Coder/deepseek-harness
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

refactor agent pre-step inbox lifecycle

Browse Source
This commit is contained in:
_Kerman
2026-07-31 19:21:16 +08:00
parent c2ff9ddec8
commit fcc2b5e282
267 changed files with 2052 additions and 1546 deletions
Download Patch File Download Diff File Expand all files Collapse all files

47
docs/core-data-structures/core.md
View File

@@ -485,7 +485,7 @@ Source: [`packages/core/agent/src/types.ts`](../../packages/core/agent/src/types
type InboxTarget = 'next-turn' | 'next-step'
```
Every pending occurrence is its `UserMessage`; `MessageId` is the sole identity. `Inbox.splice(target, start, deleteCount, inserted, outcome?)` uses standard splice coordinates, rejects duplicate pending message ids, and records the normalized mutation as durable `agent/inbox/spliced`. Replaying those events reconstructs both `nextTurn` and `nextStep`, including edits, insertion, admission, and cancellation.
Every pending occurrence is its `UserMessage`; `MessageId` is the sole identity. `Inbox.append`, `prepend`, `update`, `remove`, and `splice` record normalized durable `agent/inbox/spliced` mutations and reject duplicate pending ids. Ordinary removals are cancellations. `claim(target)` atomically removes the proposed step batch through pure deletion splices; the loop separately emits per-message claimed notifications. Whole-queue consumers such as UI projections reconstruct `nextTurn` and `nextStep` from the durable splices, while consumers following one message use the exact `agent/inbox/inserted`, `claimed`, and `discarded` notifications.
```ts type-equiv
/** Options for {@link Agent.cancel}. */
@@ -568,18 +568,18 @@ interface Agent {
steer(message: UserMessage): void
/**
* Append model-facing context without running the model. Admission or an
* open turn stages it at the next safe log position; outside that window it
* appends immediately without opening a turn. If admission closes without a
* turn, a context-only boundary appends immediately; context staged beside
* steering remains pending with it.
* Queue model-facing context for the next pre-step without waking the
* driver. Collecting and running drivers claim it at the nearest later
* step boundary; idle drivers leave it pending until follow-up or steering
* wakes them. It may miss a request whose pre-step already claimed its
* batch. Cancellation or disposal may discard pending context.
* @param message - identified injected context and its producer provenance.
*/
inject(message: UserMessage): void
}
```
`AgentStatus` is `'idle' | 'running'`, and `SessionId` is branded. Disposal removes the agent from the registry and emits `agent/disposed`; it is not a terminal status value. `running` describes the driver-wide drain interval and may span consecutive queued turns; it does not prove a turn is still open. `followup()` returns no handle: its `MessageId` identifies durable inbox and admission facts, not a later assistant output or turn ending. `whenIdle()` observes the whole agent, so callers may call a receipt-to-idle interval a run only when they explicitly own that interval ([decision](../../.agents/notes/implemented/architecture/2026-07-30-followup-enqueue-and-owned-runs.md)). `AgentOptions` is merge-extensible: core declares `provider?`, `model?`, and `maxTokens?` (dispatch requires provider and model after `agent/request`). When present, `maxTokens` must be a positive safe integer and caps every conversation-model request; omission allows the exact-model adapter default to materialize before the request header, or otherwise leaves provider behavior unchanged. Persona belongs to `dsh-system-prompt`: an agent-scoped `deployment:persona` may shadow the global default.
`AgentStatus` is `'idle' | 'running'`, and `SessionId` is branded. Disposal removes the agent from the registry and emits `agent/disposed`; it is not a terminal status value. `running` describes the driver-wide drain interval and may span consecutive queued turns; it does not prove a turn is still open. `followup()` returns no handle: its `MessageId` identifies durable inbox insertion, claim, and discard facts, not a later assistant output or turn ending. `whenIdle()` observes the whole agent, so callers may call a receipt-to-idle interval a run only when they explicitly own that interval ([decision](../../.agents/notes/implemented/architecture/2026-07-30-followup-enqueue-and-owned-runs.md)). `AgentOptions` is merge-extensible: core declares `provider?`, `model?`, and `maxTokens?` (dispatch requires provider and model after `agent/request`). When present, `maxTokens` must be a positive safe integer and caps every conversation-model request; omission allows the exact-model adapter default to materialize before the request header, or otherwise leaves provider behavior unchanged. Persona belongs to `dsh-system-prompt`: an agent-scoped `deployment:persona` may shadow the global default.
The cause is a TypeScript-enforced same-process input. An active cancellation holder copies it into the runtime-only `AbortSignal.reason`; a signal grants cooperating listeners no classification authority. Durable `turn/end` retains the coarse `{ kind: 'aborted' }` outcome; request provenance would require a separate durable event rather than overloading the terminal result.
@@ -591,22 +591,31 @@ The process-local initiator carried by `ctx.agents` is the exact `Agent` above,
## Interception decisions
Prompt decisions use the same identified `UserMessage` shape as durable user-role input. The allowed batch is authoritative and preserves every message's identity and provenance. Hook bridges map their native decision fields onto this typed result.
Pre-step decisions use the same identified `UserMessage` shape as durable user-role input. The entered batch is authoritative and preserves every message's identity and provenance. Hook bridges map their native decision fields onto this typed result.
Source: [`packages/core/agent/src/types.ts`](../../packages/core/agent/src/types.ts)
`agent/prompt-submit` returns a `PromptDecision` before a turn opens. Allow supplies the complete admitted batch; block rejects admission without creating turn events and must choose whether to discard the claimed messages. Messages not claimed by that admission remain pending:
`agent/pre-step` receives the exclusive claimed batch and the proposed step's coordinates and cancellation signal. The initial proposal runs before its turn opens; a tool continuation may submit an empty claimed batch between steps:
```ts type-equiv
/**
* Prompt interception result. An allowed batch replaces the submitted
* messages; a listener wrapping `next()` preserves that batch unless it
* intentionally replaces it. A blocked batch explicitly chooses whether to
* discard the claimed messages; unclaimed work remains pending.
*/
type PromptDecision =
| { kind: 'allow'; messages: UserMessage[] }
| { kind: 'block'; reason: string; discardClaimed: boolean }
/** Coordinates and cancellation for a proposed step. */
interface PreStepContext {
/** Turn that will own the step. */
readonly turn: number
/** Step proposed by the loop. */
readonly step: number
/** Current turn cancellation signal. */
readonly signal: AbortSignal
}
```
It returns a `PreStepDecision`. Reject opens no step. Enter supplies the complete message batch appended after `step/start`; claimed messages omitted by the final decision remain removed, while input inserted after the claim stays pending:
```ts type-equiv
/** Whether and with which messages the loop enters a proposed step. */
type PreStepDecision =
| { kind: 'reject' }
| { kind: 'enter'; messages: UserMessage[] }
```
`agent/request-error` runs after a failed model step closes and before its turn closes. Listeners can repair durable state or await policy work while the failed turn's signal is still live. A handling listener returns `{ kind: 'retry' }` without calling `next()`; the default `undefined` leaves the failure terminal.
@@ -616,7 +625,7 @@ type PromptDecision =
type RequestErrorAction = { kind: 'retry' } | undefined
```
`agent/step` is the single serial boundary before request derivation. `agent/turn-stopping` runs when a turn has no tool or steering continuation, before one final steering drain.
`agent/pre-step` is the single serial boundary before request derivation. `agent/turn-stopping` runs when a turn has no tool or steering continuation, before one final steering drain.
`agent/session-start` carries a `SessionStartSource` (why the session lifecycle began; a bridge keys its SessionStart matcher on it):