refactor(agent): remove message metadata channel

This commit is contained in:
_Kerman
2026-07-24 14:05:33 +08:00
parent 7d5c8b12c0
commit 5c7505b208
55 changed files with 1610 additions and 502 deletions

View File

@@ -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
2026-07-22-unified-send-and-coalesced-user-messages.md: 9eb355128b217a0ea8dc09daf4e83334f6aeaa10
2026-07-22-unified-send-and-coalesced-user-messages.zh.md: 679c9100aa49777b7b601725bdfc077f5f2dab0e
2026-07-22-unified-send-and-coalesced-user-messages.md: dbd157ad4c81f278cf1417765816497b9e3568df
2026-07-22-unified-send-and-coalesced-user-messages.zh.md: dea32f142c4ef005ff04d00ef8c8ae1ea19c9705

View File

@@ -8,23 +8,23 @@ English | [中文](2026-07-22-unified-send-and-coalesced-user-messages.zh.md)
The agent's public driving surface had grown three near-parallel verbs — `send`, `steer`, `inject` — each with its own options type, its own live event story, and its own durable event. `send` and `steer` both queued a frozen inbox record and emitted `agent/queued`; `inject` bypassed the inbox and wrote a separate `context/message` durable event. The three verbs actually vary along only two independent axes: which queue an item joins (a whole new turn versus the active turn) and whether the item makes the model run. Encoding that 2×2 as three hand-written methods hid the symmetry, made "queue a turn without waking the driver" unreachable, and left `cancel()` with no way to abort a turn while preserving queued work.
Separately, `context/message` and `user/message` had converged: the surface projected both as verbatim user-role content, and the only real difference was that injected context carried `source`/`meta` and was "not a prompt." Two event types for one projection meant every consumer branched on event type to answer "is this a human prompt?", and the goal system used the type split as a side channel (round-zero state changes were `context/message`, admitted rounds were `user/message`).
Separately, `context/message` and `user/message` had converged: the surface projected both as verbatim user-role content, and the only real difference was that injected context carried a non-user `source` and was "not a prompt." Two event types for one projection meant every consumer branched on event type to answer "is this a human prompt?", and the goal system used the type split as a side channel (round-zero state changes were `context/message`, admitted rounds were `user/message`).
## Decision
**One primitive, three preset aliases.** `Agent` is now an abstract class whose single abstract `send(content, { target, wakeup, source, contexts, meta })` covers the (`target` × `wakeup`) matrix. `followup` (`next-turn`/wakeup), `steer` (`next-step`/wakeup), and `inject` (`next-step`/no-wakeup) are concrete delegates on the base class, so concrete drivers implement `send` once and inherit the ergonomic presets. `wakeup` means "make the model run": wake a parked driver for a `next-turn` item, or force a continuation for a running `next-step` item. `send` defaults to `{ target: 'next-turn', wakeup: true }`, so every prior bare `agent.send(content)` call keeps its exact behavior. `next-turn`/no-wakeup (queue without waking) is now representable with no alias and no current caller.
**One primitive, three preset aliases.** `Agent` is now an abstract class whose single abstract `send(content, { target, wakeup, source, contexts })` covers the (`target` × `wakeup`) matrix. `followup` (`next-turn`/wakeup), `steer` (`next-step`/wakeup), and `inject` (`next-step`/no-wakeup) are concrete delegates on the base class, so concrete drivers implement `send` once and inherit the ergonomic presets. `wakeup` means "make the model run": wake a parked driver for a `next-turn` item, or force a continuation for a running `next-step` item. `send` defaults to `{ target: 'next-turn', wakeup: true }`, so every prior bare `agent.send(content)` call keeps its exact behavior. `next-turn`/no-wakeup (queue without waking) is now representable with no alias and no current caller.
**inject keeps its mechanism.** The `next-step`/no-wakeup path is exactly the old `inject`: durable model-facing context appended at the current log position (deferred behind an executing tool batch), or a one-shot `injection` turn when idle. It bypasses the FIFOs entirely and defaults its source to `{ kind: 'plugin', plugin: '' }`, never `{ kind: 'user' }`.
**context/message is gone.** Injected context is now a `user/message` whose `source` is a non-`user` kind (plugin or goal). `PromptMessageData` gained the optional `meta` that `context/message` carried. The surface, derivation, and `SurfaceEventType` drop `context/message`; consumers that need "is this a human prompt?" read `source.kind === 'user'` instead of the event type. This keeps goal-authority's human-authority check exactly as strict as before — an injected message defaults to a plugin source and can never satisfy `source.kind === 'user'`.
**context/message is gone.** Injected context is now a `user/message` whose `source` is a non-`user` kind. Typed source variants carry any domain-specific durable provenance. The surface, derivation, and `SurfaceEventType` drop `context/message`; consumers that need "is this a human prompt?" read `source.kind === 'user'` instead of the event type. This keeps goal-authority's human-authority check exactly as strict as before — an injected message defaults to a plugin source and can never satisfy `source.kind === 'user'`.
**Goal replay disambiguates by round, not type.** A goal state change is a round-zero goal-sourced `user/message` carrying `goal/change` metadata; a positive round is an admitted continuation prompt. `decodeGoalEvent` now takes a `user/message` and still fails loud on goal metadata under a non-goal source or a goal source lacking metadata.
**Goal replay disambiguates by round, not type.** A goal state change is a round-zero goal-sourced `user/message` whose source carries the complete change; a positive round is an admitted continuation prompt. `decodeGoalEvent` takes a `user/message` and fails loud when goal-state content and its typed source disagree.
**`send` returns an id.** `send` (and the aliases) return an opaque branded `AgentMessageId` for the accepted message; `send`'s previous return was `void`.
**Three inbox events replace agent/queued.** `agent/inbox/enqueue` (an item entered a FIFO), `agent/inbox/dequeue` (the driver claimed one), and `agent/inbox/discard` (`cancel()` dropped pending items) each carry an `AgentMessage` — the accepted message including its returned `id`, `target`/`wakeup`, source, and contexts — so a caller can correlate a queued item with its lifecycle. Injection never touches a FIFO and emits none of these. Every FIFO entry publishes an enqueue, including the loop-authored continuation-reason steer (`agent/turn-continuation` returning `{ action: 'continue', reason }`), so the ledger stays balanced with its later dequeue or discard. The `dsh-agent` invariant companion asserts FIFO conservation: a per-agent outstanding count that dequeue and discard can never drive negative.
**cancel gains keepInbox.** `cancel(cause?, { keepInbox? })`; when true it aborts the active turn but preserves queued and steering items (no discard event, and un-started work is not dropped).
**cancel gains keepInbox.** `cancel(cause, { keepInbox? })`; callers choose the cause explicitly, and `keepInbox: true` aborts the active turn while preserving queued and steering items (no discard event, and un-started work is not dropped).
## Alternatives considered
@@ -36,7 +36,7 @@ Separately, `context/message` and `user/message` had converged: the surface proj
The delivery surface is now one primitive plus three self-documenting presets, and the (`target` × `wakeup`) matrix makes previously-unreachable combinations explicit. One durable message type serves prompts, injected context, and goal rounds, so the surface projection and every "human prompt?" check simplify to a `source` test. The cost: `Agent` became an abstract class, so object-literal test fakes must supply `followup` and cannot spread a class-typed value without re-casting (prototype methods are non-enumerable); the goal fold's channel split moved from event type to `source.round`; and every consumer that filtered `context/message` now filters `user/message` by source. The turn-enclosure and reconstruction invariants are unchanged — an idle injection still wraps a one-shot turn, now emitting `user/message` instead of `context/message`.
`wakeup` is the "should the model run" signal, so the inbox distinguishes `hasWakingQueued` (drives the loop and idle/quiescence decisions) from `hasQueued` (anything to dequeue): a lone `next-turn`/no-wakeup item stays parked at idle and rides along the next waking send, and `whenIdle`/`cancel` settle quiescence off the waking signal (a lone quiet item takes `whenIdle`'s fast path, so no waiter is ever left hanging). `SendOptions.meta` on a queued or steering send is carried onto the durable `user/message`/`steering/message`, matching injection; it is intentionally not on the live `AgentMessage` event, which carries only routing facts. Every FIFO exit publishes exactly one lifecycle event: a terminal stop that drops pending steering emits `agent/inbox/discard` for it — both at the in-turn stop point and on the post-turn drain of late steering — and a loop-authored continuation reason is snapshotted and frozen like a public send. `gen-cordis-api` collects exported classes (public members, body-stripped) so the now-class `Agent` and its transitive shapes still appear in the model-facing API catalog.
`wakeup` is the "should the model run" signal, so the inbox distinguishes waking queued work from anything available to dequeue: a lone `next-turn`/no-wakeup item stays parked at idle and rides along the next waking send, and `whenIdle`/`cancel` settle quiescence off the waking signal. Every FIFO exit publishes exactly one lifecycle event, while domain-specific durable facts travel in typed message sources rather than a parallel metadata channel. `gen-cordis-api` collects exported classes (public members, body-stripped) so the now-class `Agent` and its transitive shapes still appear in the model-facing API catalog.
## Related

View File

@@ -8,23 +8,23 @@ Status: implemented
agent 的对外驱动接口逐渐长出三个近乎平行的动词——`send`、`steer`、`inject`——各自带有独立的选项类型、独立的实时事件叙事,以及独立的持久事件。`send` 和 `steer` 都会把一条冻结的 inbox 记录入队并发出 `agent/queued`;`inject` 则绕过 inbox,写入一条独立的 `context/message` 持久事件。这三个动词实际上只沿两条独立的轴变化:一个队列项加入哪个队列(一个全新的轮次,还是当前活跃的轮次),以及这个队列项是否让模型运行。把这个 2×2 编码成三个手写方法,掩盖了其中的对称性,让“排入一个轮次但不唤醒驱动器”无法表达,也让 `cancel()` 无从在保留排队工作的前提下中止一个轮次。
另外,`context/message` 与 `user/message` 已经趋同:对外接口把二者都投影为逐字的 user 角色内容,唯一真正的区别是注入的上下文携带 `source`/`meta` 且“不是提示词”。一个投影对应两种事件类型,意味着每个消费方都要根据事件类型分支来回答“这是不是一条人类提示词?”,而 goal 系统把这种类型区分当作侧信道使用(第 0 轮的状态变更是 `context/message`,已准入的轮次是 `user/message`)。
另外,`context/message` 与 `user/message` 已经趋同:对外接口把二者都投影为逐字的 user 角色内容,唯一真正的区别是注入的上下文携带非 user `source` 且“不是提示词”。一个投影对应两种事件类型,意味着每个消费方都要根据事件类型分支来回答“这是不是一条人类提示词?”,而 goal 系统把这种类型区分当作侧信道使用(第 0 轮的状态变更是 `context/message`,已准入的轮次是 `user/message`)。
## 决策
**一个原语,三个预设别名。** `Agent` 现在是一个抽象类,其唯一的抽象方法 `send(content, { target, wakeup, source, contexts, meta })` 覆盖 (`target` × `wakeup`) 矩阵。`followup`(`next-turn`/wakeup)、`steer`(`next-step`/wakeup)和 `inject`(`next-step`/no-wakeup)是基类上的具体委托方法,因此具体驱动器只需实现一次 `send`,就能继承这些好用的预设。`wakeup` 意为“让模型运行”:为一个 `next-turn` 队列项唤醒处于停泊状态的驱动器,或为一个运行中的 `next-step` 队列项强制继续执行。`send` 默认使用 `{ target: 'next-turn', wakeup: true }`,因此此前每一次裸调用 `agent.send(content)` 都保持完全相同的行为。`next-turn`/no-wakeup(入队但不唤醒)现在可以表达,只是没有别名,也没有当前调用方。
**一个原语,三个预设别名。** `Agent` 现在是一个抽象类,其唯一的抽象方法 `send(content, { target, wakeup, source, contexts })` 覆盖 (`target` × `wakeup`) 矩阵。`followup`(`next-turn`/wakeup)、`steer`(`next-step`/wakeup)和 `inject`(`next-step`/no-wakeup)是基类上的具体委托方法,因此具体驱动器只需实现一次 `send`,就能继承这些好用的预设。`wakeup` 意为“让模型运行”:为一个 `next-turn` 队列项唤醒处于停泊状态的驱动器,或为一个运行中的 `next-step` 队列项强制继续执行。`send` 默认使用 `{ target: 'next-turn', wakeup: true }`,因此此前每一次裸调用 `agent.send(content)` 都保持完全相同的行为。`next-turn`/no-wakeup(入队但不唤醒)现在可以表达,只是没有别名,也没有当前调用方。
**inject 保留其机制。** `next-step`/no-wakeup 路径正是旧的 `inject`:在当前日志位置追加的持久、面向模型的上下文(在执行中的工具批处理之后延迟处理),或在空闲时的一次性 `injection` 轮次。它完全绕过 FIFO 队列,并把来源默认设为 `{ kind: 'plugin', plugin: '' }`,绝不是 `{ kind: 'user' }`。
**context/message 已移除。** 注入的上下文现在是一条 `user/message`,其 `source` 为非 `user` 类别(plugin 或 goal)。`PromptMessageData` 新增了 `context/message` 原本携带的可选 `meta`。对外接口、派生逻辑和 `SurfaceEventType` 都不再包含 `context/message`;需要判断“这是不是一条人类提示词?”的消费方改为读取 `source.kind === 'user'`,而不是事件类型。这让 goal-authority 的人类授权检查与此前一样严格——注入的消息默认使用 plugin 来源,永远无法满足 `source.kind === 'user'`。
**context/message 已移除。** 注入的上下文现在是一条 `user/message`,其 `source` 为非 `user` 类别。类型化 source 变体携带所有特定于领域的持久 provenance。对外接口、派生逻辑和 `SurfaceEventType` 都不再包含 `context/message`;需要判断“这是不是一条人类提示词?”的消费方改为读取 `source.kind === 'user'`,而不是事件类型。这让 goal-authority 的人类授权检查与此前一样严格——注入的消息默认使用 plugin 来源,永远无法满足 `source.kind === 'user'`。
**goal 回放靠轮次而非类型来区分。** 一次 goal 状态变更是一条第 0 轮、来源为 goal 的 `user/message`,携带 `goal/change` 元数据;正数轮次则是一条已准入的继续执行提示词。`decodeGoalEvent` 现在接收一条 `user/message`,并仍会在非 goal 来源携带 goal 元数据、或 goal 来源缺少元数据时立即报错。
**goal 回放靠轮次而非类型来区分。** 一次 goal 状态变更是一条第 0 轮、来源为 goal 的 `user/message`,其 source 携带完整变更;正数轮次则是一条已准入的继续执行提示词。`decodeGoalEvent` 接收一条 `user/message`,并在 goal 状态内容与其类型化 source 不一致时立即报错。
**`send` 返回一个 id。** `send`(以及其别名)为被接受的消息返回一个不透明的 branded `AgentMessageId`;`send` 此前的返回值是 `void`。
**三个 inbox 事件取代 agent/queued。** `agent/inbox/enqueue`(一个队列项进入某个 FIFO)、`agent/inbox/dequeue`(驱动器认领了一个)和 `agent/inbox/discard`(`cancel()` 丢弃了待处理项)都携带一条 `AgentMessage`——即被接受的消息,包含其返回的 `id`、`target`/`wakeup`、来源和上下文——因此调用方可以把一个排队项与其生命周期关联起来。注入从不触及 FIFO,也不发出这些事件中的任何一个。每一次 FIFO 入队都会发布一个 enqueue 事件,包括由 loop 生成的携带继续原因的 steer(`agent/turn-continuation` 返回 `{ action: 'continue', reason }`),因此账目会与其后的 dequeue 或 discard 保持平衡。`dsh-agent` 的不变量配套断言 FIFO 守恒:一个按 agent 计的未结算计数,dequeue 和 discard 永远无法把它压到负数。
**cancel 新增 keepInbox。** `cancel(cause?, { keepInbox? })`;当其为 true 时,它中止活跃轮次,但保留排队项和 steering 项(不发出 discard 事件,尚未启动的工作也不会被丢弃)。
**cancel 新增 keepInbox。** `cancel(cause, { keepInbox? })`;调用方显式选择 cause,且 `keepInbox: true` 会中止活跃轮次,同时保留排队项和 steering 项(不发出 discard 事件,尚未启动的工作也不会被丢弃)。
## 考虑过的替代方案
@@ -36,7 +36,7 @@ agent 的对外驱动接口逐渐长出三个近乎平行的动词——`send`
投递接口现在是一个原语加三个自解释的预设,(`target` × `wakeup`) 矩阵把此前无法表达的组合显式化。一种持久消息类型同时服务提示词、注入的上下文和 goal 轮次,因此对外接口的投影和每一处“是否人类提示词?”检查都简化为一次 `source` 判断。代价是:`Agent` 变成了抽象类,因此对象字面量形式的测试替身必须提供 `followup`,且无法在不重新做类型转换的情况下展开一个类类型的值(原型方法不可枚举);goal 折叠的通道区分从事件类型改到了 `source.round`;此前过滤 `context/message` 的每个消费方现在改为按来源过滤 `user/message`。轮次封闭与重建的不变量保持不变——空闲状态下的一次注入仍然封装成一个一次性轮次,只是现在发出 `user/message` 而非 `context/message`。
`wakeup` 是“模型是否应当运行”的信号,因此 inbox 区分 `hasWakingQueued`(驱动 loop 以及空闲/静默判定)与 `hasQueued`(是否有任何可 dequeue 的项):一个孤立的 `next-turn`/no-wakeup 队列项会停泊在空闲状态,并随下一次唤醒 send 一同带出,而 `whenIdle`/`cancel` 依据唤醒信号来结算静默(一个孤立的静默项走 `whenIdle` 的快速路径,因此不会有等待者被永久挂起)。排队 send 或 steering send 上的 `SendOptions.meta` 会被带到持久的 `user/message`/`steering/message` 上,与注入保持一致;它有意不放在实时的 `AgentMessage` 事件上,后者只携带路由事实。每一次 FIFO 退出都恰好发布一个生命周期事件:一次会丢弃待处理 steering 项的终止性停止会为它发出 `agent/inbox/discard`——既在轮次内的停止点,也在轮次结束后对迟到 steering 的清空时——而由 loop 生成的继续原因会像一次对外 send 那样被快照并冻结。`gen-cordis-api` 收集导出的类(公开成员,剥除方法体),因此如今已是类的 `Agent` 及其传递涉及的形状仍会出现在面向模型的 API 目录中。
`wakeup` 是“模型是否应当运行”的信号,因此 inbox 会区分能唤醒的排队工作与任何可 dequeue 的项:一个孤立的 `next-turn`/no-wakeup 队列项会停泊在空闲状态,并随下一次唤醒 send 一同带出,而 `whenIdle`/`cancel` 依据唤醒信号来结算静默。每一次 FIFO 退出都恰好发布一个生命周期事件,特定于领域的持久事实则通过类型化消息 source 传递,而非通过平行的元数据通道。`gen-cordis-api` 收集导出的类(公开成员,剥除方法体),因此如今已是类的 `Agent` 及其传递涉及的形状仍会出现在面向模型的 API 目录中。
## 相关

View File

@@ -16,7 +16,7 @@ The canonical surface separates transformable policy, around-dispatch control, a
- `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/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-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. Its narrower type does not carry attached contexts.
### The tool pipeline gives each phase one kind of authority