text uses `Model wait`.
|
||||
@@ -0,0 +1,25 @@
|
||||
# Agent Note：TUI 消息头部计时
|
||||
|
||||
Status: implemented
|
||||
|
||||
[English](2026-07-24-tui-message-header-timing.md) | 中文
|
||||
|
||||
## 问题
|
||||
|
||||
编辑器旁的轮次计时会在用户滚动时从 transcript（文本记录）中消失，且要等到编辑器状态渲染后才能出现。整轮聚合值还会掩盖工具调用之后各后续模型请求的延迟。
|
||||
|
||||
## 决策
|
||||
|
||||
每个模型步骤都在 `step/start` 时（即第一个流式分片到达之前）创建一个 assistant 头部。头部立即显示 `Model wait` 并以 100 ms 分辨率刷新；随着会话事件使该步骤在不同阶段之间切换，头部再加入互斥的 `Thinking`、`Response` 和 `Tools` 时间桶。
|
||||
|
||||
`step/end` 冻结头部并附上本地完成时间戳。transcript 回放从持久事件时间戳派生出相同的计时。空步骤和纯工具步骤同样保留头部；当重试处理重建活跃会话表层时，失败的实时输出与其头部一并撤除。
|
||||
|
||||
提示区上下文（prompt context）只保留排队中的 steering（中途引导）状态。计时归属于产生它的模型步骤，而不是编辑器或整个轮次。
|
||||
|
||||
## 考虑过的替代方案
|
||||
|
||||
把计时留在编辑器旁能保持布局稳定，但在 scrollback 和会话恢复中看不到各步骤的延迟。增加第二条状态行会让同一指标出现在两处。把第一个时间桶标为 `TTFT` 更紧凑，但依赖协议术语；`Model wait` 直接陈述用户可见的含义，而不宣称第一个分片总是文本。
|
||||
|
||||
## 后果
|
||||
|
||||
用户在模型输出之前就能得到可见反馈，并能比较工具或重试之后的每次请求。以 100 ms 分辨率刷新会在模型步骤活跃期间带来更多终端渲染。内部计时状态沿用既有的 `ttft` 名称，因为它精确标识所计量的时间桶；只有渲染文本使用 `Model wait`。
|
||||
@@ -0,0 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-24-tui-prompt-status-indicator.md: 8d469c3b0627325f373ca8f8d4d23d67bb09e342
|
||||
2026-07-24-tui-prompt-status-indicator.zh.md: 0dee8d1e4e7ce5a6f0299f9b7cd1595f63bd6e08
|
||||
@@ -0,0 +1,33 @@
|
||||
# Agent Note: TUI prompt status indicator
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-07-24-tui-prompt-status-indicator.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
While a turn runs, the input prompt shows only its static `dsh>` prefix. The assistant header carries the elapsed timing, but the editor row — where the user's attention rests — gives no live signal of what the agent is doing right now: waiting for the first token, thinking, responding, or running tools.
|
||||
|
||||
## Decision
|
||||
|
||||
While the agent is running, a phase-specific glyph replaces the `>` caret of the built-in `${indicator}` prompt value. The `inputPrompt` theme template defaults to `${symbol} ${indicator}`, where the built-in `${symbol}` value holds the `dsh` label and `${indicator}` holds the caret slot with its trailing gap before the cursor; the template literal space separates them, rendering `dsh <glyph> ` in every state. The phase is the open step's active timing bucket, derived from the same session events and rules that drive the [message header timing](2026-07-24-tui-message-header-timing.md) — no new phase model. One glyph per bucket: `◍` model wait (pre-first-token), `✻` thinking, `●` responding, `⚙` tools. A running turn with no open step falls back to the model-wait glyph; an idle agent restores the plain `>`.
|
||||
|
||||
The glyph occupies the caret's exact column with the same display width every frame, so the cursor never shifts as the phase changes or the glyph animates. Activity is conveyed by a brightness pulse, not by appearing and disappearing: a four-frame triangle wave (dim → normal → bold → normal) wraps the accent-colored glyph in the true SGR intensity codes (2 and 1) — never the palette's semantic `dim` role, which on a light scheme is a color the glyph's own accent would override — so the pulse survives every terminal scheme. The render-clock cadence is 250 ms per frame, a fixed presentation rhythm alongside the sibling 100 ms status refresh, not a deployment choice. The running-status timer refreshes every 100 ms tick unconditionally rather than only when a streaming component exists, so the pulse animates even during the pre-first-token wait.
|
||||
|
||||
The caret and its animation are their own `${indicator}` value, separate from the `${symbol}` label, so the `inputPrompt` template composes the two: `${symbol} ${indicator}` reads as `dsh <caret>`. Configurability lives at that template — a deployment reorders or drops either value, and omitting `${indicator}` opts out of the running indicator. The glyph set, the pulse, and the `dsh` label are fixed in code — not per-deployment fields — matching the fixed timing-bucket labels they mirror.
|
||||
|
||||
The built-in `${symbol}`/`${indicator}` updates ride the renders the TUI already drives on every state change that can move a value (`agent/status`, session events, the 100 ms running-status timer, async model-context resolution). A prompt value that changes on its own schedule — a plugin-owned `${custom}` fragment — instead redraws through the registry's coalesced change notification, which the renderer subscribes to directly rather than through a Cordis event ([registry](2026-07-24-configurable-tui-prompt-theme.md)).
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
**Prepend the glyph before `dsh>` as its own `${status}` token.** Rejected: a leading token shifts the whole prompt — and the cursor — right by two columns whenever it appears, and collapses back when it clears. Replacing the caret keeps the cursor column fixed.
|
||||
|
||||
**A blinking glyph that appears and disappears.** Rejected: on/off blanking still moves nothing horizontally once the glyph owns the caret column, but the empty frames read as flicker. A brightness pulse animates continuously while the character stays put.
|
||||
|
||||
**A per-phase spinner animation** (rotating frames). Rejected: the four phases are already distinguished by their glyph shapes; swapping the character per frame would conflate "which phase" with "still working". The pulse animates intensity while the shape stays a stable phase signal, reusing the existing 100 ms status timer.
|
||||
|
||||
**A new phase state machine in the TUI.** Rejected: the header-timing machinery already replays the open step's active bucket from session events. Deriving the glyph from that bucket keeps one source of truth for "what phase is this step in".
|
||||
|
||||
## Consequences
|
||||
|
||||
The user gets a live, glanceable phase signal in the caret they are already watching, with no horizontal movement of the cursor or the prompt. The pulse costs terminal renders on every 100 ms tick for the whole running turn, not only while a streaming component is mounted. The glyph mapping and the pulse are fixed in code, not configurable, matching the fixed timing-bucket labels they mirror.
|
||||
@@ -0,0 +1,33 @@
|
||||
# Agent Note：TUI 提示区状态指示器
|
||||
|
||||
Status: implemented
|
||||
|
||||
[English](2026-07-24-tui-prompt-status-indicator.md) | 中文
|
||||
|
||||
## 问题
|
||||
|
||||
轮次运行期间，输入提示区只显示其静态的 `dsh>` 前缀。assistant 头部承载已用计时，但编辑器所在的这一行——也就是用户注意力所在之处——对 agent 此刻正在做什么没有任何实时信号：是在等待第一个 token、思考、响应，还是在运行工具。
|
||||
|
||||
## 决策
|
||||
|
||||
agent 运行期间，一个按阶段区分的字形会替换内置 `${indicator}` 提示区值中的 `>` 光标符。`inputPrompt` 主题模板默认为 `${symbol} ${indicator}`，其中内置 `${symbol}` 值承载 `dsh` 标签，`${indicator}` 承载光标符槽位及其在光标前的尾随间隙；模板中的字面空格将两者隔开，在每种状态下渲染为 `dsh <字形> `。阶段取自当前打开步骤的活跃计时桶，其派生所依据的会话事件与规则和[消息头部计时](2026-07-24-tui-message-header-timing.md)相同——没有引入新的阶段模型。每个桶对应一个字形：`◍` 等待模型（第一个 token 之前）、`✻` 思考、`●` 响应、`⚙` 工具。运行中但没有打开步骤的轮次回退到等待模型的字形；agent 空闲时则恢复为纯 `>`。
|
||||
|
||||
字形占据光标符所在的同一列，且每一帧的显示宽度都相同，因此无论阶段切换还是字形动画，光标都不会移动。活动状态由亮度脉动传达，而不是靠出现和消失：一个四帧三角波（暗 → 正常 → 亮 → 正常）用真正的 SGR 强度码（2 与 1）包裹带 accent 色的字形——绝不使用调色板语义上的 `dim` 角色，因为在浅色 scheme 下它是一种颜色，会被字形自身的 accent 色覆盖——因此脉动在任何终端 scheme 下都能保留。渲染时钟节拍为每帧 250 ms，是与配套的 100 ms 状态刷新并列的固定呈现节奏，而非部署选项。运行状态计时器每 100 ms 无条件刷新一次，而不再只在存在流式组件时刷新，因此即使在第一个 token 之前的等待期间，脉动也能持续。
|
||||
|
||||
光标符及其动画自成一个 `${indicator}` 值，与 `${symbol}` 标签分离，因此 `inputPrompt` 模板将二者组合：`${symbol} ${indicator}` 读作 `dsh <光标符>`。可配置性位于该模板——部署可重排或丢弃任一值，省略 `${indicator}` 即退出运行指示器。字形集、脉动以及 `dsh` 标签都固定在代码中——不是逐部署字段——与它们映射的固定计时桶标签一致。
|
||||
|
||||
内置 `${symbol}`/`${indicator}` 的更新搭乘 TUI 本就在每次可能改变某个值的状态变化（`agent/status`、会话事件、100 ms 运行状态计时器、异步模型上下文解析）时驱动的渲染。而一个自行变化的值——插件拥有的 `${custom}` 片段——则通过注册表的合并变更通知重绘，而渲染器直接订阅它，而非通过 Cordis 事件（参见[注册表](2026-07-24-configurable-tui-prompt-theme.md)）。
|
||||
|
||||
## 考虑过的替代方案
|
||||
|
||||
**把字形作为自己的 `${status}` token 前置在 `dsh>` 之前。** 已否决：前置 token 每次出现都会把整个提示区——连同光标——向右移动两列，清除时又缩回。在尾随的 `${indicator}` 槽位替换光标符能让光标列保持固定。
|
||||
|
||||
**出现又消失的闪烁字形。** 已否决：一旦字形占据光标符所在列，开/关式的空白帧在水平方向上不再移动任何东西，但空帧读起来像闪烁。亮度脉动让字符保持不动的同时持续做动画。
|
||||
|
||||
**按阶段的 spinner 动画**（旋转帧）。已否决：四个阶段已经通过各自的字形形状区分；逐帧切换字符会把「哪个阶段」与「仍在工作」混为一谈。脉动只改变强度做动画，而形状始终是稳定的阶段信号，且复用了既有的 100 ms 状态计时器。
|
||||
|
||||
**在 TUI 中新建阶段状态机。** 已否决：头部计时机制已从会话事件回放出当前打开步骤的活跃桶。从该桶派生字形，能让「这个步骤处于哪个阶段」保持单一事实来源。
|
||||
|
||||
## 后果
|
||||
|
||||
用户在自己本就注视的光标符处获得可一眼掌握的实时阶段信号，且光标与提示区都没有水平移动。脉动的代价是整个运行轮次内每 100 ms 一次的终端渲染，而不再只在流式组件挂载期间。字形映射与脉动都固定在代码中、不可配置，与其所对应的固定计时桶标签一致。
|
||||
@@ -0,0 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-24-tui-prompt-workspace-label.md: c45c60c6554766cca01076b229956a7bc7d98d48
|
||||
2026-07-24-tui-prompt-workspace-label.zh.md: 170dba83becf4b529679e7db9c7c84a6de7dec13
|
||||
@@ -0,0 +1,34 @@
|
||||
# Agent Note: The prompt context combines directory and branch
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-07-24-tui-prompt-workspace-label.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
The idle prompt context rendered the working directory and `git:<branch>` as separate segments. In task worktrees, the directory can already identify the checkout, while the prefixed branch segment consumed additional horizontal space and was discarded independently on narrower terminals.
|
||||
|
||||
## Decision
|
||||
|
||||
- The prompt context renders the working directory and available Git branch as one workspace label: `<directory> (<branch>)`.
|
||||
- The directory remains bold and accented; the parenthesized branch remains muted.
|
||||
- The combined workspace label has the highest retention priority and is clipped as one segment when it exceeds the terminal width.
|
||||
- Outside a Git worktree or on detached HEAD, the label remains the directory alone.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
**Keep `git:<branch>` as a separate segment.** Rejected: the prefix and separator use more columns without adding meaning in this context.
|
||||
|
||||
**Show only the branch.** Rejected: the session working directory determines where tools operate and remains the primary prompt context.
|
||||
|
||||
**Derive a special worktree root label.** Rejected: the existing formatted directory and Git branch already provide the two relevant facts without adding repository-layout assumptions.
|
||||
|
||||
## Consequences
|
||||
|
||||
- A typical checkout renders as `~/git/tui-staging (tui-staging)`.
|
||||
- Narrow terminals retain or clip directory and branch together instead of dropping the branch independently.
|
||||
- Embedding-provided `TuiRuntime.formatCwd` labels compose with the branch in the same form.
|
||||
|
||||
## Testing
|
||||
|
||||
`packages/ui/tui/tests/tui.spec.ts` pins home, absolute, formatted, and narrow workspace labels. Package-local and runnable-example TUI snapshots verify the assembled prompt context.
|
||||
@@ -0,0 +1,34 @@
|
||||
# Agent Note：提示区上下文合并显示目录与分支
|
||||
|
||||
Status: implemented
|
||||
|
||||
[English](2026-07-24-tui-prompt-workspace-label.md) | 中文
|
||||
|
||||
## 问题
|
||||
|
||||
空闲提示区上下文（prompt context）把工作目录和 `git:<branch>` 作为两个独立片段渲染。在任务 worktree 中，目录本身往往已能标识当前检出，而带前缀的分支片段额外占用横向空间，且在较窄的终端上会被单独丢弃。
|
||||
|
||||
## 决策
|
||||
|
||||
- 提示区上下文把工作目录和可用的 Git 分支渲染为一个工作区标签（workspace label）：`<directory> (<branch>)`。
|
||||
- 目录仍为加粗强调色；括号内的分支仍为弱化色。
|
||||
- 合并后的工作区标签具有最高保留优先级，超出终端宽度时作为一个整体片段裁剪。
|
||||
- 不在 Git worktree 中或处于 detached HEAD 时，标签仍只显示目录。
|
||||
|
||||
## 考虑过的替代方案
|
||||
|
||||
**保留 `git:<branch>` 作为独立片段。** 否决：前缀和分隔符占用更多列宽，在此上下文中却不增加信息。
|
||||
|
||||
**只显示分支。** 否决：会话工作目录决定工具在哪里运行，仍是提示区上下文的首要信息。
|
||||
|
||||
**派生一个特殊的 worktree 根目录标签。** 否决：现有的格式化目录和 Git 分支已经提供了这两项相关信息，无需引入对仓库布局的假设。
|
||||
|
||||
## 后果
|
||||
|
||||
- 典型的检出渲染为 `~/git/tui-staging (tui-staging)`。
|
||||
- 窄终端把目录和分支作为整体保留或裁剪，而不是单独丢弃分支。
|
||||
- 嵌入方通过 `TuiRuntime.formatCwd` 提供的标签以同样的形式与分支组合。
|
||||
|
||||
## 测试
|
||||
|
||||
`packages/ui/tui/tests/tui.spec.ts` 固定了主目录、绝对路径、格式化及窄终端下的工作区标签。包内快照与可运行示例的 TUI 快照验证了组装后的提示区上下文。
|
||||
@@ -2,5 +2,5 @@
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-10-session-query-service.md: 42d12fe2c5e34e71a6166816857b9ced52a61a95
|
||||
2026-07-10-session-query-service.zh.md: 2c8d322ca6099db1c8ddbb8a02efbc8729e83dbf
|
||||
2026-07-24-tui-shell-prompt-editor.md: bba03e788b92692f534fd97e66035757e9c74356
|
||||
2026-07-24-tui-shell-prompt-editor.zh.md: 1897d11292ec3b189245169956ac327f0b81b0e2
|
||||
@@ -0,0 +1,33 @@
|
||||
# Agent Note: TUI shell-prompt editor
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-07-24-tui-shell-prompt-editor.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
The upstream pi-tui editor always renders horizontal frame rows. That presentation separates input from the transcript but occupies two terminal rows and does not resemble the command-oriented input used by shells.
|
||||
|
||||
## Decision
|
||||
|
||||
The TUI presents a two-line prompt. A DSH-owned context line shows the working directory, running-turn timing, optional Git branch, current model, token totals, cache hit rate, and context pressure as independently prioritized segments. Narrow terminals omit lower-priority segments while retaining the directory, followed by running timing when it is present. The second line uses a fixed-width `dsh> ` prefix and equal-width continuation indent; its running steer/cancel guidance is placeholder text that disappears when input begins.
|
||||
|
||||
The pinned `@earendil-works/pi-tui` package carries a pnpm patch that adds `frame: "none"` and fixed-width prompt prefixes to `EditorOptions`. The default remains the upstream horizontal frame, so only the DSH editor opts into the behavior. Prefixes must have equal visible widths; construction fails when they differ. Input, explicit newlines, autocomplete, cursor placement, and scroll indicators share the reduced first-row width; automatically wrapped rows render no prefix, so their text starts at the editor's left padding, occupies the prefix columns, and wraps at the full content width.
|
||||
|
||||
The patch stays limited to the published editor JavaScript and declarations. Keeping the exact dependency pin makes installation either apply the known patch or fail rather than silently dropping the presentation.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
**Filter the rendered editor output in a wrapper.** This would depend on recognizing ANSI-styled border and scroll-indicator rows and distinguishing autocomplete output from input output, all of which are undocumented render details.
|
||||
|
||||
**Vendor the complete pi-tui package.** The project updates frequently, while this change needs only a localized editor rendering option. Owning the full source and synchronization process would add disproportionate maintenance.
|
||||
|
||||
**Keep the horizontal frame.** This avoids dependency customization but retains the presentation the change is intended to replace.
|
||||
|
||||
## Consequences
|
||||
|
||||
The editor and context use two rows instead of the framed editor plus footer, with one blank row separating the prompt area from conversation cards. The persistent presentation omits session identity and tool-card mode; `/status` and commands retain those details. Input layout and autocomplete lose six columns to the prompt prefix, but wrapped text uses the otherwise blank prefix columns. Borderless scrolling uses standalone `↑ N more` and `↓ N more` rows.
|
||||
|
||||
The internal segment representation establishes width priorities without exposing a public customization language. Future Starship-like configuration can build on it after the default modules and overflow behavior have production evidence.
|
||||
|
||||
A pi-tui upgrade requires reviewing and reapplying or retiring the patch. TUI terminal snapshots pin the assembled presentation, including context modules, prompt color, alignment, cursor placement, and autocomplete width.
|
||||
@@ -0,0 +1,33 @@
|
||||
# Agent Note: TUI shell 提示符编辑器
|
||||
|
||||
Status: implemented
|
||||
|
||||
[English](2026-07-24-tui-shell-prompt-editor.md) | 中文
|
||||
|
||||
## 问题
|
||||
|
||||
上游 pi-tui 编辑器始终渲染横向边框行。这种呈现方式虽然把输入区与 transcript（文本记录）分隔开，却占用两行终端高度，也不像 shell 中面向命令的输入形态。
|
||||
|
||||
## 决策
|
||||
|
||||
TUI 呈现两行提示符。DSH 自有的上下文行把工作目录、运行中轮次的计时、可选的 Git 分支、当前模型、token 总量、缓存命中率与上下文压力显示为各自独立分配优先级的段（segment）。窄终端会省略低优先级的段，但保留目录；运行中计时存在时，其保留优先级仅次于目录。第二行使用固定宽度的 `dsh> ` 前缀与等宽的续行缩进；agent 运行期间提示 steering（中途引导）与取消的引导文字是占位文本，开始输入后即消失。
|
||||
|
||||
固定版本的 `@earendil-works/pi-tui` 包（package）携带一个 pnpm 补丁，为 `EditorOptions` 增加 `frame: "none"` 与固定宽度的提示符前缀。默认值仍是上游的横向边框，因此只有 DSH 编辑器选择启用该行为。两个前缀的可见宽度必须相等；宽度不同时构造会失败。输入、显式换行、自动补全、光标定位和滚动指示共用缩减后的首行宽度；自动折行产生的行不渲染前缀，其文本从编辑器左侧留白处开始，占用前缀列，并按完整内容宽度折行。
|
||||
|
||||
补丁范围仅限已发布的编辑器 JavaScript 与类型声明。依赖保持精确的版本固定，使安装要么应用已知补丁，要么直接失败，而不会静默丢掉这种呈现方式。
|
||||
|
||||
## 曾考虑的替代方案
|
||||
|
||||
**在包装层过滤编辑器的渲染输出。** 这需要识别带 ANSI 样式的边框行与滚动指示行，并区分自动补全输出与输入输出，而这些都是未见于文档的渲染细节。
|
||||
|
||||
**vendor 完整的 pi-tui 包。** 该项目更新频繁，而本次改动只需要一个局部的编辑器渲染选项。接手全部源码及其同步流程会带来不成比例的维护成本。
|
||||
|
||||
**保留横向边框。** 这可以避免定制依赖，但保留的正是本次改动想要替换的呈现方式。
|
||||
|
||||
## 后果
|
||||
|
||||
编辑器与上下文共占两行，取代原先带边框的编辑器加页脚，提示符区域与对话卡片之间以一行空行分隔。常驻呈现不含会话标识与工具卡片模式；`/status` 与各命令仍保留这些细节。输入布局与自动补全因提示符前缀占位而损失六列宽度，但折行后的文本会占用原本留空的前缀列。无边框滚动使用独立的 `↑ N more` 与 `↓ N more` 行。
|
||||
|
||||
段的内部表示确立了宽度优先级，而未暴露公开的定制语言。待默认模块与溢出行为积累生产环境证据后，未来可在其上构建类似 Starship 的配置。
|
||||
|
||||
升级 pi-tui 时需要评审该补丁，并重新应用或将其退役。TUI 终端快照固定组装后的呈现效果，包括上下文模块、提示符颜色、对齐、光标定位和自动补全宽度。
|
||||
@@ -0,0 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-07-24-web-session-model-selector.md
|
||||
2026-07-24-web-session-model-selector.md: 4bd52bba8d3ba16c01fc59d2b561516f6fd0bb87
|
||||
2026-07-24-web-session-model-selector.zh.md: 35f3525b80e7d27fb573e854aefc317ed354df53
|
||||
@@ -0,0 +1,43 @@
|
||||
# Agent Note: Session model selection in the Web composer
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-07-24-web-session-model-selector.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
The Web conversation displayed and sent through the Host's fixed provider/model route without exposing that route or letting a user change it. The TUI already had a session-local route target, but copying its presentation or hardcoding DeepSeek models in the browser would split model discovery and step-boundary semantics across front doors. A switch made while a response is running also needs one atomic boundary: prompt variables and request routing cannot observe different targets.
|
||||
|
||||
## Decision
|
||||
|
||||
The Web Host reuses `installAgentLlmTarget` for every created or resumed agent. The provider/model/reasoning target starts from the latest `request/header` when the session has used a model, otherwise from the Host default route. `session.selectModel` changes the session-local mutable target, and prompt assembly captures it with request routing; a switch during a running step therefore applies to the next assembled step. The next consumed target persists through the existing full `request/header` snapshot, while a choice that has not reached a request remains process-local.
|
||||
|
||||
The session RPC domain exposes a `session.models` directory and `session.selectModel`. The directory is built dynamically from the LLM registry and grouped by provider; each listed model's exact metadata adds adapter-owned reasoning effort ids, names, descriptions, and optional default. Provider catalogs and exact metadata load concurrently by provider and fail independently, so successful groups remain usable alongside retryable failure records. Catalog membership stays advisory: the current model is inserted as an unlisted row when its registered provider omits it, while exact resolution decides whether a route and explicit effort are available. Selection uses `resolveCallConfig` to reject unsupported effort ids and materialize an adapter-configured default before updating the target.
|
||||
|
||||
The browser `ModelService` owns one `ModelDirectory` per live session. Its snapshot contains the current complete target, grouped catalog, provider failures, operation error, and `idle`/`loading`/`ready`/`selecting`/`error` state. Mounting primes the trigger label and each menu open refreshes the directory. Directory and selection calls share an operation generation so older responses cannot replace a newer result; connection reset discards the process-local projection before restoring the Host target. Failures retain the previous current target and usable groups.
|
||||
|
||||
`@deepseek-ai/dsh-client-ui-conversation` declares the session-scoped single slot `conversation.input.model` as a child of its composer-bar entry. InputBar renders the seat in its trailing controls immediately before the pending indicator and primary button; the seat receives the bar's `locked` owner prop and session scope. `@deepseek-ai/dsh-client-ui-model` occupies that seat and also contributes `/model` over the same directory. Its compact trigger displays the catalog model name and effective reasoning label, falling back to ids when metadata is absent. The upward menu first offers Model and, when the current exact model supports it, Effort; Model drills into provider groups, while Effort drills into the adapter-ordered levels. The provider-default row appears only when the adapter does not configure a model default.
|
||||
|
||||
The production browser roster is the flat config tree in `apps/cli/cordis.yml`; the model feature is one `dshClient` row rather than a package hardcoded in Web boot code. Its package manifest orders it after the runtime and command feature, while Cordis service injection waits for the conversation slot before registering the composer occupant.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
**Use separate provider and model dropdowns.** The model list depends on the provider and repeats a two-stage interaction for every change. One grouped menu keeps the provider visible as organization without lengthening the trigger or each row.
|
||||
|
||||
**Hardcode the current DeepSeek catalog in the Web client.** This would drift from registered adapters and exclude deployment-owned providers. The LLM registry remains the source of provider and model metadata, including partial lookup failures.
|
||||
|
||||
**Keep `High`/`Max` as client-local UI state.** Static DeepSeek labels cannot represent `off`, pi-ai provider vocabularies, adapter defaults, validation, resume, or the next provider request. Exact-model metadata owns the selectable vocabulary, and the session target owns the selected id.
|
||||
|
||||
**Make the selection a global default.** A global mutation would unexpectedly redirect other open conversations. The target belongs to one live session, while Host configuration remains the default for sessions without a logged request.
|
||||
|
||||
**Reject changes while an agent is running.** The shared atomic target already separates the assembled step from the next selection. Keeping the selector available lets the user prepare the following step without altering the in-flight request.
|
||||
|
||||
**Persist every click as a new session event.** A choice is not model-visible until prompt assembly consumes it. Persisting unused UI intent would add a durable event that does not reconstruct a model request; the existing `request/header` records the first request that actually uses the route.
|
||||
|
||||
## Consequences
|
||||
|
||||
Any Host-backed Web conversation, including a blank session, can switch among dynamically discovered provider groups and adapter-owned reasoning levels without displaying duplicated `provider/model` labels. The current consumed target survives resume and reconnect; catalog names remain presentation-only, while selection and persistence use provider/model/effort ids. A provider catalog or exact-metadata outage degrades only that group. Route changes can reduce provider-side cache reuse, but the selector adds no prompt content and does not disturb the in-flight step. A model without reasoning metadata has no Effort row.
|
||||
|
||||
## Testing
|
||||
|
||||
Host tests pin grouped discovery, catalog and exact-metadata failure isolation, logged effort restoration, unlisted current targets, unsupported effort rejection, default materialization, and next-assembly switching. Client tests pin the shared directory, reconnect restoration, and complete-target submission. Component tests pin dynamic effort labels, descriptions, provider-default exposure, and effort submission. The keyless built-app fixture loads the production model plugin, selects OpenAI's GPT-5 and its Max effort, sends a turn, and verifies that the next generated response reports both ids.
|
||||
@@ -0,0 +1,43 @@
|
||||
# Agent Note: Web 对话输入区的会话模型选择
|
||||
|
||||
Status: implemented
|
||||
|
||||
[English](2026-07-24-web-session-model-selector.md) | 中文
|
||||
|
||||
## 问题
|
||||
|
||||
Web 对话原本通过 Host 固定的提供方与模型路由显示并发送消息，既不呈现该路由，也不允许用户更改。TUI 已经具备会话级的路由目标，但如果照搬其呈现方式，或在浏览器中硬编码 DeepSeek 模型，就会让模型发现逻辑和步骤边界语义分散到不同前门中。响应运行期间发生的切换还需要一个原子边界：提示词变量与请求路由不能观测到不同的目标。
|
||||
|
||||
## 决策
|
||||
|
||||
Web Host 为每个新建或恢复的 agent（智能体）复用 `installAgentLlmTarget`。如果会话已经使用过模型，提供方／模型／推理（reasoning）目标从最新的 `request/header` 开始；否则采用 Host 默认路由。`session.selectModel` 会更改会话级可变目标，提示词组装则将该目标与请求路由一并捕获，因此运行中步骤发生的切换会应用于下一个组装步骤。下一条实际采用的目标通过现有的完整 `request/header` 快照持久化；尚未进入请求的选择则仅保存在当前进程中。
|
||||
|
||||
会话 RPC 领域公开 `session.models` 模型目录与 `session.selectModel`。该目录从 LLM（大语言模型）注册表动态构建，并按提供方分组；每个已列出模型的精确元数据还会加入由适配器持有的推理强度 ID、名称、说明和可选默认值。各提供方的目录与精确元数据会按提供方并发加载，且彼此独立失败，因此成功加载的分组仍可与可重试的失败记录一同使用。模型是否位于目录仅供参考：如果当前模型的已注册提供方没有列出该模型，系统会将其作为未列出行插入；精确解析则决定路由与显式推理强度是否可用。选择操作通过 `resolveCallConfig` 拒绝不支持的推理强度 ID，并在更新目标前具体化适配器配置的默认值。
|
||||
|
||||
浏览器中的 `ModelService` 为每个实时会话持有一个 `ModelDirectory`。其快照包含当前完整目标、分组目录、提供方失败记录、操作错误，以及 `idle`、`loading`、`ready`、`selecting`、`error` 状态。挂载时会预先填充触发器标签，此后每次打开菜单都会刷新目录。目录与选择调用共用操作代次，防止较早响应覆盖较新结果；连接重置会先丢弃当前进程中的投影，再恢复 Host 目标。失败时保留先前的当前目标和可用分组。
|
||||
|
||||
`@deepseek-ai/dsh-client-ui-conversation` 将会话作用域的单实例 slot `conversation.input.model` 声明为其输入栏 entry 的子 slot。InputBar 在尾部控件区将该 seat 渲染于 pending 指示器与主按钮之前；该 seat 接收输入栏的 `locked` owner prop 与会话作用域。`@deepseek-ai/dsh-client-ui-model` 占用该 seat，并在同一目录上提供 `/model`。其紧凑型触发器显示目录中的模型名称与生效的推理强度标签；元数据缺失时则回退到相应 ID。向上展开的菜单首先提供 Model，并在当前精确模型支持时提供 Effort；Model 可深入提供方分组，Effort 可深入适配器排序的级别。仅当适配器没有配置模型默认值时，才显示提供方默认值行。
|
||||
|
||||
生产环境的浏览器名册是 `apps/cli/cordis.yml` 中的平铺 config tree；模型功能对应其中一行 `dshClient` 配置项，而不是 Web boot 代码中硬编码的包。其包 manifest（元数据清单）将加载顺序置于运行时与命令功能之后；Cordis 服务注入则等待 conversation slot 可用，再注册 composer 占用方。
|
||||
|
||||
## 考虑过的替代方案
|
||||
|
||||
**分别使用提供方与模型下拉框。** 模型列表依赖提供方，每次更改都需要经过两阶段交互。单个分组菜单仍以提供方组织模型，同时不会增加触发器或各行的显示长度。
|
||||
|
||||
**在 Web 客户端中硬编码当前 DeepSeek 目录。** 该目录会与已注册适配器发生偏离，也会排除部署自有的提供方。LLM 注册表继续作为提供方与模型元数据的真源，也负责呈现部分查询失败。
|
||||
|
||||
**将 `High`／`Max` 保留为客户端本地 UI 状态。** 静态 DeepSeek 标签无法覆盖 `off`、pi-ai 的提供方词汇、适配器默认值与校验，也不能参与恢复或下一次提供方请求。精确模型元数据拥有可选词汇，会话目标则拥有已选择的 ID。
|
||||
|
||||
**将选择设为全局默认值。** 全局变更会意外改道其他已打开的对话。目标仅属于一个实时会话；对于没有已记录请求的会话，Host 配置仍是默认值。
|
||||
|
||||
**agent 运行期间拒绝更改。** 共享原子目标已经将当前组装步骤与下一次选择分离。保持选择器可用，可以让用户为下一个步骤预先选择模型，而不会改变正在执行的请求。
|
||||
|
||||
**将每次点击作为新的会话事件持久化。** 只有提示词组装采用某项选择后，该选择才对模型可见。持久化尚未使用的 UI 意图，会增加一个无法重建模型请求的持久事件；现有 `request/header` 会记录首次实际使用该路由的请求。
|
||||
|
||||
## 影响
|
||||
|
||||
任何由 Host 支撑的 Web 对话（包括空白会话）都可以在动态发现的提供方分组和适配器持有的推理级别之间切换，而无需显示重复的 `provider/model` 标签。当前实际采用的目标会在恢复和重连后保留；目录名称仅用于呈现，而选择和持久化使用提供方／模型／推理强度 ID。某个提供方的目录或精确元数据不可用时，只有相应分组会降级。路由变更可能降低提供方侧的缓存复用率，但选择器不会添加任何提示词内容，也不会干扰正在执行的步骤。没有推理元数据的模型不显示 Effort 行。
|
||||
|
||||
## 测试
|
||||
|
||||
Host 测试固定分组发现、目录与精确元数据失败隔离、已记录推理强度恢复、当前未列出目标、不支持的推理强度拒绝、默认值具体化，以及切换仅影响下一次组装。客户端测试固定共享目录、重连恢复与完整目标提交。组件测试固定动态推理强度标签、说明、提供方默认值展示与推理强度提交。无密钥 built-app fixture（测试前置数据）加载生产模型插件，选择 OpenAI 的 GPT-5 及其 Max 推理强度，发起一个轮次，并验证下一条生成的响应会报告两个 ID。
|
||||
@@ -1,6 +0,0 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write
|
||||
2026-07-26-code-mode-trajectory-waterfall-spans.md: fe4dcc25dbf211cf69e0d33937cf87a7482852e2
|
||||
2026-07-26-code-mode-trajectory-waterfall-spans.zh.md: aaae06b1fca1b5587d06aa7704adec421d2b2c27
|
||||
@@ -1,31 +0,0 @@
|
||||
# Agent Note: Code Mode sub-calls in the trajectory and waterfall views
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-07-26-code-mode-trajectory-waterfall-spans.zh.md)
|
||||
|
||||
> Scope: the final PR of the Code Mode UI stack — sub-dispatch rendering in the two non-chat views. Chat nesting is owned by the [sub-call rows note](2026-07-26-code-mode-chat-subcall-rows.md); the timing this consumes is the [live-parallel note](2026-07-26-code-mode-live-parallel-dispatch.md)'s start/settle pair.
|
||||
|
||||
## Problem
|
||||
|
||||
Trajectory and waterfall still rendered a `run_code` turn as one opaque Tool cell / one node-count bar. The chat view got nested sub-rows in the earlier PRs, but the two analytical views — whose whole purpose is structure and timing — showed none of the sub-call structure and none of the per-sub-call wall time the dispatch pair now records. Waterfall sub-spans were deliberately deferred until that pair existed: a span without real timing would have been a lie.
|
||||
|
||||
## Decision
|
||||
|
||||
**Trajectory: `subtool` cells interleaved after their parent Tool cell. Waterfall: real-time sub-lanes under the owning turn row.**
|
||||
|
||||
- **Trajectory**: the layout fold takes the snapshot's `codeDispatches` index; after each Tool cell whose `callId` has dispatches (assistant-block calls, orphan results, and running calls alike), it interleaves one `subtool` cell per sub-dispatch in start order — indexes stay sequential across the interleave. A settled sub-call's duration is its start/settle pair (`durationSeconds(sub.time, sub.callTime)`); a running one shows the em dash, exactly the native in-flight convention. The new cell kind wears a `Sub` tag (business tint) and a 28px indent so nesting reads at a glance.
|
||||
- **Waterfall**: `deriveSubSpans` folds the dispatch index into per-turn lanes with REAL timing — each parent's dispatch window is first start → last settle, and every lane's offset/width is its fraction of that window, so parallel sub-calls (PR3) visibly overlap. Each lane carries a `timing` provenance tag: `measured` (pair observed), `running` (settle pending — extends to the window end at reduced opacity), or `unknown` (settle-only replay window, `callTime: null` — drawn hollow and titled "duration unknown", never a fabricated 0 ms). Lanes draw under the owning turn's bar row, scaled into a fixed lane budget.
|
||||
- Both views read `codeDispatches` through the standard snapshot hook — no new wire data, no new stores; replay renders identically to live by construction.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
**Fold sub-calls into the turn-span node counts (weight the existing bars).** Rejected: it hides exactly the structure this stack exists to show, and node-count weighting is already flagged as a stand-in (deviation ledger #3).
|
||||
|
||||
**A dedicated sub-call panel instead of in-view nesting.** Rejected: the stack's settled UX is nesting under the parent everywhere; a separate panel would diverge from chat and double the selection plumbing.
|
||||
|
||||
**Defer waterfall lanes until the P-III duration-lane redesign.** Rejected: the sub-lane timing is real today (the pair), and the fraction-of-window rendering is independent of whatever the turn-level lanes become; deferring would strand the stack's timing payoff.
|
||||
|
||||
## Consequences
|
||||
|
||||
The waterfall carries the first REAL wall-time rendering in the client (turn bars remain node-count stand-ins — the contrast is deliberate and labeled by hover titles). Trajectory cell indexes now count sub-calls, so `#N` totals grow on Code Mode turns. Specs pin the interleave order and durations, the running em-dash arm, window fractions (offsets/widths), the running-lane extension, the unknown-timing (settle-only) lane, and the rendered lane under the turn row; the built-client Code Mode fixture snapshot additionally pins both tabs' assembled rendering (sub-cells with real +0.8s durations, measured lanes).
|
||||
@@ -1,31 +0,0 @@
|
||||
# Agent Note：trajectory 与 waterfall 视图中的 Code Mode 子调用
|
||||
|
||||
Status: implemented
|
||||
|
||||
[English](2026-07-26-code-mode-trajectory-waterfall-spans.md) | 中文
|
||||
|
||||
> 范围：Code Mode UI 堆叠 PR（Pull Request）链的最后一个 PR，涵盖两个非 chat 视图中的子分发渲染。chat 的嵌套归[子调用行 Agent Note](2026-07-26-code-mode-chat-subcall-rows.md)所有；本篇所消费的计时即[实时并行 Agent Note](2026-07-26-code-mode-live-parallel-dispatch.md)的 start/settle 事件对。
|
||||
|
||||
## 问题
|
||||
|
||||
trajectory 过去仍把一个 `run_code` 轮次渲染为单个不透明的 Tool 单元格，waterfall 则渲染为一根节点计数条。chat 视图在此前的几个 PR 中已获得嵌套子行，但这两个分析视图（其全部意义恰恰是结构与计时）过去既不显示任何子调用结构，也不显示分发事件对如今已记录的逐子调用墙钟时间。waterfall 的子调用 span 曾被刻意推迟到该事件对存在之后：没有真实计时的 span 就是在撒谎。
|
||||
|
||||
## 决策
|
||||
|
||||
**trajectory：`subtool` 单元格穿插在其父 Tool 单元格之后。waterfall：所属轮次行之下、带真实计时的子泳道（sub-lane）。**
|
||||
|
||||
- **trajectory**：布局 fold 接收快照的 `codeDispatches` 索引；凡某个 Tool 单元格的 `callId` 名下存在分发（assistant 块内的调用、孤儿结果与运行中的调用一视同仁），fold 就在该单元格之后按启动顺序为每个子分发穿插一个 `subtool` 单元格，索引在整个穿插序列中保持连续编号。已结算子调用的耗时来自其 start/settle 事件对（`durationSeconds(sub.time, sub.callTime)`）；运行中的子调用则显示破折号，与原生的进行中约定完全一致。新增的单元格类型带有 `Sub` 标签（business 色调）与 28px 缩进，嵌套关系一眼可辨。
|
||||
- **waterfall**：`deriveSubSpans` 把分发索引折叠成带真实计时的逐轮次泳道：每个父调用的分发窗口为首个 start → 最后一个 settle，每条泳道的偏移/宽度即其在该窗口中的占比，因此并行的子调用（PR3）会肉眼可见地重叠。每条泳道带有 `timing` 来源标记：`measured`（观察到了成对事件）、`running`（settle 未到 — 以较低不透明度延伸至窗口末端）或 `unknown`（回放窗口只含 settle、`callTime: null` — 画成空心并以「duration unknown」为悬停标题，绝不伪造 0 ms）。泳道绘制在所属轮次的条形行之下，并缩放进固定的泳道预算。
|
||||
- 两个视图都经由标准的快照 hook 读取 `codeDispatches`：没有新的 wire 数据，也没有新的 store；回放的渲染由构造保证与实时完全一致。
|
||||
|
||||
## 曾考虑的替代方案
|
||||
|
||||
**把子调用折入轮次 span 的节点计数（给既有的条加权）。** 否决：它隐藏的恰恰是本堆叠 PR 链存在就是为了展示的结构，而且节点计数加权本就已被标记为占位（偏差账本 #3）。
|
||||
|
||||
**用专用的子调用面板取代视图内嵌套。** 否决：本堆叠 PR 链已敲定的 UX 是处处嵌套在父级之下；独立面板会与 chat 发生偏差，还会让选中接线翻倍。
|
||||
|
||||
**把 waterfall 泳道推迟到 P-III 的时长泳道重新设计。** 否决：子泳道的计时如今已是真实的（即那对事件），而按窗口占比的渲染与轮次级泳道将来的形态无关；推迟只会让本堆叠 PR 链的计时收益搁浅。
|
||||
|
||||
## 后果
|
||||
|
||||
waterfall 承载了 client 中第一处真实的墙钟时间渲染（轮次条仍是节点计数的占位；这一反差是有意为之，并由悬停标题标注）。trajectory 的单元格索引现在会把子调用计入，因此 Code Mode 轮次上的 `#N` 总数会随之增大。spec 锁定穿插顺序与耗时、运行中的破折号分支、窗口占比（偏移/宽度）、运行中泳道的延伸、unknown 计时（仅 settle）泳道，以及轮次行之下实际渲染出的泳道；构建产物级的 Code Mode fixture 快照另行锁定两个标签页的组装后渲染（带真实 +0.8s 耗时的子单元格、measured 泳道）。
|
||||
@@ -0,0 +1,6 @@
|
||||
# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each
|
||||
# side as of the last confirmed-consistent state. Both languages carry equal authority;
|
||||
# after editing either side, bring the other along and re-record with:
|
||||
# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-07-27-assistant-timing-header-trailing.md
|
||||
2026-07-27-assistant-timing-header-trailing.md: a315a0620c63f25660220e55cf5187d7117c14d1
|
||||
2026-07-27-assistant-timing-header-trailing.zh.md: 84b8612337a345912371e37952195e4602f7ca25
|
||||
@@ -0,0 +1,25 @@
|
||||
# Agent Note: Assistant timing line renders after the message body
|
||||
|
||||
Status: implemented
|
||||
|
||||
English | [中文](2026-07-27-assistant-timing-header-trailing.zh.md)
|
||||
|
||||
## Problem
|
||||
|
||||
The TUI assistant message opened with a single header line joining the `Assistant` label and the step-timing string (`Assistant · Model wait 0.0s · Completed …`). Placing the timing before the body pushed the durations away from the answer they describe and, once completed, buried the reply's first line under a metadata line the reader scans past.
|
||||
|
||||
## Decision
|
||||
|
||||
**Split the label from the timing; render the timing as the message's trailing line.**
|
||||
|
||||
`AssistantMessageComponent` (packages/ui/tui/src/index.ts) now emits the bold `Assistant` label as the first line and appends the dim timing string (already assembled by `StreamingAssistantComponent.rebuild()` as `header`, including the `· Completed …` suffix when settled) as the last child, after reasoning and text. The timing content, bucket-hiding, and completion-time behavior are unchanged — only its position moved from the top to the bottom of the message.
|
||||
|
||||
## Alternatives considered
|
||||
|
||||
**Move the whole header line (label included) to the end.** Rejected: the `Assistant` label orients the reader to who is speaking and belongs at the top like the `You` label; only the timing metadata benefits from trailing placement.
|
||||
|
||||
**Keep the timing inline but below the label as a second top line.** Rejected: that still separates the durations from the completed answer and keeps two metadata lines between the prompt and the reply.
|
||||
|
||||
## Consequences
|
||||
|
||||
Each assistant message reads label → reasoning → answer → timing, so completed timing sits next to the reply it measures. The keyless TUI snapshot suite was refreshed to pin the new layout across every fixture; four `tui.spec.ts` assertions that matched the old inline `Assistant · Model wait …` string now assert the label and timing separately, since the two no longer render contiguously.
|
||||
Some files were not shown because too many files have changed in this diff Show More
Reference in New Issue
Block a user