Merge branch 'worktree/agent-execution-context-rfc' into worktree/explicit-turn-signal

packages/core/agent/README.md

@@ -35,6 +35,8 @@ Most interception points are cooperative waterfalls returning seam-specific deci
 
 Every asynchronous turn seam receives the same explicit `AbortSignal` for that turn. Listeners may cooperate with cancellation but must not retain the signal to control another turn; ambient `ctx.agentExecution` identity carries no liveness or cancellation authority. The signal and typed cancellation contract are defined by the [explicit turn cancellation RFC](../../../docs/rfc/implemented/architecture/2026-07-16-explicit-turn-cancellation.md).
 
+`PromptDecision.additionalContexts` is an array so every injected context keeps its own source, envelope, and metadata. A `ContinuationDecision` reason is narrower: it becomes a `steering/message`, not a `context/message`, and therefore carries only content and source.
+
 Turn and step boundaries and the model token stream are durable `session/event` facts rather than mirrored `agent/*` notifications. Consumers read `turn/*`, `step/*`, and `assistant/chunk` from the session feed; tool policy and outcome observation belong to the complete pipeline documented by [`dsh-tools`](../tools/README.md).
 
 ### Agent interface (`types.ts`)
@@ -43,7 +45,7 @@ The handle every plugin programs against:
 
 - `agent.send(content, options?)` — queue a message; starts a turn when idle. Content and resolved source become one detached, deeply frozen lossless-JSON record before `agent/queued` and enqueue; invalid data throws synchronously, and caller or notification-listener in-place mutation cannot change the log or model input (`agent/prompt-submit` still rewrites by returning replacement content).
 - `agent.steer(content, options?)` — steer a running turn (inject between steps); uses the same owned acceptance boundary and behaves like `send` when idle
-- `agent.inject(content, options?)` — inject in-session context (context/message event); the next request sees it. Does not run the model. While a turn is open it joins that turn; while idle it is wrapped in a one-shot `injection` turn so every event stays turn-enclosed ([the turn-enclosure invariant](../../../docs/rfc/implemented/architecture/2026-06-15-turn-enclosure-invariant.md))
+- `agent.inject(content, options?)` — inject in-session context (`context/message` event); the next request sees it. `options.envelope` defaults to the canonical `<context>` framing and may be `'raw'` when the caller owns a complete familiar frame; `options.meta` persists opaque JSON state without rendering it. Does not run the model. While a turn is open it joins that turn; while idle it is wrapped in a one-shot `injection` turn so every event stays turn-enclosed ([the turn-enclosure invariant](../../../docs/rfc/implemented/architecture/2026-06-15-turn-enclosure-invariant.md))
 - `agent.cancel(cause?)` — cancel ALL pending work: clears the queued + steering FIFOs, aborts the active turn, and drops queued work not yet claimed by the driver. `AgentCancelCause` is the runtime-only `{ kind: 'user' } | { kind: 'parent' }`; omission means `user`, the first cause wins for an active turn, and ACP `session/cancel` maps to `user`. `normalizeAgentCancelCause()` provides the same strict detached-value boundary used by the concrete loop: validation is synchronous even while idle, accepts only an exact plain object, and returns a frozen detached cause. After validation, `agent.cancel()` is a safe no-op when no work exists.
 - `agent.whenIdle()` — resolve once the agent reaches quiescence after settling out of `running` (idle → immediately; disposed → awaits the loop exit). A non-owner's quiescence-observation hook: it observes the work settling WITHOUT tearing the agent down. Teardown is separate — a lifecycle owner stops and unregisters via `AgentHandle.dispose()`, which awaits the loop exit directly.
 - `agent.session`, `agent.status`, `agent.options`, `agent.id`

packages/core/agent/src/types.ts

@@ -10,7 +10,7 @@ import type { Context } from 'cordis'
 import type { Scoped } from '@deepseek-ai/dsh-scope'
 import type { ContentBlock, LlmCallConfig, Message, MessageSource } from '@deepseek-ai/dsh-llm'
 import type {} from '@deepseek-ai/dsh-system-prompt'
-import type { Session } from '@deepseek-ai/dsh-session'
+import type { ContextEnvelope, JsonValue, Session } from '@deepseek-ai/dsh-session'
 
 /** Identifies one live agent in the registry. */
 export type AgentId = Branded<'AgentId'>
@@ -23,6 +23,7 @@ export type AgentId = Branded<'AgentId'>
 export function AgentId(id: string): AgentId {
   return id as AgentId
 }
+
 declare module '@deepseek-ai/dsh-system-prompt' {
   interface AssembleContext {
     /** Agent for this assembly; absent on diagnostics. When present, `scope` must identify the same agent. */
@@ -32,7 +33,9 @@ declare module '@deepseek-ai/dsh-system-prompt' {
 
 /** Merge-extensible agent creation options. Persona belongs to system-prompt sections. */
 export interface AgentOptions {
-  /** Model name (must have a registered adapter at call time). */
+  /** Provider route (must have a registered adapter at call time). */
+  provider?: string
+  /** Model id interpreted by the selected provider adapter. */
   model?: string
 }
 
@@ -41,6 +44,14 @@ export interface SendOptions {
   source?: MessageSource
 }
 
+/** Options specific to durable synthetic context injection. */
+export interface InjectOptions extends SendOptions {
+  /** Keep the canonical context tag, or send caller-owned framing verbatim. */
+  envelope?: ContextEnvelope
+  /** Opaque JSON state retained in the session event but hidden from the model. */
+  meta?: JsonValue
+}
+
 /**
  * An agent's lifecycle state, emitted on every transition as `agent/status`:
  * `idle` (parked, waiting for queued work), `running` (a turn is in progress),
@@ -53,21 +64,25 @@ export type AgentStatus = 'idle' | 'running' | 'disposed'
 export interface HookContext {
   content: ContentBlock[]
   source: MessageSource
+  /** Keep the canonical context tag, or use caller-owned framing verbatim. */
+  envelope?: ContextEnvelope
+  /** Opaque JSON state retained in the session event but hidden from the model. */
+  meta?: JsonValue
 }
 
 /**
- * Prompt interception result. `allow.content` replaces the prompt and
- * `additionalContext` becomes a separate context message. `block` records a
+ * Prompt interception result. `allow.content` replaces the prompt and each
+ * `additionalContexts` entry becomes a separate context message. `block` records a
  * durable `prompt/blocked`; an all-blocked batch ends a zero-step rejected turn.
  */
 export type PromptDecision =
-  | { kind: 'allow'; content?: ContentBlock[]; additionalContext?: HookContext }
+  | { kind: 'allow'; content?: ContentBlock[]; additionalContexts?: HookContext[] }
   | { kind: 'block'; reason: string }
 
 /** Turn continuation override; a continue reason is recorded as next-step steering in the same turn. */
 export type ContinuationDecision =
   | { action: 'stop' }
-  | { action: 'continue'; reason?: HookContext }
+  | { action: 'continue'; reason?: { content: ContentBlock[]; source: MessageSource } }
 
 /**
  * The terminal subset of {@link ContinuationDecision}. A listener on
@@ -173,7 +188,7 @@ export interface Agent {
    * turn joins it at the current log position. Disposal awaits idle checkpoints;
    * flush failures are reported through `agent/error`, not thrown to the caller.
    */
-  inject(content: ContentBlock[], options?: SendOptions): void
+  inject(content: ContentBlock[], options?: InjectOptions): void
 
   /**
    * Clear queued and steering work, including work waiting to start, and abort