Merge origin/master into feat/remove-str-replace-editor-from-minimal

.agents/notes/implemented/architecture/2026-06-13-capability-seams.i18n.yaml

@@ -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/architecture/2026-06-13-capability-seams.md
-2026-06-13-capability-seams.md: 46a2c39e927e859c7eb95956d8586f3bf04c7b1c
-2026-06-13-capability-seams.zh.md: f44e3e68d2153149435b0fd0aaa5fd121cf3ecad
+2026-06-13-capability-seams.md: 3c552c474b499b9f1c9f60242f4773750faadafb
+2026-06-13-capability-seams.zh.md: 4e6100a2ce557d91fa18b5630c267f37f1bc0f00

.agents/notes/implemented/architecture/2026-06-13-capability-seams.md

@@ -15,7 +15,7 @@ This is distinct from "who provides vs. needs a capability at runtime", which Co
 A swappable capability has **three roles**:
 
 1. **Service Definition** — the Cordis `Service` and vocabulary types owning `ctx.<key>` and depending only on the vocabulary the contract needs (e.g. `dsh-shell`: `ShellExecutor`, `ShellRunResult`, `ShellProcess`). A definition may be an abstract class or a concrete registry service; it is never a TypeScript `interface`.
-2. **Service Provider** — a plugin that supplies or registers an implementation (e.g. `dsh-bash-local`: subprocesses, process-group kills, spill-file truncation). Sandboxed and remote providers are sibling packages implementing or registering against the same Service Definition.
+2. **Service Provider** — a plugin that supplies or registers an implementation (e.g. `dsh-bash-local`: subprocesses, provider-managed range termination, spill-file truncation). The [native-containment decision](2026-08-28-subprocess-native-containment.md) owns the local provider's OS-specific range mechanics. Sandboxed and remote providers are sibling packages implementing or registering against the same Service Definition.
 3. **Consumer** — what the model and plugins program against (e.g. `dsh-tool-bash`: the `bash` schema, with background handles registered into the generic job runtime). Consumers inject the service key and never import provider-specific types.
 
 The role names use title case: **Service Definition**, **Service Provider**, and **Consumer**. Generic uses of `provider` and `consumer` remain lowercase.

.agents/notes/implemented/architecture/2026-06-13-capability-seams.zh.md

@@ -15,7 +15,7 @@ harness 具有可替换的能力，包括 shell 执行和模型提供方。一
 一项可替换的能力包含**三个角色**：
 
 1. **Service Definition**——拥有 `ctx.<key>` 的 Cordis `Service` 和词汇类型，仅依赖约定所需的词汇（例如 `dsh-shell`：`ShellExecutor`、`ShellRunResult`、`ShellProcess`）。Service Definition 可以是抽象类，也可以是具体的注册表服务；绝不是 TypeScript `interface`。
-2. **Service Provider**——提供或注册实现的插件（例如 `dsh-bash-local`：子进程、进程组 kill、spill 文件截断）。沙箱化和远程 Service Provider 是依据同一 Service Definition 实现或注册的兄弟包。
+2. **Service Provider**——提供或注册实现的插件（例如 `dsh-bash-local`：子进程、由提供方管理的范围终止、spill 文件截断）。[原生 containment 决策](2026-08-28-subprocess-native-containment.zh.md)负责本地提供方的 OS 特有范围机制。沙箱化和远程 Service Provider 是依据同一 Service Definition 实现或注册的兄弟包。
 3. **Consumer**——模型和插件编程所面向的内容（例如 `dsh-tool-bash`：`bash` schema，后台句柄注册到通用任务运行时）。Consumer 注入服务键，从不导入 Service Provider 特有的类型。
 
 角色名使用标题式大小写：**Service Definition**、**Service Provider** 和 **Consumer**。泛指的 `provider` 和 `consumer` 仍使用小写。

.agents/notes/implemented/architecture/2026-07-05-prompt-variables-and-tool-guidance-ownership.i18n.yaml

@@ -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/architecture/2026-07-05-prompt-variables-and-tool-guidance-ownership.md
-2026-07-05-prompt-variables-and-tool-guidance-ownership.md: f4364453c5ddded2fc0cb1d733732059feee6632
-2026-07-05-prompt-variables-and-tool-guidance-ownership.zh.md: d8ee8ef906d02d5a96ae7db3fb705685618223ee
+2026-07-05-prompt-variables-and-tool-guidance-ownership.md: 35bb7c6fabc85ae6f93bdbb67e13910eea627ca3
+2026-07-05-prompt-variables-and-tool-guidance-ownership.zh.md: 53d69cf45c02f6534334561b626d2c2ae6087c05

.agents/notes/implemented/architecture/2026-07-05-prompt-variables-and-tool-guidance-ownership.md

@@ -32,7 +32,7 @@ Plugins register `{{name}}` values through `ctx.systemPrompt.variable(name, prov
 
 ### Persona as the order-0 section
 
-`dsh-system-prompt` owns `harness:identity` at first-party order `-1000` and the configured `deployment:persona` at order 0, so both survive a replacement loop. Prompt rendering has one path, `renderPrompt(assembly)`, and the routed request header therefore records the exact prompt later replayed by `ctx.tokenMeter` for compaction pressure. An agent-scoped `deployment:persona` shadows the global default and lets subagent providers install a persona before publication. The [`dsh-system-prompt` README](../../../../packages/core/system-prompt/README.md) owns the sparse named placements for identity, policy, tool guidance, generated protocol, and final-output obligations.
+`dsh-system-prompt` owns `harness:identity` at first-party order `-1000` and the configured `deployment:persona-prefix` at order 0, so both survive a replacement loop. Prompt rendering has one path, `renderPrompt(assembly)`, and the routed request header therefore records the exact prompt later replayed by `ctx.tokenMeter` for compaction pressure. An agent-scoped `deployment:persona-prefix` shadows the global default and lets subagent providers install a persona before publication. The [`dsh-system-prompt` README](../../../../packages/core/system-prompt/README.md) owns the sparse named placements for identity, policy, tool guidance, generated protocol, and final-output obligations.
 
 ### Tool guidance ownership
 

.agents/notes/implemented/architecture/2026-07-05-prompt-variables-and-tool-guidance-ownership.zh.md

@@ -32,7 +32,7 @@ Status: implemented
 
 ### Persona 作为 order-0 section
 
-`dsh-system-prompt` 拥有 first-party order 为 `-1000` 的 `harness:identity` 和 order 为 0 的配置 `deployment:persona`，因此两者在循环被替换时仍然存活。提示词渲染只有一条路径 `renderPrompt(assembly)`，已路由请求 header 因此会记录准确的提示词，稍后由 `ctx.tokenMeter` 为压缩（compaction）压力回放。agent 作用域的 `deployment:persona` 遮蔽全局默认值，允许 subagent 提供方在发布前安装 persona。[`dsh-system-prompt` README](../../../../packages/core/system-prompt/README.zh.md)规定身份、策略、工具指导、生成协议和最终输出义务的稀疏具名位置。
+`dsh-system-prompt` 拥有 first-party order 为 `-1000` 的 `harness:identity` 和 order 为 0 的配置 `deployment:persona-prefix`，因此两者在循环被替换时仍然存活。提示词渲染只有一条路径 `renderPrompt(assembly)`，已路由请求 header 因此会记录准确的提示词，稍后由 `ctx.tokenMeter` 为压缩（compaction）压力回放。agent 作用域的 `deployment:persona-prefix` 遮蔽全局默认值，允许 subagent 提供方在发布前安装 persona。[`dsh-system-prompt` README](../../../../packages/core/system-prompt/README.zh.md)规定身份、策略、工具指导、生成协议和最终输出义务的稀疏具名位置。
 
 ### 工具指导归属
 

.agents/notes/implemented/architecture/2026-07-05-reconstructable-requests.i18n.yaml

@@ -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/architecture/2026-07-05-reconstructable-requests.md
-2026-07-05-reconstructable-requests.md: bc8ba640c400b18598f18aa303f2bd1b5c5b9cdc
-2026-07-05-reconstructable-requests.zh.md: de1802aac83f1e0980172d2d541a093f2d729e4a
+2026-07-05-reconstructable-requests.md: bca93a60bf07484d73f1faf50359b72a0d00b9a3
+2026-07-05-reconstructable-requests.zh.md: c9d2a4a5d05456df8b0bd065bade8a41dd7e4e84

.agents/notes/implemented/architecture/2026-07-05-reconstructable-requests.md

@@ -24,7 +24,7 @@ Prefix-cache stability is corollary #1, not the headline: an append-only log pro
 
 `EpochHeader` records the request's non-history state: call config, rendered system prompt, and tool schemas, with empty values canonicalized to absence. Adapter-supplied effort and token defaults retain their `adapterDefaults` provenance; a Web model selection restored from the log omits an adapter-owned effort so the next resolution cannot reclassify the same effective config as an explicit selection and a false change. `request/header` always writes a full snapshot: the first loop instance uses reason `initial`, later instances use `resume`, an in-instance change uses `change`, and an unchanged envelope beginning an explicitly declared message series or following a surface replacement uses `series`. A `change` snapshot carries `startsSeries: true` when the changed request also starts a series, preserving the two independent facts without a duplicate header. Ordinary append-only later Turns, further same-series Steps, and retries inherit the latest snapshot. `foldRequestHeader` selects the latest snapshot. Legacy `request/header-delta` events and the removed `fallback` reason are rejected when appended or loaded.
 
-Each proposed step first claims its inbox batch and runs `agent/pre-step`. Rejection opens no step; enter opens `step/start`, records the final message batch as `user/message` events, and may use `startsRequestSeries: true` to declare a distinct series. The step then assembles the system prompt and tools, while `agent/request` may replace only the frozen call-config seed. The loop records the owed initial, resume, change, or series full snapshot, builds `GenerateOptions` from derived messages and that header, and deep-freezes it while leaving `AbortSignal` live. The first call config starts from explicit `AgentOptions`, preserving fork overrides and resume reconfiguration; later calls start from the folded header.
+Each proposed step first claims its inbox batch and runs `agent/pre-step`. Rejection opens no step; enter opens `step/start`, records the final message batch as `user/message` events, and may use `startsRequestSeries: true` to declare a distinct series. The step then assembles the system prompt and tools, while `agent/request` may replace only the frozen call-config seed. The loop records the owed initial, resume, change, or series full snapshot, builds `GenerateOptions` from derived messages and that header, and freezes it while leaving `AbortSignal` live. The [request-freeze provenance decision](../simplification/2026-09-06-agent-request-freeze-provenance.md) owns reuse of completed message freezes and per-request local header freezing. The first call config starts from explicit `AgentOptions`, preserving fork overrides and resume reconfiguration; later calls start from the folded header.
 
 **The open step is the reconstruction boundary.** Its entered `user/message` batch and any newly written `request/header` precede request dispatch. Injection after the atomic claim joins a later request, while a listener that must affect this request returns messages through `agent/pre-step`. Header reconstruction selects the step's `request/header`, or carries the prior snapshot when no new header is written.
 

.agents/notes/implemented/architecture/2026-07-05-reconstructable-requests.zh.md

@@ -24,7 +24,7 @@ Status: implemented
 
 `EpochHeader` 记录请求的非历史状态：调用配置、渲染后的系统提示词和工具 schema，空值规范化为缺失。适配器提供的推理强度与 token 默认值会保留其 `adapterDefaults` 来源信息；Web 从日志恢复模型选择时会省略适配器持有的推理强度，因此下一次解析不会把相同的有效配置重新归类为显式选择并产生虚假变更。`request/header` 始终写入完整快照：首个循环实例使用 reason `initial`，后续实例使用 `resume`，实例内变更使用 `change`，内容未变的封装显式开启消息序列或跟随表层替换时使用 `series`。如果发生变化的请求同时开启序列，`change` 快照会携带 `startsSeries: true`，无需重复 header 即可保留这两个独立事实。普通的仅追加后续 Turn、同一序列内后续的 Step 与重试沿用最新快照。`foldRequestHeader` 选择最新快照。旧的 `request/header-delta` 事件和已移除的 `fallback` reason 在追加或加载时都会被拒绝。
 
-每个拟议步骤先领取其 inbox 批次，再运行 `agent/pre-step`。reject 不打开步骤；enter 打开 `step/start`，把最终消息批次记录为 `user/message` 事件，并可使用 `startsRequestSeries: true` 声明独立序列。随后步骤组装系统提示词与工具，`agent/request` 只能替换冻结的调用配置种子。循环记录所需的 initial、resume、change 或 series 完整快照，从派生消息与该 header 构建 `GenerateOptions`，对其深度冻结但保持 `AbortSignal` 活跃。首次调用配置从显式的 `AgentOptions` 出发，保留 fork 覆盖和恢复重配置；后续调用从折叠后的 header 出发。
+每个拟议步骤先领取其 inbox 批次，再运行 `agent/pre-step`。reject 不打开步骤；enter 打开 `step/start`，把最终消息批次记录为 `user/message` 事件，并可使用 `startsRequestSeries: true` 声明独立序列。随后步骤组装系统提示词与工具，`agent/request` 只能替换冻结的调用配置种子。循环记录所需的 initial、resume、change 或 series 完整快照，从派生消息与该 header 构建 `GenerateOptions`，冻结请求但保持 `AbortSignal` 活跃。[请求冻结来源证明决策](../simplification/2026-09-06-agent-request-freeze-provenance.zh.md)拥有消息完整冻结的复用规则和每次请求的本地 header 冻结规则。首次调用配置从显式的 `AgentOptions` 出发，保留 fork 覆盖和恢复重配置；后续调用从折叠后的 header 出发。
 
 **已打开步骤是重建边界。** 进入步骤的 `user/message` 批次与任何新写入的 `request/header` 都位于请求分派之前。原子领取后发生的注入加入后续请求；必须影响本次请求的监听器则通过 `agent/pre-step` 返回消息。header 重建选择该步骤的 `request/header`，或在无新 header 写入时沿用前一个快照。
 

.agents/notes/implemented/architecture/2026-07-06-timeout-deadline-library.i18n.yaml

@@ -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/architecture/2026-07-06-timeout-deadline-library.md
-2026-07-06-timeout-deadline-library.md: 24dcabfd039e553af28fd3cc1d2c7f717ee66509
-2026-07-06-timeout-deadline-library.zh.md: 5f31cf21606345bb323ce3f839b5cb939def5a48
+2026-07-06-timeout-deadline-library.md: 6f726938731f18386105ff1180c79fef981c95bf
+2026-07-06-timeout-deadline-library.zh.md: 48bd7b4b91efd0045c03e3e1e1e9e37656b4dcf6

.agents/notes/implemented/architecture/2026-07-06-timeout-deadline-library.md

@@ -8,11 +8,11 @@ English | [中文](2026-07-06-timeout-deadline-library.zh.md)
 
 Timeout handling was drifting apart across the tool-bearing capabilities, and the divergence was not superficial — it was the same logic re-implemented three ways, each with its own subtle correctness burden.
 
-- **bash** (then in the bash-local implementation's `run.ts`) had a full, correct timeout inside the process plumbing: a config-clamped `timeoutMs`, two independent triggers — a `killTimer` for the timeout and an `onAbort` listener for upstream cancellation — each calling one `kill()` closure that escalates SIGTERM→grace→SIGKILL on the process group, and two orthogonal outcome booleans (`timedOut`, `aborted`) latched independently. After this consolidation, the plumbing — [packages/subprocess/subprocess-local/src/spawn.ts](../../../../packages/subprocess/subprocess-local/src/spawn.ts) — only reacts to aborts; [packages/shell/bash-local/src/index.ts](../../../../packages/shell/bash-local/src/index.ts) owns the fused deadline and the `timedOut`/`aborted` classification.
+- **bash** (then in the bash-local implementation's `run.ts`) had a full, correct timeout inside the process plumbing: a config-clamped `timeoutMs`, two independent triggers — a `killTimer` for the timeout and an `onAbort` listener for upstream cancellation — each calling one `kill()` closure that drove the subprocess termination path, and two orthogonal outcome booleans (`timedOut`, `aborted`) latched independently. After this consolidation, the plumbing — [packages/subprocess/subprocess-local/src/spawn.ts](../../../../packages/subprocess/subprocess-local/src/spawn.ts) — only reacts to aborts; [packages/shell/bash-local/src/index.ts](../../../../packages/shell/bash-local/src/index.ts) owns the fused deadline and the `timedOut`/`aborted` classification.
 - **web_fetch** ([packages/web/web-fetch-http/src/provider.ts](../../../../packages/web/web-fetch-http/src/provider.ts)) had a correct but *hand-rolled* timeout: it constructed an `AbortController`, wired `setTimeout(() => controller.abort(new WebError(…, 'WEB_FETCH_TIMEOUT')))`, manually added and removed the upstream-signal listener, cleared the timer in a `finally`, and recovered the timeout reason from `signal.reason` in a `translateAbortOrNetwork` helper because the reader surfaces a bare `AbortError`.
 - **web_search** ([packages/web/tool-web/src/search.ts](../../../../packages/web/tool-web/src/search.ts)) had **no timeout at all**: `WebSearchRequest` ([packages/web/web/src/types.ts](../../../../packages/web/web/src/types.ts)) carries no `timeoutMs` field, and each provider's `search()` only forwards `exec.signal`. (web_search stays untimed here — see Consequences.)
 
-Each new external-process or network tool re-derived the same four things — clamp the requested value, start a timer, fuse the timeout with upstream cancellation, and distinguish "timed out" from "cancelled" on the way out — and the fusion and reason-recovery are exactly the parts that are easy to get subtly wrong (web_fetch's `signal.reason` dance is evidence). At the same time, the *termination* each performs is irreducibly different: bash kills an OS process group (work runs in a child process, outside this runtime, reachable only by signal), while web aborts an in-process `fetch` (undici tears down the socket). There is no single mechanism that can stop all of them.
+Each new external-process or network tool re-derived the same four things — clamp the requested value, start a timer, fuse the timeout with upstream cancellation, and distinguish "timed out" from "cancelled" on the way out — and the fusion and reason-recovery are exactly the parts that are easy to get subtly wrong (web_fetch's `signal.reason` dance is evidence). At the same time, the *termination* each performs is irreducibly different: bash asks its subprocess provider to terminate an OS-owned range, while web aborts an in-process `fetch` and lets undici tear down the socket. The [native-containment decision](2026-08-28-subprocess-native-containment.md) owns the local scope, Job, and fallback mechanisms; there is no single mechanism that can stop every capability's work.
 
 ## Decision
 
@@ -87,17 +87,17 @@ export function timeoutOf(x: AbortSignal | { reason?: unknown }, code?: string):
 | The default/max *values* | the capability's config |
 | The timeout `code` string | the capability (`WEB_FETCH_TIMEOUT` ≠ `BASH_TIMEOUT`) |
 
-The signal only *notifies*; termination is always the listener's job, and the listener differs by capability. bash writes its own `addEventListener('abort', kill)` because the OS process lives outside this runtime and nothing else will kill it; web hands `d.signal` to `fetch` and undici tears down the socket. This is why file read/write/edit take **no** `timeoutMs`: a local syscall is best-effort-abortable at most, a timeout could not force `fsync`/`rename` to stop, and adding one would be an implicit default that violates explicit-over-implicit. Both reference agents leave file I/O untimed for the same reason.
+The signal only *notifies*; termination is always the listener's job, and the listener differs by capability. bash writes its own `addEventListener('abort', kill)` because the OS process lives outside this runtime and its subprocess provider must drive the owned range to settlement; web hands `d.signal` to `fetch` and undici tears down the socket. This is why file read/write/edit take **no** `timeoutMs`: a local syscall is best-effort-abortable at most, a timeout could not force `fsync`/`rename` to stop, and adding one would be an implicit default that violates explicit-over-implicit. Both reference agents leave file I/O untimed for the same reason.
 
 ### How each capability consumes it
 
 - **web_fetch** — the tool stays validate-and-forward; the provider's hand-rolled controller + `setTimeout` + manual listener + `finally` + `signal.reason` recovery is replaced by provider-owned `deadline`/`timeoutOf`. A pre-aborted upstream signal still throws `WEB_ABORTED` up front; otherwise `fetch` runs against the fused `d.signal`, and `translateAbortOrNetwork` classifies a thrown error by the signal (`timeoutOf` → `WEB_FETCH_TIMEOUT`, else aborted → `WEB_ABORTED`, else network → `WEB_PROVIDER_ERROR`). The public error-code contract is unchanged, and `TimeoutReason` never crosses the web seam as the public error.
-- **bash** — `resolve()` clamps the request into an explicit spec. Foreground `run()` creates the deadline and passes its signal to process execution, whose existing abort listener performs the process-group kill. The executor classifies the first abort as timeout or cancellation. Background starts remain timeout-free and forward only upstream cancellation.
+- **bash** — `resolve()` clamps the request into an explicit spec. Foreground `run()` creates the deadline and passes its signal to process execution, whose abort listener calls `SubprocessHandle.terminate()` and awaits the same provider-managed range. The executor classifies the first abort as timeout or cancellation. Background starts remain timeout-free and forward only upstream cancellation.
 - **LLM adapters** — `dsh-llm-deepseek` and `dsh-llm-pi-ai` wrap actual transport iteration with `idleWatchdog`. The five-minute configured interval covers only outstanding provider demand, not time the downstream consumer spends between chunks. The direct DeepSeek adapter also pulses that outstanding demand when its SSE parser observes a comment, without yielding the comment as a `StreamChunk` or writing it to the session log. The pi-ai SDK does not expose comment activity to its adapter, so that path can rearm only when the SDK yields. The stable signal reaches `fetch` or the SDK for the whole call, so timeout closes the underlying request and maps to `TIMEOUT`, while an earlier caller abort maps to `ABORTED`.
 
 ## Consequences
 
-- `runBash`'s outcome no longer independently latches `timedOut` and `aborted`; a timeout and a user abort racing before process close now report a single first-abort cause instead of both being true. The uniform SIGTERM→grace→SIGKILL kill is unchanged, and the Service Definition type `ShellRunResult` keeps both booleans (now mutually exclusive), so `dsh-tool-bash`'s result rendering is untouched.
+- `runBash`'s outcome no longer independently latches `timedOut` and `aborted`; a timeout and a user abort racing before process close now report a single first-abort cause instead of both being true. Timeout classification does not alter provider-owned termination: local POSIX ranges use TERM→grace→KILL, while Windows ordinary ranges terminate immediately. The Service Definition type `ShellRunResult` keeps both booleans (now mutually exclusive), so `dsh-tool-bash`'s result rendering is untouched.
 - `SpawnSpec.timeoutMs` and `SpawnOutcome.timedOut`/`aborted` were removed rather than kept as always-zero/always-false vestiges: with `runBash` owning no timer and the executor owning classification, they were read nowhere. An always-0 field read by nothing is dead weight under the per-file coverage gate.
 - web_fetch shed its bespoke controller/timer/listener/reason-recovery; the classifier now keys off the deadline signal (`timeoutOf` + `aborted`) rather than the thrown error's shape, which is robust across both the request-phase reject-with-reason and the read-phase bare-`AbortError`.
 - `AbortSignal.any` and `using`/`Symbol.dispose` enter the repo for the first time here (Node ≥ 24 baseline, already met).
@@ -107,10 +107,10 @@ Out of scope, named to mark the boundary: `web_search` can gain an optional mode
 
 ## Alternatives considered
 
-**A unified timeout *plugin* / `ctx.timeout` service.** Rejected on microkernel grounds. A service that could stop any tool's work would have to understand every capability's termination mechanism (process-group SIGKILL, socket teardown, syscall-boundary checks) — the "kernel knows too much" the architecture forbids. Codex's `ExecExpiration` is scoped to the exec family precisely because the kill it drives (`killpg`) is process-family-specific; MCP and model-stream keep their own. There is no coherent middle layer that owns termination for everything, so the shared piece can only be the pure timing/classification half — a library, not a service.
+**A unified timeout *plugin* / `ctx.timeout` service.** Rejected on microkernel grounds. A service that could stop any tool's work would have to understand every capability's termination mechanism (native scope or Job termination, fallback process-group signalling, socket teardown, syscall-boundary checks) — the "kernel knows too much" the architecture forbids. Codex's `ExecExpiration` is scoped to the exec family precisely because the kill it drives (`killpg`) is process-family-specific; MCP and model-stream keep their own. There is no coherent middle layer that owns termination for everything, so the shared piece can only be the pure timing/classification half — a library, not a service.
 
 **Per-tool ad-hoc timeout, no shared code (the prior status quo, and Claude Code's choice).** Rejected because it was already producing divergence and duplicated correctness burden: web_fetch hand-rolled the exact controller/reason logic that future network/process-backed tools would each have to re-derive, and the fusion + `signal.reason` recovery are the error-prone parts. Claude Code tolerates full duplication; this repo has a single shared abort channel (`exec.signal` on every `execute`) that makes a small shared primitive strictly cleaner, so the cost/benefit differs.
 
 **A `withTimeout(promise, ms)` wrapper instead of a signal factory.** Rejected because racing a promise against a timer resolves the *tool-call* promise on deadline without stopping the underlying work — the child process or fetch socket leaks on. Handing out a signal and requiring the capability to listen is what forces a real termination path to exist. This mirrors the "dispose must reach quiescence, not just request it" defensive rule.
 
-**Keep separate bash timeout and cancellation triggers.** Rejected because one deadline signal removes the bespoke timer and standardizes classification. Racing causes report whichever abort arrived first, while the existing SIGTERM-to-SIGKILL termination path remains unchanged.
+**Keep separate bash timeout and cancellation triggers.** Rejected because one deadline signal removes the bespoke timer and standardizes classification. Racing causes report whichever abort arrived first, while the provider-owned termination path is independent of which cause won.

.agents/notes/implemented/architecture/2026-07-06-timeout-deadline-library.zh.md

@@ -8,11 +8,11 @@ Status: implemented
 
 超时处理在各个承载工具的能力之间逐渐分化，而且这种分化并非表面的：同一套逻辑被以三种方式重新实现，各自带有微妙的正确性负担。
 
-- **bash**（当时位于 bash-local 实现的 `run.ts`）在进程管道内部有一套完整、正确的超时实现：一个经配置钳位的 `timeoutMs`，两个独立触发器（用于超时的 `killTimer` 和用于上游取消的 `onAbort` 监听器），各自调用同一个 `kill()` 闭包对进程组执行 SIGTERM→宽限期→SIGKILL 升级，以及两个正交的结果布尔值（`timedOut`、`aborted`）独立锁存。经此次整合之后，这套管道——位于 [packages/subprocess/subprocess-local/src/spawn.ts](../../../../packages/subprocess/subprocess-local/src/spawn.ts)——只响应中止；[packages/shell/bash-local/src/index.ts](../../../../packages/shell/bash-local/src/index.ts) 拥有融合的 deadline 以及 `timedOut`/`aborted` 分类。
+- **bash**（当时位于 bash-local 实现的 `run.ts`）在进程管道内部有一套完整、正确的超时实现：一个经配置钳位的 `timeoutMs`，两个独立触发器（用于超时的 `killTimer` 和用于上游取消的 `onAbort` 监听器），各自调用同一个 `kill()` 闭包驱动子进程终止路径，以及两个正交的结果布尔值（`timedOut`、`aborted`）独立锁存。经此次整合之后，这套管道——位于 [packages/subprocess/subprocess-local/src/spawn.ts](../../../../packages/subprocess/subprocess-local/src/spawn.ts)——只响应中止；[packages/shell/bash-local/src/index.ts](../../../../packages/shell/bash-local/src/index.ts) 拥有融合的 deadline 以及 `timedOut`/`aborted` 分类。
 - **web_fetch**（[packages/web/web-fetch-http/src/provider.ts](../../../../packages/web/web-fetch-http/src/provider.ts)）有一套正确但*手写*的超时：构造一个 `AbortController`，连接 `setTimeout(() => controller.abort(new WebError(…, 'WEB_FETCH_TIMEOUT')))`，手动添加和移除上游信号监听器，在 `finally` 中清除定时器，并在 `translateAbortOrNetwork` 辅助函数中从 `signal.reason` 恢复超时原因（因为 reader 只抛出裸 `AbortError`）。
 - **web_search**（[packages/web/tool-web/src/search.ts](../../../../packages/web/tool-web/src/search.ts)）**完全没有超时**：`WebSearchRequest`（[packages/web/web/src/types.ts](../../../../packages/web/web/src/types.ts)）不携带 `timeoutMs` 字段，各提供方的 `search()` 只转发 `exec.signal`。（web_search 在本次设计中保持无超时——见「后果」。）
 
-每个新的外部进程或网络工具都要重新推导同样四件事：钳位请求值、启动定时器、将超时与上游取消融合、在出口处区分「超时」与「已取消」。而融合与原因恢复恰恰是最容易出微妙错误的部分（web_fetch 的 `signal.reason` 处理就是证据）。与此同时，各能力执行的*终止*操作不可归约地不同：bash 杀死一个 OS 进程组（工作运行在子进程中，在本运行时之外，只能通过信号触达），而 web 中止一个进程内的 `fetch`（undici 拆除 socket）。不存在一个能停止所有能力工作的单一机制。
+每个新的外部进程或网络工具都要重新推导同样四件事：钳位请求值、启动定时器、将超时与上游取消融合、在出口处区分「超时」与「已取消」。而融合与原因恢复恰恰是最容易出微妙错误的部分（web_fetch 的 `signal.reason` 处理就是证据）。与此同时，各能力执行的*终止*操作不可归约地不同：bash 请求其子进程提供方终止由 OS 拥有的范围，而 web 中止一个进程内的 `fetch`，由 undici 拆除 socket。[原生 containment 决策](2026-08-28-subprocess-native-containment.zh.md)负责本地 scope、Job 与 fallback 机制；不存在一个能停止所有能力工作的单一机制。
 
 ## 决策
 
@@ -87,17 +87,17 @@ export function timeoutOf(x: AbortSignal | { reason?: unknown }, code?: string):
 | 默认值/最大值*数值* | 各能力的配置 |
 | 超时 `code` 字符串 | 各能力（`WEB_FETCH_TIMEOUT` ≠ `BASH_TIMEOUT`） |
 
-信号只*通知*；终止始终是监听方的职责，而监听方因能力而异。bash 自行编写 `addEventListener('abort', kill)`，因为 OS 进程存在于本运行时之外，没有别的东西会杀死它；web 将 `d.signal` 交给 `fetch`，由 undici 拆除 socket。这也是文件读/写/编辑**不接受** `timeoutMs` 的原因：本地系统调用最多只能尽力中止，超时无法强制 `fsync`/`rename` 停止，添加超时将是一个违反「显式优于隐式」的隐式默认值。两个参考 agent（智能体）出于同样的原因对文件 I/O 不设超时。
+信号只*通知*；终止始终是监听方的职责，而监听方因能力而异。bash 自行编写 `addEventListener('abort', kill)`，因为 OS 进程存在于本运行时之外，必须由子进程提供方驱动其拥有的范围达到完全停稳；web 将 `d.signal` 交给 `fetch`，由 undici 拆除 socket。这也是文件读/写/编辑**不接受** `timeoutMs` 的原因：本地系统调用最多只能尽力中止，超时无法强制 `fsync`/`rename` 停止，添加超时将是一个违反「显式优于隐式」的隐式默认值。两个参考 agent（智能体）出于同样的原因对文件 I/O 不设超时。
 
 ### 各能力如何消费该库
 
 - **web_fetch**：工具层保持校验并转发；提供方手写的 controller + `setTimeout` + 手动监听器 + `finally` + `signal.reason` 恢复被替换为提供方自有的 `deadline`/`timeoutOf`。已预先中止的上游信号仍然立即抛出 `WEB_ABORTED`；否则 `fetch` 使用融合后的 `d.signal` 运行，`translateAbortOrNetwork` 根据信号分类抛出的错误（`timeoutOf` → `WEB_FETCH_TIMEOUT`，否则已中止 → `WEB_ABORTED`，否则网络错误 → `WEB_PROVIDER_ERROR`）。公开的错误码约定不变，`TimeoutReason` 永远不会作为公开错误跨越 web seam。
-- **bash**：`resolve()` 将请求钳位为显式规格。前台 `run()` 创建 deadline 并将其信号传给进程执行，后者既有的 abort 监听器执行进程组 kill。执行器将首个 abort 分类为超时或取消。后台启动保持无超时，仅转发上游取消。
+- **bash**：`resolve()` 将请求钳位为显式规格。前台 `run()` 创建 deadline 并将其信号传给进程执行，后者的 abort 监听器调用 `SubprocessHandle.terminate()`，并等待同一个由提供方管理的范围。执行器将首个 abort 分类为超时或取消。后台启动保持无超时，仅转发上游取消。
 - **LLM（大语言模型）适配器**：`dsh-llm-deepseek` 和 `dsh-llm-pi-ai` 用 `idleWatchdog` 包装实际的传输迭代。配置的五分钟间隔只覆盖尚未结算的提供方 demand，不包括下游消费方在分片之间花费的时间。DeepSeek 直连适配器还会在其 SSE（Server-Sent Events）解析器观察到注释时，对该项尚未结算的 demand 调用 `pulse()`；该注释既不会作为 `StreamChunk` 产出，也不会写入会话日志。pi-ai SDK 不会向其适配器暴露注释活动，因此该路径只能在 SDK 产出值时重新启动定时器。稳定信号在整个调用期间传给 `fetch` 或 SDK，因此超时会关闭底层请求并映射为 `TIMEOUT`，而更早的调用方中止映射为 `ABORTED`。
 
 ## 后果
 
-- `runBash` 的结果不再独立锁存 `timedOut` 和 `aborted`；超时与用户中止在进程关闭前竞争时，现在报告单一的首个 abort 原因，而非两者同时为 true。统一的 SIGTERM→宽限期→SIGKILL 终止路径不变，Service Definition 类型 `ShellRunResult` 保留两个布尔值（现在互斥），因此 `dsh-tool-bash` 的结果渲染不受影响。
+- `runBash` 的结果不再独立锁存 `timedOut` 和 `aborted`；超时与用户中止在进程关闭前竞争时，现在报告单一的首个 abort 原因，而非两者同时为 true。超时分类不改变由提供方管理的终止：本地 POSIX 范围使用 TERM→宽限期→KILL，Windows 普通范围则立即终止。Service Definition 类型 `ShellRunResult` 保留两个布尔值（现在互斥），因此 `dsh-tool-bash` 的结果渲染不受影响。
 - `SpawnSpec.timeoutMs` 和 `SpawnOutcome.timedOut`/`aborted` 被移除，而非作为始终为零/始终为 false 的残余保留：由于 `runBash` 不再拥有定时器且执行器负责分类，这些字段无处被读取。一个始终为 0 且无处读取的字段在逐文件覆盖率门禁下属于死代码。
 - web_fetch 去除了其定制的 controller/timer/listener/reason-recovery；分类器现在基于 deadline 信号（`timeoutOf` + `aborted`）而非抛出错误的形状来判断，这在请求阶段的 reject-with-reason 和读取阶段的裸 `AbortError` 两种情况下都是健壮的。
 - `AbortSignal.any` 和 `using`/`Symbol.dispose` 在此首次进入本仓库（Node ≥ 24 基线，已满足）。
@@ -107,10 +107,10 @@ export function timeoutOf(x: AbortSignal | { reason?: unknown }, code?: string):
 
 ## 曾考虑的替代方案
 
-**统一的超时*插件* / `ctx.timeout` 服务。** 基于微内核原则否决。一个能停止任何工具工作的服务必须理解每个能力的终止机制（进程组 SIGKILL、socket 拆除、系统调用边界检查），这正是架构所禁止的「内核知道太多」。Codex 的 `ExecExpiration` 被限定于 exec 族，正是因为它驱动的 kill（`killpg`）是进程族特有的；MCP 和模型流各自保有自己的。不存在一个连贯的中间层能为所有东西拥有终止权，因此共享部分只能是纯计时/分类那一半——一个库，而非服务。
+**统一的超时*插件* / `ctx.timeout` 服务。** 基于微内核原则否决。一个能停止任何工具工作的服务必须理解每个能力的终止机制（原生 scope 或 Job 终止、fallback 进程组信号、socket 拆除、系统调用边界检查），这正是架构所禁止的「内核知道太多」。Codex 的 `ExecExpiration` 被限定于 exec 族，正是因为它驱动的 kill（`killpg`）是进程族特有的；MCP 和模型流各自保有自己的。不存在一个连贯的中间层能为所有东西拥有终止权，因此共享部分只能是纯计时/分类那一半——一个库，而非服务。
 
 **每个工具各自实现超时，不共享代码（先前的现状，也是 Claude Code 的选择）。** 否决，因为它已经在产生分化和重复的正确性负担：web_fetch 手写了与未来网络/进程类工具各自需要重新推导的完全相同的 controller/reason 逻辑，而融合 + `signal.reason` 恢复正是容易出错的部分。Claude Code 容忍完全重复；本仓库有一个统一的共享 abort 通道（每次 `execute` 上的 `exec.signal`），使得采用一个小型共享原语明显更简洁，因此成本/收益不同。
 
 **用 `withTimeout(promise, ms)` 包装器代替信号工厂。** 否决，因为让 promise 与定时器竞争只是在截止时间到达时 resolve *工具调用*的 promise，而不会停止底层工作——子进程或 fetch socket 会泄漏。分发信号并要求能力监听，才能强制一条真实的终止路径存在。这与「dispose 必须达到完全停稳，而非仅仅请求它」的防御性规则一致。
 
-**保留 bash 独立的超时和取消触发器。** 否决，因为一个 deadline 信号移除了定制定时器并标准化了分类。发生竞争时，报告先到达的那个 abort 作为原因，而既有的 SIGTERM→SIGKILL 终止路径保持不变。
+**保留 bash 独立的超时和取消触发器。** 否决，因为一个 deadline 信号移除了定制定时器并标准化了分类。发生竞争时，报告先到达的那个 abort 作为原因，由提供方管理的终止路径不受哪个原因先胜出的影响。

.agents/notes/implemented/architecture/2026-07-08-agent-scope-contexts.i18n.yaml

@@ -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/architecture/2026-07-08-agent-scope-contexts.md
-2026-07-08-agent-scope-contexts.md: eb3f6f247bac1a1d81aa2644132c7b9cc04d602c
-2026-07-08-agent-scope-contexts.zh.md: a0f4ffb0ef80dd2fc1ee61c9ab3f4730c28c687e
+2026-07-08-agent-scope-contexts.md: 6a1fd4aed49cb8edef061c8fb6f0edcd0a09c30f
+2026-07-08-agent-scope-contexts.zh.md: 8408c4afff6075c129c6a96c47393c9c812b04b7

.agents/notes/implemented/architecture/2026-07-08-agent-scope-contexts.md

@@ -66,7 +66,7 @@ const handle = await ctx.agents.create({
   agentOptions: { model: 'model-name' },
   setup(agentCtx) {
     agentCtx.systemPrompt.section({
-      name: 'deployment:persona',
+      name: 'deployment:persona-prefix',
       order: 0,
       text: 'Review code, but do not modify files.',
     })

.agents/notes/implemented/architecture/2026-07-08-agent-scope-contexts.zh.md

@@ -66,7 +66,7 @@ const handle = await ctx.agents.create({
   agentOptions: { model: 'model-name' },
   setup(agentCtx) {
     agentCtx.systemPrompt.section({
-      name: 'deployment:persona',
+      name: 'deployment:persona-prefix',
       order: 0,
       text: 'Review code, but do not modify files.',
     })

.agents/notes/implemented/architecture/2026-07-08-tool-output-spill-files.i18n.yaml

@@ -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/architecture/2026-07-08-tool-output-spill-files.md
-2026-07-08-tool-output-spill-files.md: 4e18e9887ab636d31dfe40b1cd38e078e4b30c23
-2026-07-08-tool-output-spill-files.zh.md: b900c168d81253189c7229cc77d6400c614aa89c
+2026-07-08-tool-output-spill-files.md: a80b6cdcb0e731a687d511d38ea288a68a298173
+2026-07-08-tool-output-spill-files.zh.md: a995eafc7878b342a2164c5116da1f43ada791a5

.agents/notes/implemented/architecture/2026-07-08-tool-output-spill-files.md

@@ -22,7 +22,7 @@ A thin spill storage seam plus a default spill policy plugin, in a new `packages
 | `@deepseek-ai/dsh-spill-local` | Local backend: private, session-scoped file storage on the host filesystem. |
 | `@deepseek-ai/dsh-spill-policy` | Tool-result policy plugin: wraps final text results after dispatch and replaces oversized results with a retained preview plus a spill locator. |
 
-There is no dedicated model-facing Consumer package. The Consumer is the existing `ctx.tools` execution pipeline: `dsh-spill-policy` consumes final tool results through the `tools/post-execute` waterfall, and the model follows the backend-supplied retrieval hint for the returned locator.
+The tool-result Consumer is `dsh-spill-policy`, which consumes final tool results through the `tools/post-execute` waterfall. The model follows the backend-supplied retrieval hint for the returned locator. [Session-reference spill reuse](../bug-fix/2026-09-05-session-reference-spill-reuse.md) adds a direct storage consumer with separate preview, provenance, and failure semantics; it does not change the tool-result policy.
 
 ### Spill seam
 
@@ -33,10 +33,15 @@ interface SpillStore {
   saveText(input: SaveTextSpill): Promise<SpillRef>
 }
 
-interface SpillSource {
+type SpillSource = {
+  kind: 'tool'
   toolName: string
   callId: ToolCallId
   label: string
+} | {
+  kind: 'session-reference'
+  sessionId: SessionId
+  label: string
 }
 
 interface SaveTextSpill {

.agents/notes/implemented/architecture/2026-07-08-tool-output-spill-files.zh.md

@@ -22,7 +22,7 @@ Status: implemented
 | `@deepseek-ai/dsh-spill-local` | 本地后端：在宿主文件系统中提供私有、会话作用域的文件存储。 |
 | `@deepseek-ai/dsh-spill-policy` | 工具结果策略插件：包装分发后的最终文本结果，并以保留预览和 spill 定位符替换超大结果。 |
 
-系统不增加专用的面向模型消费方包。消费方是现有 `ctx.tools` 执行流水线：`dsh-spill-policy` 通过 `tools/post-execute` waterfall（瀑布式事件）使用最终工具结果，模型则按照后端随定位符返回的检索提示读取内容。
+工具结果消费方是 `dsh-spill-policy`，它通过 `tools/post-execute` waterfall（瀑布式事件）使用最终工具结果。模型按照后端随定位符返回的检索提示读取内容。[会话引用 spill 复用](../bug-fix/2026-09-05-session-reference-spill-reuse.zh.md)增加一个直接存储消费方，采用独立的预览、来源信息与失败语义；它不改变工具结果策略。
 
 ### spill seam
 
@@ -33,10 +33,15 @@ interface SpillStore {
   saveText(input: SaveTextSpill): Promise<SpillRef>
 }
 
-interface SpillSource {
+type SpillSource = {
+  kind: 'tool'
   toolName: string
   callId: ToolCallId
   label: string
+} | {
+  kind: 'session-reference'
+  sessionId: SessionId
+  label: string
 }
 
 interface SaveTextSpill {

.agents/notes/implemented/architecture/2026-07-10-single-file-executable-sdk-runtime-distribution.i18n.yaml

@@ -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/architecture/2026-07-10-single-file-executable-sdk-runtime-distribution.md
-2026-07-10-single-file-executable-sdk-runtime-distribution.md: f13cd35989e5ef71ca336a5f84bd7e460ab0e3ac
-2026-07-10-single-file-executable-sdk-runtime-distribution.zh.md: 294a3a3c90f033772c0e023fbf18f428ef7ece5b
+2026-07-10-single-file-executable-sdk-runtime-distribution.md: 7792e24a5869be6b7bae7787a6a481f3075ba740
+2026-07-10-single-file-executable-sdk-runtime-distribution.zh.md: 857bfaec80da962fac0403f2239e0a5e71954194

.agents/notes/implemented/architecture/2026-07-16-explicit-turn-cancellation.i18n.yaml

@@ -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/architecture/2026-07-16-explicit-turn-cancellation.md
-2026-07-16-explicit-turn-cancellation.md: ee4253ba607d3468c05b61d6fb0fbc5efcf9f9d7
-2026-07-16-explicit-turn-cancellation.zh.md: d98d2c114056c44f131a2fe989f0b1f544127976
+2026-07-16-explicit-turn-cancellation.md: 09631349226837d7f93112fe5e9c517f2fee099c
+2026-07-16-explicit-turn-cancellation.zh.md: baf039d49578895d53ffae4c6b88d29b3d1563b6

.agents/notes/implemented/architecture/2026-07-16-explicit-turn-cancellation.md

@@ -20,7 +20,7 @@ AgentLoop privately owns one `TurnCancellation` per prospective turn. It install
 
 The driver keeps only a cause-less pre-run marker for queued work cancelled before a turn is claimed. An effective `cancel()` emits the observe-only `agent/cancel-requested` notification with its resolved typed cause before clearing queued and steering work or aborting the holder; notification failures cannot veto the stop, and an idle call emits nothing. Work synchronously queued by a notification observer is included in that clear, while work queued by a later signal abort observer is latched and runs when the aborted activity converges to idle — a `disposed` cancel leaves it parked ([cancel-convergence wake latch](../bug-fix/2026-08-07-cancel-convergence-wake-latch.md)). If a `running` listener synchronously cancels old work and sends a replacement, the driver discards the aborted holder and creates a fresh one for the replacement. Repeated cancellation is first-wins for the active holder, while later calls may still clear newly queued pending work.
 
-The explicit event signatures pass a single payload object: agent-scoped events carry `agent` and `signal` in the payload with `next` last, and the remaining APIs keep `signal` immediately before a waterfall's final `next`. `PreStepContext` and `RequestFailureContext` are retired, with their fields folded into the `agent/pre-step` and `agent/request-error` payloads ([payload-object events](../../archived/architecture/2026-08-06-agent-event-payload-objects.md)). Pre-step entry, request configuration, request-error recovery, model generation, tool execution, approval, turn stopping, and subagent or workflow requests all receive the current signal. Hook bridges must also supply `RunHookOptions.signal`, so a turn cancellation reaches the bash executor's process-group kill and join boundary. `SystemPrompt.assemble()` carries `signal?: AbortSignal` in `AssembleContext` because that object is an explicit request value that can also represent signal-less assembly outside a turn. Listeners may cooperate with the signal but must not retain it to control another turn.
+The explicit event signatures pass a single payload object: agent-scoped events carry `agent` and `signal` in the payload with `next` last, and the remaining APIs keep `signal` immediately before a waterfall's final `next`. `PreStepContext` and `RequestFailureContext` are retired, with their fields folded into the `agent/pre-step` and `agent/request-error` payloads ([payload-object events](../../archived/architecture/2026-08-06-agent-event-payload-objects.md)). Pre-step entry, request configuration, request-error recovery, model generation, tool execution, approval, turn stopping, and subagent or workflow requests all receive the current signal. Hook bridges must also supply `RunHookOptions.signal`, so a turn cancellation reaches the bash executor's provider-managed termination and join boundary; the [native-containment decision](2026-08-28-subprocess-native-containment.md) owns the supported local scope, Job, and fallback mechanics. `SystemPrompt.assemble()` carries `signal?: AbortSignal` in `AssembleContext` because that object is an explicit request value that can also represent signal-less assembly outside a turn. Listeners may cooperate with the signal but must not retain it to control another turn.
 
 `ctx.agents` continues to carry only the initiating Agent. Ambient Agent presence does not imply liveness, a current turn, or cancellation authority. The cause reader is private to the loop and states the machine-private slot invariant (only `cancel()` aborts a turn controller, always with a canonical frozen cause) instead of re-validating the reason structurally; no public helper reads a cause off an arbitrary signal. Concurrent Agents isolate both their initiator identities and their turn signals; a child driver shadows the parent initiator while its parent request signal still travels through the subagent seam.
 

.agents/notes/implemented/architecture/2026-07-16-explicit-turn-cancellation.zh.md

@@ -20,7 +20,7 @@ AgentLoop 为每个待启动轮次私有地持有一个 `TurnCancellation`。它
 
 对于轮次被认领前已取消的排队工作，驱动器只保留一个不携带取消原因的运行前标记。实际生效的 `cancel()` 会先发出仅供观察的 `agent/cancel-requested` 通知并携带最终确定的类型化取消原因，然后才清除排队工作和 steering（中途引导）工作或中止持有者；通知失败不能阻止此次停止，空闲状态下调用则不发出任何通知。通知观察者同步加入队列的工作也会被这次清除，而稍后由 signal 中止观察者加入队列的工作会被锁存，并在被中止的活动收敛到空闲时执行——`disposed` 取消则将其停放（[取消收敛窗口唤醒锁存](../bug-fix/2026-08-07-cancel-convergence-wake-latch.zh.md)）。若 `running` 监听器同步取消旧工作并发送替代提示词，驱动器会丢弃已中止的持有者，并为替代提示词创建全新的持有者。同一活跃持有者上的重复取消遵循首次请求优先，后续调用仍可清除新入队的待处理工作。
 
-显式事件签名传递单个 payload 对象：agent 作用域事件在 payload 中携带 `agent` 和 `signal`，`next` 位于最后；其余 API 保持 `signal` 紧邻 waterfall（瀑布式事件）的最终 `next` 之前。`PreStepContext` 与 `RequestFailureContext` 已退役，其字段并入 `agent/pre-step` 与 `agent/request-error` 的 payload（[payload-object 事件](../../archived/architecture/2026-08-06-agent-event-payload-objects.md)）。进入 pre-step 时、请求配置、请求错误恢复、模型生成、工具执行、审批、轮次停止以及 subagent 或工作流请求都会收到当前 signal。钩子桥接器也必须提供 `RunHookOptions.signal`，使轮次取消能够到达 Bash 执行器终止进程组并等待其退出的边界。`SystemPrompt.assemble()` 在 `AssembleContext` 中携带 `signal?: AbortSignal`，因为该对象是显式请求值，也可表示轮次之外不携带 signal 的组装。监听器可以配合该 signal 取消，但不得保留它来控制其他轮次。
+显式事件签名传递单个 payload 对象：agent 作用域事件在 payload 中携带 `agent` 和 `signal`，`next` 位于最后；其余 API 保持 `signal` 紧邻 waterfall（瀑布式事件）的最终 `next` 之前。`PreStepContext` 与 `RequestFailureContext` 已退役，其字段并入 `agent/pre-step` 与 `agent/request-error` 的 payload（[payload-object 事件](../../archived/architecture/2026-08-06-agent-event-payload-objects.md)）。进入 pre-step 时、请求配置、请求错误恢复、模型生成、工具执行、审批、轮次停止以及 subagent 或工作流请求都会收到当前 signal。钩子桥接器也必须提供 `RunHookOptions.signal`，使轮次取消能够到达 Bash 执行器由提供方管理的终止与等待边界；[原生 containment 决策](2026-08-28-subprocess-native-containment.zh.md)负责受支持本地路径上的 scope、Job 与 fallback 机制。`SystemPrompt.assemble()` 在 `AssembleContext` 中携带 `signal?: AbortSignal`，因为该对象是显式请求值，也可表示轮次之外不携带 signal 的组装。监听器可以配合该 signal 取消，但不得保留它来控制其他轮次。
 
 `ctx.agents` 仍只携带发起 Agent。环境中的 Agent 并不代表存活、当前轮次或取消权限。cause 读取器是 loop 私有的，它直接陈述机器私有的 slot 不变量（只有 `cancel()` 会中止轮次控制器，且总是携带规范的冻结 cause），而不是对 reason 做结构化再校验；不存在从任意 signal 读取 cause 的公开辅助函数。并发 Agent 会同时隔离各自的发起方身份和轮次 signal；子驱动会遮蔽父发起方，而父请求 signal 仍通过 subagent seam 传递。
 

.agents/notes/implemented/architecture/2026-07-28-portable-execution-world-consumers.i18n.yaml

@@ -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/architecture/2026-07-28-portable-execution-world-consumers.md
-2026-07-28-portable-execution-world-consumers.md: 760005f1460890222c5af1ea0ec4eaf9cb29f360
-2026-07-28-portable-execution-world-consumers.zh.md: d78b04a5af785be0707d06e519819f95c524d92d
+2026-07-28-portable-execution-world-consumers.md: 787fe341e58cc212c99e0f35f07eea8e83daf000
+2026-07-28-portable-execution-world-consumers.zh.md: a558a5af64437b8743e741ace4ccf27079501721

.agents/notes/implemented/architecture/2026-07-28-portable-execution-world-consumers.md

@@ -18,13 +18,13 @@ Ordinary pipes do not cover one requirement. A persistent terminal needs PTY all
 
 The filesystem interface owns the path facts that another capability needs without exposing its opaque target identity: a canonical process path, canonical `file:` URI, and containment. Existing whole and streaming text operations remain filesystem-owned; protocol consumers enforce their own retention limits while consuming the stream.
 
-The subprocess interface owns executable lookup and process primitives: ordinary raw or collected process spawning and `spawnTerminal()`. The terminal operation is one deep primitive whose handle owns text I/O, foreground groups, signalling, and one awaited TERM-to-KILL operation that settles in-flight handle calls and reaches quiescence for every session member the provider can still observe. Its signal cancels allocation only; the published handle owns its lifetime. Prompt detection, idle inference, scrollback, sandbox policy, and owner lifecycle remain in the PTY consumer.
+The subprocess interface owns executable lookup and process primitives: ordinary raw or collected process spawning and `spawnTerminal()`. An ordinary handle keeps target identity private: `.done` reports the direct target, while `terminate()` and `waitForExit()` control and observe the same provider-managed range. The [native-containment decision](2026-08-28-subprocess-native-containment.md) owns local Linux scopes, Windows Jobs, and their disclosed fallbacks. The terminal operation is one deep primitive whose handle owns text I/O, foreground groups, signalling, and one awaited TERM-to-KILL operation that settles in-flight handle calls and reaches quiescence for every member of its provider-owned range; an observational fallback limits that range to identities it can still observe. Its signal cancels allocation only; the published handle owns its lifetime. Prompt detection, idle inference, scrollback, sandbox policy, and owner lifecycle remain in the PTY consumer.
 
 Generic consumers use that execution world:
 
 - `dsh-bash-local` continues to map Bash semantics onto ordinary `ctx.subprocess.spawn()`.
 - `dsh-lsp-stdio` reads and contains source through `ctx.fs`, resolves and launches language servers through `ctx.subprocess`, and carries provider-owned file URIs through initialization and result rendering. One provider-lifetime signal aborts filesystem and protocol work during disposal, including workspace lookup before queue ownership; its JSON-RPC, pooling, synchronization, and normalization stay unchanged.
-- `dsh-terminal-bash` maps persistent-shell semantics onto `ctx.subprocess.spawnTerminal()`. The local `node-pty` and process-inspection implementation moves into `dsh-subprocess-local`; another subprocess provider supplies the same primitive. `danger-full-access` needs no `ctx.sandbox`; a confined mode requires a same-world sandbox provider and fails before spawn when none is mounted. Prompt and silence evidence collected during asynchronous pre-write inspection is discarded when the provider write begins. Cancellation retains the send reservation while an in-flight write settles and then signals the foreground group, so late bytes or the signal cannot target a successor; an in-flight readiness poll cannot release that reservation, and a rejected write sends no signal. The absolute deadline remains armed throughout cancellation. A signal failure becomes terminal transport failure. Completion of a stale inspection resumes polling for the current send. Startup cancellation begins terminal rollback without waiting for a stalled readiness or signalling call. Close rejects new public signals and delegates provider-observable session quiescence to the handle's awaited termination operation.
+- `dsh-terminal-bash` maps persistent-shell semantics onto `ctx.subprocess.spawnTerminal()`. The local `node-pty` and process-inspection implementation moves into `dsh-subprocess-local`; another subprocess provider supplies the same primitive. `danger-full-access` needs no `ctx.sandbox`; a confined mode requires a same-world sandbox provider and fails before spawn when none is mounted. Prompt and silence evidence collected during asynchronous pre-write inspection is discarded when the provider write begins. Cancellation retains the send reservation while an in-flight write settles and then signals the foreground group, so late bytes or the signal cannot target a successor; an in-flight readiness poll cannot release that reservation, and a rejected write sends no signal. The absolute deadline remains armed throughout cancellation. A signal failure becomes terminal transport failure. Completion of a stale inspection resumes polling for the current send. Startup cancellation begins terminal rollback without waiting for a stalled readiness or signalling call. Close rejects new public signals and delegates provider-managed session quiescence to the handle's awaited termination operation.
 
 ## E2B POC boundary
 
@@ -68,6 +68,6 @@ A remote execution provider implements only its shared sandbox owner plus filesy
 
 The fundamental interfaces are wider, and a filesystem/subprocess pair must agree on one execution world. The added operations are limited to facts and lifecycle mechanics that current generic consumers require; model schemas, protocol framing, readiness policy, and presentation do not leak into the providers.
 
-The local implementation absorbs `node-pty` and platform process inspection because it owns local terminal mechanics. This moves code without weakening terminal teardown: disposal sweeps descendants before and after terminating the top-level shell, waits for exact PID-identity-fenced descendants retained during foreground inspection, and retains Linux session members that survive top-level exit. macOS cannot enumerate a POSIX session after its leader exits, so a child that reparents between inspection snapshots remains an explicit local-provider limitation rather than a reason to move process mechanics back into the PTY consumer.
+The local implementation absorbs `node-pty` and platform process inspection because it owns local terminal mechanics. On supported Linux hosts, the user-systemd scope retains descendants that call `setsid` or reparent, while process inspection continues to own foreground attribution and synchronous fallback evidence. Other hosts use the observational teardown: disposal sweeps descendants before and after terminating the top-level shell, waits for exact PID-identity-fenced descendants retained during foreground inspection, and retains Linux session members that survive top-level exit. macOS cannot enumerate a POSIX session after its leader exits, so a child that reparents between inspection snapshots remains an explicit local-provider limitation rather than a reason to move process mechanics back into the PTY consumer.
 
 The E2B composition demonstrates that a shared sandbox owner plus filesystem and subprocess adapters are sufficient to move the mutable coding world off-host while leaving higher capabilities provider-neutral. Its POC limits remain explicit: the SDK retains complete command transport in host memory, remote startup cannot publish a PID synchronously, exact terminal stdin-wait and independent signal facts are unavailable, numeric PID/PGID operations are not identity-fenced, the initial environment probe cannot hide unknown sandbox-default secrets from already-running same-UID processes, and adapter artifacts remain until sandbox deletion. These are provider constraints, not justification for compatibility shims or more E2B packages.

.agents/notes/implemented/architecture/2026-07-28-portable-execution-world-consumers.zh.md

@@ -18,13 +18,13 @@ Status: implemented
 
 文件系统接口负责其他能力需要的路径事实，同时不公开其不透明目标身份：规范化进程路径、规范化 `file:` URI 和包含关系。现有完整文本与流式文本操作仍归文件系统负责；协议消费方在消费流时执行各自的保留上限。
 
-进程管理接口负责可执行文件查找与进程原语：以原始或收集模式 spawn 普通进程，以及 `spawnTerminal()`。终端操作是一项深层原语，其句柄负责文本 I/O、前台进程组、信号发送，以及一项须等待的 TERM→KILL 操作；该操作会结算所有在途句柄调用，并使提供方仍可观察到的每个会话成员完全停稳。其信号只取消分配；句柄一经发布，便负责自身生命周期。提示符检测、空闲推断、scrollback、沙箱策略和所有者生命周期仍由 PTY 消费方负责。
+进程管理接口负责可执行文件查找与进程原语：以原始或收集模式 spawn 普通进程，以及 `spawnTerminal()`。普通句柄把 target identity 保持为私有事实：`.done` 报告 direct target，`terminate()` 与 `waitForExit()` 则控制并观察同一个由提供方管理的范围。[原生 containment 决策](2026-08-28-subprocess-native-containment.zh.md)负责本地 Linux scope、Windows Job 及其已声明的 fallback。终端操作是一项深层原语，其句柄负责文本 I/O、前台进程组、信号发送，以及一项须等待的 TERM→KILL 操作；该操作会结算所有在途句柄调用，并使提供方拥有的范围中每个成员完全停稳；观察型 fallback 只能把该范围限制为它仍可观察到的 identity。其信号只取消分配；句柄一经发布，便负责自身生命周期。提示符检测、空闲推断、scrollback、沙箱策略和所有者生命周期仍由 PTY 消费方负责。
 
 通用消费方使用该执行世界：
 
 - `dsh-bash-local` 继续把 Bash 语义映射到普通的 `ctx.subprocess.spawn()`。
 - `dsh-lsp-stdio` 通过 `ctx.fs` 读取源文件并验证包含关系，通过 `ctx.subprocess` 解析和启动语言服务器，并让由提供方负责的文件 URI 贯穿初始化与结果渲染。一个提供方生命周期信号会在资源释放期间中止文件系统与协议操作，包括取得队列所有权之前的工作区查找；其 JSON-RPC、池化、同步和规范化保持不变。
-- `dsh-terminal-bash` 把持久 shell 语义映射到 `ctx.subprocess.spawnTerminal()`。本地 `node-pty` 与进程检查实现移入 `dsh-subprocess-local`；其他进程管理提供方则提供相同原语。`danger-full-access` 不需要 `ctx.sandbox`；受限模式要求同一执行世界中存在沙箱提供方，未挂载时会在 spawn 前失败。提供方开始写入时，系统会丢弃异步写入前检查期间收集的提示符与静默证据。取消会在在途写入结算期间保留发送预留，随后向前台进程组发送信号，因此延迟字节和该信号都无法落到后续发送；在途就绪检查无法释放该预留，写入被拒绝时也不会发送信号。绝对截止时间会在整个取消期间保持启用。信号发送失败会成为终结性传输失败。陈旧检查完成后，会针对当前发送恢复轮询。启动取消会立即开始终端回滚，而不等待停滞的就绪检查或信号发送调用。关闭操作会拒绝新的公开信号，并把提供方可观察会话成员的完全停稳委托给句柄上须等待的终止操作。
+- `dsh-terminal-bash` 把持久 shell 语义映射到 `ctx.subprocess.spawnTerminal()`。本地 `node-pty` 与进程检查实现移入 `dsh-subprocess-local`；其他进程管理提供方则提供相同原语。`danger-full-access` 不需要 `ctx.sandbox`；受限模式要求同一执行世界中存在沙箱提供方，未挂载时会在 spawn 前失败。提供方开始写入时，系统会丢弃异步写入前检查期间收集的提示符与静默证据。取消会在在途写入结算期间保留发送预留，随后向前台进程组发送信号，因此延迟字节和该信号都无法落到后续发送；在途就绪检查无法释放该预留，写入被拒绝时也不会发送信号。绝对截止时间会在整个取消期间保持启用。信号发送失败会成为终结性传输失败。陈旧检查完成后，会针对当前发送恢复轮询。启动取消会立即开始终端回滚，而不等待停滞的就绪检查或信号发送调用。关闭操作会拒绝新的公开信号，并把由提供方管理的会话完全停稳委托给句柄上须等待的终止操作。
 
 ## E2B POC 边界
 
@@ -68,6 +68,6 @@ E2B 负责可变文件系统、受管命令与 Bash 进程、终端分配与终
 
 基础接口更宽，一对文件系统／进程管理提供方必须在同一个执行世界上保持一致。新增操作仅限当前通用消费方所需的事实与生命周期机制；模型 schema、协议分帧、就绪策略和呈现不会渗入提供方。
 
-本地实现承接 `node-pty` 和平台进程检查，因为它负责本地终端机制。这种代码迁移不会削弱终端拆卸：dispose（资源释放）会在终止顶层 shell 前后清理后代进程，等待前台检查期间保留下来且受精确 PID 身份围栏保护的后代进程，并继续追踪在顶层进程退出后仍存活的 Linux 会话成员。macOS 无法在 POSIX 会话 leader 退出后枚举该会话，因此在两次检查快照之间重新设定父进程的子进程仍是明确的本地提供方限制，而不是把进程机制移回 PTY 消费方的理由。
+本地实现承接 `node-pty` 和平台进程检查，因为它负责本地终端机制。在受支持的 Linux 宿主上，user-systemd scope 会保留调用 `setsid` 或发生 reparent 的后代，进程检查则继续负责前台归属与同步 fallback 证据。其他宿主使用观察型拆卸：dispose（资源释放）会在终止顶层 shell 前后清理后代进程，等待前台检查期间保留下来且受精确 PID 身份围栏保护的后代进程，并继续追踪在顶层进程退出后仍存活的 Linux 会话成员。macOS 无法在 POSIX 会话 leader 退出后枚举该会话，因此在两次检查快照之间重新设定父进程的子进程仍是明确的本地提供方限制，而不是把进程机制移回 PTY 消费方的理由。
 
 E2B 组合证明，共享沙箱所有者加上文件系统与进程管理适配器，就足以在保持上层能力与提供方无关的同时，把可变编码世界移出宿主。其 POC 限制仍明确在案：SDK 会把完整命令传输内容保留在宿主内存中；远程启动无法同步发布 PID；无法获得精确的终端 stdin 等待状态与独立信号事实；基于数值 PID/PGID 的操作没有身份围栏；初始环境探测无法向已在运行的同 UID 进程隐藏未知的沙箱默认 secret；适配器产物会一直保留到沙箱删除。这些是提供方限制，不是引入兼容性 shim 或更多 E2B 包的理由。

.agents/notes/implemented/architecture/2026-08-10-session-log-version-mechanism.i18n.yaml

@@ -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/architecture/2026-08-10-session-log-version-mechanism.md
-2026-08-10-session-log-version-mechanism.md: 98eb220e49457d3a2783edefce13d053c9b94025
-2026-08-10-session-log-version-mechanism.zh.md: 8853e6ed9a1a1ed43193cfe0949fffdbb9b4a26f
+2026-08-10-session-log-version-mechanism.md: 0f7f70b5ad6ecb2445729b1aa61fb3b295fc4ddb
+2026-08-10-session-log-version-mechanism.zh.md: a6d58505ad9fdb6068a1afe48f7250f020d20a9e

.agents/notes/implemented/architecture/2026-08-10-session-log-version-mechanism.md

@@ -14,13 +14,13 @@ Session logs must be upgradable after release, and the runtime that ships first
 
 **The writer decides bumps, not the reader.** A bump is required exactly when an old runtime could no longer handle a new log with full semantic correctness. "Parses without error" is not the bar: silently skipping content that shapes reconstruction is a wrong read. Only structural changes qualify — header shape, event envelope, core event semantics, the surface mechanism (`SurfaceEventType` set, `SurfaceOp` variants). When unsure, bump: a near-identity upgrader is almost free, a missed bump silently corrupts old readers.
 
-**Read rules by direction.** Equal version: read normally. Newer than the reader: refuse, name the direction ("written by a newer harness — upgrade"), and point at the raw log artifact so the user can still see the text (`SessionFormatUnsupportedError`, distinct from `SessionPersistenceCorruptionError` because nothing is damaged). Older than the reader: every event-body operation first runs the complete adjacent chain in memory, leaves the source path, bytes, and inode unchanged, exclusively publishes only the final current generation under its canonical versioned filename, and reopens it before current restoration. Header-only listing remains non-mutating and reports the numerically highest canonical generation. Catalog generation and module initialization reject a missing adjacent step, so a published first-party build never exposes a partial historical chain. Retained lower generations are not automatic fallback or a downgrade compatibility promise.
+**Read rules by direction.** Equal version: read normally. Newer than the reader: refuse, name the direction ("written by a newer harness — upgrade"), and point at the raw log artifact so the user can still see the text (`SessionFormatUnsupportedError`, distinct from `SessionPersistenceCorruptionError` because nothing is damaged). Older than the reader: every event-body operation first runs the complete adjacent chain in memory and leaves the source path, bytes, and inode unchanged. Read handles may consume that current logical result directly; a write open exclusively publishes the final current generation under its canonical versioned filename before append. Header-only listing remains non-mutating and reports the numerically highest canonical generation. Catalog generation and module initialization reject a missing adjacent step, so a published first-party build never exposes a partial historical chain. Retained lower generations are not automatic fallback or a downgrade compatibility promise.
 
 **A per-event `ignorable` marker covers vocabulary growth, so ordinary event additions never bump the version.** The event vocabulary is decided by which plugins are mounted, which a single version integer cannot describe. A reader meeting an unrecognized event type refuses to interpret the log unless the event carries `ignorable: true` in its envelope. The default is *required*: forgetting the marker over-refuses a resumable session (an inconvenience), while a default of ignorable would make the same mistake silently resume a gutted one (a safety failure). The architecture makes this sound: model-visible content flows only through the three `surfaceOp`-marked surface event types plus the `request/header`/`request/context` folds, so the dangerous unknowns are exactly the non-surface events that change how the rest of the log is read (`session/end-seed` is the existing example).
 
 ## Consequences
 
-What shipped in v0 (release 0812): direction-aware refusal with the raw-log path; the unknown-event guard against a generated known-vocabulary list (`KNOWN_SESSION_EVENT_TYPES`, emitted by `gen-persistence-catalog` from every `SessionEventMap` merge and kept fresh by `verify-persistence-catalog`); the `ignorable` envelope field accepted by seed validation, JSONL, and the BFF wire schema. V1 adds the static adjacent catalog, the identity v0-to-v1 edge, header-only descriptors, exact-generation JSONL publication, and current-only restoration described in [Released Session formats](2026-08-31-released-session-format-migrations.md). V2 keeps the physical codec neutral to ordinary event vocabulary and payload additions: the adjacent edge freezes its released source and target inventories, while equal-version restoration applies the installed known-event set and current payload semantics. First-party writers do not set `ignorable` through `Session.append`, while a repository-external plugin is a current consumer; equal-version retention lives in the [external-plugin retention decision](2026-08-30-retain-ignorable-external-session-events.md), and the stricter historical rule lives in the [alpha migration refusal decision](2026-08-31-alpha-historical-unknown-event-refusal.md). The unknown-type guard remains read-side because append-time vocabulary refusal would stall a live session's durability. JSONL classifies foreign versions from the minimal raw header before current-header or event parsing, so a structurally different future format reports the upgrade direction instead of "corrupt".
+What shipped in v0 (release 0812): direction-aware refusal with the raw-log path; the unknown-event guard against a generated known-vocabulary list (`KNOWN_SESSION_EVENT_TYPES`, emitted by `gen-persistence-catalog` from every `SessionEventMap` merge and kept fresh by `verify-persistence-catalog`); the `ignorable` envelope field accepted by seed validation, JSONL, and the BFF wire schema. V1 adds the static adjacent catalog, the identity v0-to-v1 edge, header-only descriptors, exact-generation JSONL publication, and current-only restoration described in [Released Session formats](2026-08-31-released-session-format-migrations.md). [Historical Session read preparation](2026-09-05-read-only-session-migration-preparation.md) owns the JSONL timing between in-memory restoration and write publication. V2 keeps the physical codec neutral to ordinary event vocabulary and payload additions: the adjacent edge freezes its released source and target inventories, while equal-version restoration applies the installed known-event set and current payload semantics. First-party writers do not set `ignorable` through `Session.append`, while a repository-external plugin is a current consumer; equal-version retention lives in the [external-plugin retention decision](2026-08-30-retain-ignorable-external-session-events.md), and the stricter historical rule lives in the [alpha migration refusal decision](2026-08-31-alpha-historical-unknown-event-refusal.md). The unknown-type guard remains read-side because append-time vocabulary refusal would stall a live session's durability. JSONL classifies foreign versions from the minimal raw header before current-header or event parsing, so a structurally different future format reports the upgrade direction instead of "corrupt".
 
 ## Alternatives considered
 

.agents/notes/implemented/architecture/2026-08-10-session-log-version-mechanism.zh.md

@@ -14,13 +14,13 @@ Session log 在发布后必须能升级格式，而最先发布的运行时决
 
 **升不升版本由写入方决定，与读取方能力无关。**当且仅当老运行时无法在语义上完全正确地处理新日志时才必须升版本。"解析不报错"不是标准：静默跳过影响重建的内容就是读错。只有结构性变更够得上这条线：header 形状、事件信封、核心事件语义、surface 机制（`SurfaceEventType` 集合、`SurfaceOp` 变体）。拿不准就升：近似恒等的升级器几乎没有成本，漏升一次会让老读取器静默读坏。
 
-**读取规则按方向区分。**版本相等：正常读。比读取器新：拒绝，说明方向（"由更新的 harness 写入，请升级"），并给出原始日志文件的路径，用户仍能看到文本（`SessionFormatUnsupportedError`，与 `SessionPersistenceCorruptionError` 区分，因为数据没有损坏）。比读取器旧：每个事件正文操作先在内存中运行完整相邻链，保持源路径、字节与 inode 不变，只在规范具名版本文件下排他发布最终当前 generation，再在当前恢复前重新打开。仅 header 的列表保持不变更，并报告数值最高的规范 generation。catalog 生成与模块初始化会拒绝缺失的相邻步骤，因此已发布第一方 build 绝不会暴露不完整历史链。保留的低 generation 不是自动 fallback，也不构成 downgrade compatibility 承诺。
+**读取规则按方向区分。**版本相等：正常读。比读取器新：拒绝，说明方向（"由更新的 harness 写入，请升级"），并给出原始日志文件的路径，用户仍能看到文本（`SessionFormatUnsupportedError`，与 `SessionPersistenceCorruptionError` 区分，因为数据没有损坏）。比读取器旧：每个事件正文操作先在内存中运行完整相邻链，并保持源路径、字节与 inode 不变。读句柄可以直接使用该 current 逻辑结果；写 open 则在 append 前把最终 current generation 排他发布到其规范版本文件名。仅 header 的列表保持不变更，并报告数值最高的规范 generation。catalog 生成与模块初始化会拒绝缺失的相邻步骤，因此已发布第一方 build 绝不会暴露不完整历史链。保留的低 generation 不是自动 fallback，也不构成 downgrade compatibility 承诺。
 
 **逐事件的 `ignorable` 标记吸收词汇表增长，普通的新增事件永远不用升版本。**事件词汇表由挂载了哪些插件决定，单个版本整数描述不了它。读取器遇到不认识的事件类型时拒绝解读日志，除非该事件的信封带 `ignorable: true`。默认为必需：忘写标记的后果是把一个本可恢复的会话拒绝过头（体验问题），而默认可忽略会让同样的疏忽静默恢复出残缺会话（安全事故）。架构保证了这条规则成立：模型可见内容只经三种带 `surfaceOp` 标记的 surface 事件加 `request/header`、`request/context` 折叠进入重建，危险的未知事件恰好是那些不进 surface 但改变日志其余部分解读方式的事件（`session/end-seed` 是现存例子）。
 
 ## 影响
 
-v0（0812 发布）交付的内容：分方向的拒绝并带原始日志路径；基于生成的已知词汇清单（`KNOWN_SESSION_EVENT_TYPES`，由 `gen-persistence-catalog` 从所有 `SessionEventMap` 声明合并生成，`verify-persistence-catalog` 保证新鲜）的未知事件守卫；`ignorable` 信封字段被种子校验、JSONL 和 BFF 线上 schema 接受。V1 添加静态相邻 catalog、恒等 v0-to-v1 迁移边、仅 header descriptor、精确代际 JSONL 发布与[已发布 Session 格式](2026-08-31-released-session-format-migrations.zh.md)定义的当前专用恢复。V2 让物理 codec 对普通事件词汇与 payload 新增项保持中立：相邻迁移边冻结 released source 与 target 清单，同版本恢复则应用已安装的 known-event set 与当前 payload 语义。第一方 writer 不通过 `Session.append` 设置 `ignorable`，而一个仓库外插件仍依赖该字段；同版本保留由[外部插件保留决策](2026-08-30-retain-ignorable-external-session-events.zh.md)定义，更严格的历史规则由 [alpha 迁移拒绝决策](2026-08-31-alpha-historical-unknown-event-refusal.zh.md)定义。未知类型守卫仍只在读取侧生效，因为 append 时的词汇拒绝会中断活跃 Session 的持久化。JSONL 会在当前 header 或事件解析前从最小原始 header 分类外来版本，因此结构完全不同的未来格式会报告升级方向而不是"损坏"。
+v0（0812 发布）交付的内容：分方向的拒绝并带原始日志路径；基于生成的已知词汇清单（`KNOWN_SESSION_EVENT_TYPES`，由 `gen-persistence-catalog` 从所有 `SessionEventMap` 声明合并生成，`verify-persistence-catalog` 保证新鲜）的未知事件守卫；`ignorable` 信封字段被种子校验、JSONL 和 BFF 线上 schema 接受。V1 添加静态相邻 catalog、恒等 v0-to-v1 迁移边、仅 header descriptor、精确代际 JSONL 发布与[已发布 Session 格式](2026-08-31-released-session-format-migrations.zh.md)定义的当前专用恢复。[历史 Session 只读迁移准备](2026-09-05-read-only-session-migration-preparation.zh.md)负责内存恢复与写入发布之间的 JSONL 时序。V2 让物理 codec 对普通事件词汇与 payload 新增项保持中立：相邻迁移边冻结 released source 与 target 清单，同版本恢复则应用已安装的 known-event set 与当前 payload 语义。第一方 writer 不通过 `Session.append` 设置 `ignorable`，而一个仓库外插件仍依赖该字段；同版本保留由[外部插件保留决策](2026-08-30-retain-ignorable-external-session-events.zh.md)定义，更严格的历史规则由 [alpha 迁移拒绝决策](2026-08-31-alpha-historical-unknown-event-refusal.zh.md)定义。未知类型守卫仍只在读取侧生效，因为 append 时的词汇拒绝会中断活跃 Session 的持久化。JSONL 会在当前 header 或事件解析前从最小原始 header 分类外来版本，因此结构完全不同的未来格式会报告升级方向而不是"损坏"。
 
 ## 曾考虑的替代方案
 

.agents/notes/implemented/architecture/2026-08-22-single-dsh-application-launcher.i18n.yaml

@@ -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/architecture/2026-08-22-single-dsh-application-launcher.md
-2026-08-22-single-dsh-application-launcher.md: 21fc99790c506f4e3f6e8b3ab572121e1cf1a882
-2026-08-22-single-dsh-application-launcher.zh.md: 4175bd675889ec01db46ebca157aff35c9ef9cab
+2026-08-22-single-dsh-application-launcher.md: 630040c75c4c20c57b5e26288663265174331ca5
+2026-08-22-single-dsh-application-launcher.zh.md: 9bd51e6c4e8b7ab60cbabc384431bc4e22a6b522

.agents/notes/implemented/architecture/2026-08-22-single-dsh-application-launcher.md

@@ -46,7 +46,7 @@ Direct SDK use follows normal Harness-home resolution: explicit `dshHome`, inher
 
 ### Python runtime
 
-The Python runtime wheel packages the ordinary `@deepseek-ai/dsh` CLI from `node_modules/@deepseek-ai/dsh/lib/bin.js` through the private `dsh-python-runtime-closure` deploy manifest. The Python client selects `dsh --profile sdk` by default, ordered patch files, and an explicit Harness home; the runnable example under `python/sdk/examples` selects `sdk-minimal`. The installed `dsh` console command exposes the same profile grammar and the separately packaged `web` application.
+The Python runtime wheel stages [`python/sdk-runtime/runtime-bootstrap.mjs`](../../../../python/sdk-runtime/runtime-bootstrap.mjs) as the `dsh-python-runtime-closure` entry. Its ordinary branch calls the public CLI export; a provider-private selector dispatches to the internal subprocess runner before CLI parsing and is not an application entry point. The [native-containment decision](2026-08-28-subprocess-native-containment.md) owns that private dispatch. The Python client selects `dsh --profile sdk` by default, ordered patch files, and an explicit Harness home; the runnable example under `python/sdk/examples` selects `sdk-minimal`. The installed `dsh` console command exposes the same profile grammar and the separately packaged `web` application.
 
 The executable family is `deepseek-harness-sdk-runtime-<platform>-<arch>`. The SDK wire, wheel and import distribution names, sidecar names, and wire identity `deepseek-harness-sdk-runtime` remain stable. The SDK package family is `@deepseek-ai/dsh-sdk-client`, `@deepseek-ai/dsh-sdk-protocol`, and `@deepseek-ai/dsh-sdk-jsonrpc-server`; `@deepseek-ai/dsh-acp` remains the ACP protocol plugin. There is no Python-specific Node application, checked-in complete config, compatibility package, forwarding executable, fallback parser, or SDK/ACP launcher alias. [docs/architecture.md](../../../../docs/architecture.md) owns this launch, and the [`python/sdk-runtime` README](../../../../python/sdk-runtime/README.md) owns the Windows carrier.
 

.agents/notes/implemented/architecture/2026-08-22-single-dsh-application-launcher.zh.md

@@ -46,7 +46,7 @@ SDK 用户通过 profile 自定义插件。`dsh plugin --profile <name> ...` 管
 
 ### Python 运行时
 
-Python 运行时 wheel 通过私有 `dsh-python-runtime-closure` 部署 manifest，打包来自 `node_modules/@deepseek-ai/dsh/lib/bin.js` 的普通 `@deepseek-ai/dsh` CLI。Python 客户端默认选择 `dsh --profile sdk`、有序 patch 文件与显式 Harness home；`python/sdk/examples` 下的可运行示例选择 `sdk-minimal`。安装的 `dsh` 控制台命令暴露相同 profile 语法与单独打包的 `web` 应用。
+Python 运行时 wheel 将 [`python/sdk-runtime/runtime-bootstrap.mjs`](../../../../python/sdk-runtime/runtime-bootstrap.mjs) 暂存为 `dsh-python-runtime-closure` 入口。其普通分支调用公开 CLI export；提供方私有选择会在 CLI 解析前分派到内部子进程 runner，而不是应用入口。[原生 containment 决策](2026-08-28-subprocess-native-containment.zh.md)负责该私有分派。Python 客户端默认选择 `dsh --profile sdk`、有序 patch 文件与显式 Harness home；`python/sdk/examples` 下的可运行示例选择 `sdk-minimal`。安装的 `dsh` 控制台命令暴露相同 profile 语法与单独打包的 `web` 应用。
 
 可执行文件族是 `deepseek-harness-sdk-runtime-<platform>-<arch>`。SDK 协议格式、wheel 与 import 分发名称、伴随文件名称，以及协议 identity `deepseek-harness-sdk-runtime` 保持稳定。SDK 包族是 `@deepseek-ai/dsh-sdk-client`、`@deepseek-ai/dsh-sdk-protocol` 与 `@deepseek-ai/dsh-sdk-jsonrpc-server`；`@deepseek-ai/dsh-acp` 继续作为 ACP 协议插件。仓库不保留 Python 专用 Node 应用、检入的完整配置、兼容包、转发可执行文件、后备解析器或 SDK／ACP 启动别名。[docs/architecture.md](../../../../docs/architecture.zh.md)负责该启动方式，[`python/sdk-runtime` README](../../../../python/sdk-runtime/README.zh.md)负责 Windows 载体。
 

.agents/notes/implemented/architecture/2026-08-25-session-observations-and-projection-owned-client-state.i18n.yaml

@@ -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/architecture/2026-08-25-session-observations-and-projection-owned-client-state.md
-2026-08-25-session-observations-and-projection-owned-client-state.md: 71c0aa92ddb2ad5bacb238dcc7f6e44e33f84305
-2026-08-25-session-observations-and-projection-owned-client-state.zh.md: f8d4a629501dd354e1210bccab1a3a3862b8c6c3
+2026-08-25-session-observations-and-projection-owned-client-state.md: ff5628c3051086159d532b71ca3e98c1b41a51f8
+2026-08-25-session-observations-and-projection-owned-client-state.zh.md: e433df28fec86564fe1c1e9e88686a5abe6b9549

.agents/notes/implemented/architecture/2026-08-25-session-observations-and-projection-owned-client-state.md

@@ -51,7 +51,7 @@ Every owner disposes its observation. `retain()` creates another lease over the
 
 ### Source resolution and lifetime
 
-An observation binds all returned fields to one lifecycle witness. Callers do not combine a header from corpus listing, events from persistence, and projections from a later live Session. The selected header and event prefix produce the cursor and projection snapshot together.
+An observation binds all returned fields to one lifecycle witness. Callers do not combine a header from corpus listing, events from persistence, and projections from a later live Session. The selected header and event prefix produce the cursor and projection snapshot together. A live observation fixes its cut as the log length at read time and materializes `events` on the first access; the log only appends, so that prefix is identical however late a consumer reads it, and a consumer that needs only the header, cursor, or projections never copies the log.
 
 Live preference is checked both before and after a cold borrow. The second check closes the race in which an Agent attaches while persistence is loading. If persistence itself reports that a live source won but that source has already detached by the time SessionQuery examines it, resolution restarts instead of publishing an unowned reference.
 

.agents/notes/implemented/architecture/2026-08-25-session-observations-and-projection-owned-client-state.zh.md

@@ -51,7 +51,7 @@ flowchart LR
 
 ### 数据源解析与生命周期
 
-一份 observation 把所有返回字段绑定到同一 lifecycle witness。调用方不会把 corpus list 的 header、persistence 的 events 和稍后 live Session 的 projections 拼在一起。选中的 header 与事件前缀共同产生 cursor 和 projection snapshot。
+一份 observation 把所有返回字段绑定到同一 lifecycle witness。调用方不会把 corpus list 的 header、persistence 的 events 和稍后 live Session 的 projections 拼在一起。选中的 header 与事件前缀共同产生 cursor 和 projection snapshot。live observation 在读取时以日志长度固定 cut，并在首次访问时才物化 `events`；日志只会追加，所以无论消费者多晚读取，该前缀都完全相同，而只需要 header、cursor 或 projections 的消费者永远不会复制日志。
 
 系统在 cold borrow 前后都检查 live 优先级。第二次检查封住 persistence 加载期间 Agent 完成 attach 的竞态。如果 persistence 报告由 live source 胜出，但 SessionQuery 检查时该 source 已经 detach，解析会重新开始，而不是发布一份无人持有的引用。
 

.agents/notes/implemented/architecture/2026-08-28-subprocess-native-containment.i18n.yaml

@@ -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/architecture/2026-08-28-subprocess-native-containment.md
+2026-08-28-subprocess-native-containment.md: 0c03884bba67dab6e2e38f96ce2e874ed62ba00f
+2026-08-28-subprocess-native-containment.zh.md: b7850e1c4fee06dfeb1a05c968d132de5994686f

.agents/notes/implemented/architecture/2026-08-28-subprocess-native-containment.md

@@ -0,0 +1,83 @@
+# Agent Note: Native owners contain escaped subprocess descendants
+
+Status: implemented
+
+English | [中文](2026-08-28-subprocess-native-containment.zh.md)
+
+## Problem
+
+Detached POSIX process groups, Windows direct-parent traversal, and PTY descendant scans describe only members that remain observable through one process relationship. A child can call `setsid`, reparent, or outlive its direct parent and leave those ranges, so terminating the apparent tree can return while work, ports, or files remain active. A direct target result also does not prove that every descendant has stopped.
+
+The ordinary subprocess handle cannot solve this gap by publishing a PID. Linux scope setup and the Windows Job runner establish target identity asynchronously, a PID does not name the complete managed range, and consumers would be forced to infer whether startup committed. The public result, range ownership, private startup protocol, and packaged entry therefore need separate owners.
+
+## Decision
+
+`LocalSubprocessRuntime` selects one provider-private managed-range owner before a target can execute. Eligible Linux ordinary and PTY launches enter a transient user-systemd scope; eligible Windows ordinary launches enter an unnamed kill-on-close Job owned by a private runner. Unsupported hosts use the existing weaker fallback with one provider-lifetime warning. The provider never replays a target after a selected native path may have executed it.
+
+An ordinary `SubprocessHandle` has no PID or public startup state. `.done` reports the direct target result or startup/provider failure, `terminate()` signals the selected range, and `waitForExit()` succeeds only after that same range is proven empty. `SubprocessTerminalHandle.pid` remains part of the terminal contract because PTY identity and foreground inspection require it.
+
+### Linux scope and one-shot bootstrap
+
+The first eligible Linux ordinary or PTY call in one runtime deeply checks the exact runner entry, the libc `execve` and `fcntl` bindings, the readable user manager, and literal-argv transient-scope support. Failed deep probes are retried, while the first success is cached. Each later eligible call still performs a lightweight user-manager reachability probe before target execution. Once selected, a scope, protocol, state-query, or pre-exec failure is reported through that launch and never switches to fallback.
+
+The parent creates one 0700 directory with a complete 0600 `launch-request.json` containing the final target cwd and environment. The private `DSH_SUBPROCESS_RUNNER` value locates that request while the runner starts from the provider cwd and a bootstrap-safe environment. `systemd-run --user --scope --quiet --collect --expand-environment=no` registers its process in the scope, then the one-shot bootstrap removes and validates the request, changes to the target cwd, restores the complete target environment, resolves a bare executable with the target PATH rules, clears `FD_CLOEXEC` on fd 0 through fd 2, and calls libc `execve()` with the original argv. The bootstrap becomes the target in place and preserves its inherited stdio; it does not remain as a supervisor.
+
+Request consumption or a manager observation of a loaded unit establishes scope ownership. Unit absence before either fact remains unresolved while the direct launcher is running. If that launcher exits while the request remains unconsumed, the direct result rejects with the startup failure while range observation records that the scope never existed and resolves the empty-range wait. The parent checks this unresolved interval every 50 milliseconds; after establishment, state queries back off exponentially to the existing 5-second systemctl bound. Each query reads both `LoadState` and `ActiveState`: loaded `inactive` or `failed`, or an established unit becoming `not-found`/`inactive` or otherwise collected away, proves the range empty. `active`, `activating`, `reloading`, and `deactivating` remain nonterminal. Unknown or malformed combinations and unreadable manager results reject `waitForExit()` instead of claiming quiescence. `terminate()` wakes a sleeping observer for an immediate recheck, and settlement cancels the losing backoff sleep. A strict sibling `startup-error.json` carries only request/bootstrap or target pre-exec failure, and the parent removes this spawn's private paths at observable lifecycle completion.
+
+The ordinary target result still comes from the same child process. The PTY path uses the same request and bootstrap without a resident runner, so the `node-pty` PID, process group, session leader, controlling terminal, foreground `inputWaiting`, `/dev/tty`, readiness, and direct terminal outcome retain their existing meanings while scope membership covers `setsid` and reparented descendants.
+
+### Windows runner and Job
+
+The Windows parent starts the provider runner from a bootstrap cwd and environment, gives it the original target argv after a private `--` delimiter, and waits for Node's runner `spawn` event before sending exactly one start request. A pre-spawn runner error preserves that direct launch failure while proving that no Job range existed, so the empty-range wait succeeds; post-spawn infrastructure failure remains uncertain and rejects range settlement. Node IPC otherwise carries idempotent terminate control and exactly one result. Runner fd 0 through fd 2 are isolated, fd 3 carries IPC, and fd 4 through fd 6 carry target stdin, stdout, and stderr. Fd 4 inherits the platform null-device descriptor when stdin is ignored and uses a pipe otherwise. The shared Win32 layer maps fds 4 through 6 to OS handles through Node's exported `uv_get_osfhandle()`, rejects null plus the unsigned `UV_INVALID_OS_FILE_HANDLE` and `UV_INVALID_FILE_DESCRIPTOR` sentinels exposed by Koffi, temporarily enables inheritance on valid handles, and passes them through `STARTF_USESTDHANDLES`. `spawnCurrentTokenJobProcess` requires a separately resolved `applicationName` and a complete target environment, which it sends as a sorted, double-NUL-terminated UTF-16LE block with `CREATE_UNICODE_ENVIRONMENT`, including `=X:` drive entries, without mutating the runner environment. After the suspended target enters the Job and resumes, the runner closes only fd 4 through fd 6; it never mutates or destroys Node's standard streams. The parent returns the pipe carriers as the ordinary handle's stdio, and user bytes never pass through IPC.
+
+The runner is the sole owner of the target process handle and unnamed Job handle. `spawnCurrentTokenJobProcess` creates the target suspended, assigns it to a kill-on-close Job that disallows active breakaway, and resumes it only after assignment. The runner polls the direct process for the target exit code and the Job for active-process count. It exits successfully only after the direct result has been delivered through the IPC send callback and the Job has reported zero active processes; the parent maps only that clean exit to successful `waitForExit()`.
+
+The parent permanently latches a validated numeric `target-exit` as soon as it arrives, before the existing stdout/stderr close or bounded-drain barrier settles. `.done` waits only for that stdio barrier and then returns the latched result. A later Job query, range-settlement failure, IPC loss, or abnormal runner exit rejects only `waitForExit()` and cannot replace the direct result. Infrastructure failure before a valid target result rejects `.done`, but that rejection does not reveal whether target execution began. On disconnect or result-send failure, the runner stops protocol work, terminates and closes its only Job handle, and exits nonzero. Closing the last Job handle kills remaining members but does not convert the disconnected path into a successful quiescence proof.
+
+### Private dispatch and protocol
+
+Source launches execute the package runner entry through the TypeScript source launcher, built launches resolve the `@deepseek-ai/dsh-subprocess-local/runner` export, and the Python SDK single-file executable enters through [`python/sdk-runtime/runtime-bootstrap.mjs`](../../../../python/sdk-runtime/runtime-bootstrap.mjs). The Python runtime owns that bootstrap. It calls the public CLI export when the private selector is absent; otherwise it removes the selector and dispatches to the same subprocess runner core. The public `dsh` argument parser has no hidden runner mode, `apps/cli` builds only its public entry, and packaging ships no second Node executable.
+
+The selector is a per-spawn locator or sentinel, not a credential or persistent format. Linux uses one strict request plus one optional strict startup-error file. Windows uses one IPC channel with closed `start` and `terminate` requests and exactly two result branches: `target-exit` with a numeric `exitCode`, and `error` with required `name` and `message` plus only optional `code`, `syscall`, and `path`; the parent derives `signal: null`. Pre-commit cancellation uses the same ordinary `error` record. The cancellation reason never crosses the wire, so a parent cancellation latch restores its first local reason exactly, including `null` or `undefined`. Missing, extra, mistyped, or unknown fields fail closed. Target environments may contain the selector name, including Windows case variants, because the provider transmits target state separately and restores it only after private selection is consumed.
+
+### Fallback and cleanup
+
+Linux falls back before target execution when the exact bootstrap, modern readable user-systemd manager, or literal-argv scope is unavailable. Windows ordinary launch falls back when the runner entry, Win32 bindings, or current-token Job probe is unavailable. macOS ordinary launch, Windows ConPTY, and other unsupported hosts retain their existing PGID, `taskkill /T`, or identity-fenced PTY observation. The warning states that descendants escaping those observable relationships are not guaranteed to terminate or delay `waitForExit()`.
+
+Normal Cordis disposal starts direct-result and range observation independently, requests termination, and waits for every owned range. Consumer teardown does not inspect an ordinary PID; it retains the original operation or startup error while attempting terminate and final wait, and preserves cleanup failures in the consumer's existing error order. A confirmed empty range permanently disables later signalling against stale identities.
+
+During a JavaScript-observable host exit, `LocalSubprocessRuntime` synchronously force-terminates every still-live handle without promises or timers. Linux sends the existing direct fallback kill and the exact scope kill; Windows kills the runner so its only Job handle closes; PTY fallback scans remain best effort. Per-handle failures are contained and do not change the host's exit result. Termination modes in which JavaScript cannot run remain outside this listener's guarantee.
+
+## Existing decisions and supersession
+
+This note owns the current native-containment mechanism. It partially updates the provider and no-PID facts in the [subprocess seam](../../archived/architecture/2026-07-26-subprocess-seam.md), the Linux teardown facts in [persistent PTY sessions](../feature/2026-07-16-persistent-pty-sessions.md), the native targets used by [synchronous host-exit cleanup](../../archived/bug-fix/2026-08-11-synchronous-subprocess-exit-cleanup.md), the ordinary consumer of [shared Win32 process primitives](../../archived/architecture/2026-08-19-shared-win32-process-primitives.md), and the private entry selected by the [Python SDK profile runtime](../../archived/architecture/2026-08-23-python-sdk-dsh-profile-runtime.md). Each note retains its other decision and remains active.
+
+## Verification
+
+- Provider and Linux protocol suites pin synchronous NUL rejection before launch side effects, strict request/error decoding, target cwd and complete environment restoration, private-variable collision, symlink-sensitive PATH traversal with preserved argv, close-on-exec removal for inherited stdio, pre-exec error ownership, failed-deep-probe retry plus successful-deep-probe caching with per-call manager checks, the three scope-establishment states including an exited launcher with an unconsumed request, `LoadState`/`ActiveState` parsing, `reloading`, terminate wake-up with losing-delay cancellation, bounded established-scope backoff, and exactly-once PTY managed-owner cleanup.
+- Windows protocol and Win32 suites pin exactly two result branches, numeric-only target exits, ordinary-error start cancellation with raw parent-local reasons, the reduced `name`/`message`/`code`/`syscall`/`path` error record, the fixed `2`/`3`/`267` to `ENOENT`, `740` to `EACCES`, `5` to `EPERM`, `193` to `EFTYPE`, and remaining-code to `UNKNOWN` mapping, start delivery after runner spawn, empty-range settlement after pre-spawn failure, explicit ordinally sorted target environment blocks with `=C:` preservation and double-NUL termination, `uv_get_osfhandle()` carrier mapping and unsigned invalid-sentinel rejection, the null-device ignored-stdin carrier and piped non-ignored stdin, result-send and IPC-disconnect failures, direct-result latching before stdio settlement, active-process quiescence, and unique handle cleanup.
+- Real Linux user-systemd tests run one ordinary and one `node-pty` `setsid`/reparent scenario through the production entry. They prove scope signalling and collection, bare executable lookup, escaped-descendant termination, range settlement, and unchanged PTY PID, session, controlling-terminal, foreground-input, `/dev/tty`, readiness, and startup-failure semantics.
+- Native Windows tests prove suspended creation, Job assignment before resume, inherited stdio, default descendant inheritance, direct result, termination, active-process zero, abnormal/disconnected runner cleanup, kill-on-close, and synchronous host-exit termination. Source, built, and Python packaged smokes enter the same runner core.
+- Public seam types, local and E2B providers, LSP and subagent consumers, shell fixtures, READMEs, the Cordis catalog, and the keyless subprocess API snapshot contain no ordinary PID; terminal PID remains.
+
+## Alternatives considered
+
+**Keep PID, make it optional, or add a public `started` promise.** Rejected because each representation exposes an asynchronous provider identity that does not name the managed range and invites consumers to infer startup or quiescence from the wrong fact.
+
+**Extend process-group, session, parent-tree, or PID scanning.** Rejected because a process can leave each observed relationship; broader SID signalling can also reach unrelated processes when a PTY helper shares a launcher session. Native OS membership is persistent and independently queryable.
+
+**Let the parent own or reopen the Windows Job.** Rejected because copied handles, named Jobs, `OpenJobObject`, process-handle handoff, and completion ports create multiple lifecycle owners without improving the direct-result contract. One runner can own target creation, Job membership, result production, and final handle closure.
+
+**Carry control or results through target stdio or files on Windows.** Rejected because user bytes and EOF must remain authoritative to existing Node streams, while result files or polling introduce a second result owner. One IPC channel separates control from target stdio.
+
+**Parse a hidden runner argument in the public CLI or ship another Node executable.** Rejected because either choice expands the public application grammar or distribution surface. A packaging-only bootstrap keeps one physical executable and two private logical entries.
+
+**Cache the complete Linux native decision without another reachability check.** Rejected because user-manager availability can change between spawns. The runtime caches only the first successful deep bootstrap/scope probe and lightly rechecks manager reachability for every later eligible call.
+
+**Recover a failed native launch by replaying the command.** Rejected because an ambiguous failure may occur after target execution and replay can therefore execute the command twice.
+
+## Consequences
+
+Supported Linux ordinary and PTY launches and Windows ordinary launches retain descendants through process-group escape and direct-parent exit, while direct target results remain independent from range quiescence. The cost is a per-spawn Linux manager check and scope/request or Windows runner/IPC/Job lifecycle, plus explicit failure when the selected owner cannot prove settlement.
+
+Fallback hosts continue to run commands but carry a visible weaker guarantee. Windows ConPTY, macOS native containment, active breakaway descendants, old or absent user-systemd environments, target replay, persistent runner recovery, and termination paths where JavaScript cannot execute remain outside this decision.

.agents/notes/implemented/architecture/2026-08-28-subprocess-native-containment.zh.md

@@ -0,0 +1,83 @@
+# Agent Note: 原生 owner 收容逃逸的 subprocess 后代
+
+Status: implemented
+
+[English](2026-08-28-subprocess-native-containment.md) | 中文
+
+## Problem
+
+detached POSIX 进程组、Windows direct-parent 遍历与 PTY 后代扫描只能描述仍可通过某种进程关系观察到的成员。子进程可以调用 `setsid`、发生 reparent，或比 direct parent 存活更久并离开这些 range，因此终止表面进程树后，工作、端口或文件仍可能保持活跃。direct target result 也不能证明全部后代已经停止。
+
+普通 subprocess 句柄无法通过发布 PID 解决这个缺口。Linux scope setup 与 Windows Job runner 会异步建立 target identity，PID 不表示完整 managed range，消费方也会被迫推断 startup 是否已经提交。因此，公共结果、range ownership、私有启动协议与打包入口需要各自明确的 owner。
+
+## Decision
+
+`LocalSubprocessRuntime` 会在 target 执行前选择一个 provider 私有的 managed-range owner。符合条件的 Linux 普通命令与 PTY 进入临时 user-systemd scope；符合条件的 Windows 普通命令进入由私有 runner 拥有的 unnamed kill-on-close Job。不支持的宿主使用既有较弱 fallback，并在 provider 生命周期内只警告一次。选定的 native 路径一旦可能已经执行 target，provider 绝不重放 target。
+
+普通 `SubprocessHandle` 没有 PID 或公共 startup 状态。`.done` 报告 direct target result 或 startup／provider failure，`terminate()` 向所选 range 发送信号，`waitForExit()` 只有在同一 range 被证明为空后才成功。`SubprocessTerminalHandle.pid` 继续属于终端约定，因为 PTY identity 与前台检查需要它。
+
+### Linux scope 与 one-shot bootstrap
+
+同一 runtime 第一次符合条件的 Linux 普通或 PTY 调用会深入检查准确 runner 入口、libc `execve` 与 `fcntl` bindings、可读的 user manager 与保留 literal argv 的 transient-scope 支持。失败的深度 probe 会重试，第一次成功则会缓存。之后每次符合条件的调用仍会在 target 执行前轻量检查 user manager 是否可达。native 路径一旦选定，scope、协议、状态查询或 pre-exec failure 都由本次启动报告，绝不切换到 fallback。
+
+parent 创建一个 0700 目录，其中的完整 0600 `launch-request.json` 保存最终 target cwd 与环境。私有 `DSH_SUBPROCESS_RUNNER` 值负责定位该 request，runner 则从 provider cwd 与 bootstrap-safe 环境启动。`systemd-run --user --scope --quiet --collect --expand-environment=no` 先把自身进程注册到 scope，再由 one-shot bootstrap 删除并校验 request、切换到 target cwd、恢复完整 target 环境、按 target PATH 规则解析裸可执行文件、清除 fd 0 至 fd 2 的 `FD_CLOEXEC`，并使用原始 argv 调用 libc `execve()`。bootstrap 会原地成为 target 并保留继承的 stdio，不作为常驻 supervisor。
+
+request 被消费或 manager 已观察到 loaded unit 都能建立 scope ownership。在这两项事实出现前，只要 direct launcher 仍在运行，unit absence 就保持未决。如果 launcher 退出时 request 仍未消费，direct result 会以 startup failure reject，而 range observation 会记录 scope 从未存在，并成功结算 empty-range wait。parent 每 50 毫秒检查一次这段未决区间；建立后，状态查询按指数增长间隔退避，最多达到既有的 5 秒 systemctl 上限。每次查询同时读取 `LoadState` 与 `ActiveState`：loaded `inactive` 或 `failed`，以及已经建立的 unit 变为 `not-found`/`inactive` 或被 collect 卸载，都能证明 range 为空。`active`、`activating`、`reloading` 与 `deactivating` 仍是非终态。未知或 malformed 组合以及不可读的 manager 结果会使 `waitForExit()` reject，而不是宣称完全停稳。`terminate()` 会唤醒正在休眠的 observer 立即复查，结算时会取消未胜出的退避 sleep。严格的同目录 `startup-error.json` 只承载 request／bootstrap 或 target pre-exec failure，parent 会在可观察生命周期完成时移除本次 spawn 的私有路径。
+
+普通 target result 仍来自同一个 child process。PTY 路径复用同一 request 与 bootstrap，但不增加常驻 runner，因此 `node-pty` PID、进程组、session leader、控制终端、前台 `inputWaiting`、`/dev/tty`、readiness 与 direct terminal outcome 保留既有含义，同时 scope membership 覆盖 `setsid` 与 reparent 后代。
+
+### Windows runner 与 Job
+
+Windows parent 从 bootstrap cwd 与环境启动 provider runner，把原始 target argv 放在私有 `--` 分隔符之后，并等待 Node 的 runner `spawn` 事件后才发送恰好一条 start request。runner 在 spawn 前报错时，direct launch failure 会原样保留，同时证明 Job range 从未存在，因此 empty-range wait 成功；spawn 后的 infrastructure failure 仍是不确定状态，会使 range settlement reject。除此之外，Node IPC 还承载幂等 terminate control 与恰好一个 result。runner 的 fd 0 至 fd 2 相互隔离，fd 3 承载 IPC，fd 4 至 fd 6 承载 target stdin、stdout 与 stderr。忽略 stdin 时，fd 4 继承平台 null-device descriptor；其他模式使用 pipe。共享 Win32 层通过 Node 导出的 `uv_get_osfhandle()` 把 fd 4 至 fd 6 映射为 OS handle，拒绝 null 以及 Koffi 暴露的 unsigned `UV_INVALID_OS_FILE_HANDLE` 与 `UV_INVALID_FILE_DESCRIPTOR` sentinel，临时启用有效 handle 的继承，并通过 `STARTF_USESTDHANDLES` 传入。`spawnCurrentTokenJobProcess` 要求单独解析的 `applicationName` 与完整 target 环境，并使用 `CREATE_UNICODE_ENVIRONMENT` 传入排序、双 NUL 结尾的 UTF-16LE 块，其中包括 `=X:` 驱动器条目，而不修改 runner 环境。suspended target 进入 Job 并恢复后，runner 只关闭 fd 4 至 fd 6；它绝不改写或销毁 Node 标准流。parent 把 pipe carrier stream 作为普通句柄的 stdio 返回，用户字节绝不经过 IPC。
+
+runner 是 target process handle 与 unnamed Job handle 的唯一 owner。`spawnCurrentTokenJobProcess` 以 suspended 状态创建 target，把它分配给不允许 active breakaway 的 kill-on-close Job，并只在分配后恢复。runner 轮询 direct process 获取 target exit code，并轮询 Job 获取 active-process count。只有 direct result 已通过 IPC send callback 交付且 Job 已报告零 active process 后，runner 才成功退出；parent 只把这次 clean exit 映射成成功的 `waitForExit()`。
+
+parent 会在收到经过校验、只含数字的 `target-exit` 时立即永久锁存它，此时既有 stdout／stderr close 或有界 drain barrier 可能尚未完成。`.done` 只继续等待该 stdio barrier，随后返回已锁存的结果。后续 Job query、range settlement failure、IPC loss 或 runner 异常退出只会使 `waitForExit()` reject，不能替换 direct result。在有效 target result 到达前发生 infrastructure failure 才会使 `.done` reject，但该 rejection 不会说明 target 是否已经开始执行。disconnect 或 result-send failure 会让 runner 停止协议工作、终止并关闭自己唯一的 Job handle，然后以非零状态退出。最后一个 Job handle 关闭会终止剩余成员，但不会把 disconnected 路径改写成成功的完全停稳证明。
+
+### 私有分派与协议
+
+source 启动通过 TypeScript source launcher 执行包内 runner 入口，built 启动解析 `@deepseek-ai/dsh-subprocess-local/runner` export，Python SDK 单文件可执行程序则从 [`python/sdk-runtime/runtime-bootstrap.mjs`](../../../../python/sdk-runtime/runtime-bootstrap.mjs) 进入。该 bootstrap 由 Python runtime 拥有；私有 selector 不存在时，它调用公共 CLI export；否则会删除 selector，并分派到同一 subprocess runner core。公共 `dsh` 参数解析器没有隐藏 runner mode，`apps/cli` 只构建公共入口，打包也不提供第二个 Node 可执行程序。
+
+selector 是 per-spawn locator 或 sentinel，不是凭据或持久格式。Linux 使用一个严格 request 与一个可选严格 startup-error 文件。Windows 使用一条 IPC channel，承载闭集的 `start` 与 `terminate` request，以及恰好两个 result 分支：只含数字 `exitCode` 的 `target-exit`，以及必含 `name`、`message` 且只允许可选 `code`、`syscall`、`path` 的 `error`；parent 会派生 `signal: null`。提交前取消使用同一种普通 `error` record。取消 reason 不跨 wire 传递，因此 parent cancellation latch 会原样恢复第一个本地 reason，包括 `null` 或 `undefined`。缺失、额外、类型错误或未知字段都会 fail closed。target 环境可以包含 selector 名称及其 Windows 大小写变体，因为 provider 会单独传递 target 状态，并且只在私有选择值消费后才恢复该状态。
+
+### Fallback 与 cleanup
+
+准确 bootstrap、现代且可读的 user-systemd manager，或保留 literal argv 的 scope 不可用时，Linux 会在 target 执行前进入 fallback。runner 入口、Win32 bindings 或 current-token Job probe 不可用时，Windows 普通启动会进入 fallback。macOS 普通启动、Windows ConPTY 与其他不受支持的宿主保留既有 PGID、`taskkill /T` 或带身份围栏的 PTY 观察机制。warning 会明确说明：逃离这些可观察关系的后代不保证被终止，也不保证延迟 `waitForExit()`。
+
+正常 Cordis dispose 会独立启动 direct-result 与 range observation、请求终止，并等待每个自有 range。消费方 teardown 不检查普通 PID；它会保留原始 operation 或 startup error，同时尝试 terminate 与 final wait，并按消费方既有错误顺序保留 cleanup failure。range 一旦被确认为空，就会永久禁止后续向陈旧 identity 发送信号。
+
+在 JavaScript 可观察的 host exit 期间，`LocalSubprocessRuntime` 会同步强制终止每个仍存活的句柄，不使用 Promise 或 timer。Linux 会发送既有 direct fallback kill 与准确 scope kill；Windows 会终止 runner，使其唯一 Job handle 关闭；PTY fallback 扫描仍是 best effort。每个句柄的失败相互隔离，也不改变宿主退出结果。JavaScript 无法运行的终止形态不属于该 listener 的保证。
+
+## Existing decisions and supersession
+
+本 Note 拥有当前 native containment 机制。它局部更新了[subprocess seam](../../archived/architecture/2026-07-26-subprocess-seam.md)中的 provider 与 no-PID 事实、[持久化 PTY 会话](../feature/2026-07-16-persistent-pty-sessions.zh.md)中的 Linux teardown 事实、[宿主退出同步清理](../../archived/bug-fix/2026-08-11-synchronous-subprocess-exit-cleanup.md)使用的 native target、[共享 Win32 process primitives](../../archived/architecture/2026-08-19-shared-win32-process-primitives.md)的 ordinary 消费方，以及[Python SDK profile 运行时](../../archived/architecture/2026-08-23-python-sdk-dsh-profile-runtime.md)选择的私有入口。每份 Note 都保留其余决策并继续处于 active 状态。
+
+## Verification
+
+- provider 与 Linux 协议测试套件固定同步 NUL 拒绝发生在启动副作用之前、严格 request／error 解码、target cwd 与完整环境恢复、私有变量碰撞、保留 argv 且对 symlink 敏感的 PATH 遍历、为继承 stdio 清除 close-on-exec、pre-exec error ownership、失败深度 probe 重试与成功深度 probe 缓存及逐调用 manager 检查、三种 scope 建立状态（包括 launcher 退出且 request 未消费）、`LoadState`／`ActiveState` 解析、`reloading`、带未胜出 delay 取消的 terminate wake-up、建立后有上限的退避，以及 PTY managed-owner 恰好一次 cleanup。
+- Windows 协议与 Win32 测试套件固定恰好两个 result 分支、只含数字的 target exit、使用普通 error 的 start cancellation 与 parent 原样保留的本地 reason、缩减到 `name`／`message`／`code`／`syscall`／`path` 的 error record、固定的 `2`／`3`／`267` 到 `ENOENT`、`740` 到 `EACCES`、`5` 到 `EPERM`、`193` 到 `EFTYPE` 及其余 code 到 `UNKNOWN` 的映射、runner spawn 后才发送 start、spawn 前 failure 的 empty-range settlement、按序数显式排序的 target 环境块及 `=C:` 保留和双 NUL 结尾、`uv_get_osfhandle()` carrier 映射与 unsigned invalid sentinel 拒绝、null-device ignored-stdin carrier 与非 ignore stdin pipe、result-send 与 IPC-disconnect failure、stdio settlement 前的 direct-result 锁存、active-process 完全停稳，以及唯一 handle cleanup。
+- 真实 Linux user-systemd 测试会分别通过生产入口运行一条普通命令与一条 `node-pty` `setsid`／reparent 场景。它们证明 scope signalling 与 collection、裸可执行文件查找、逃逸后代终止、range settlement，以及不变的 PTY PID、session、控制终端、前台输入、`/dev/tty`、readiness 与 startup-failure 语义。
+- native Windows 测试证明 suspended creation、resume 前 Job assignment、继承 stdio、默认后代继承、direct result、termination、active-process zero、异常／disconnected runner cleanup、kill-on-close 与同步 host-exit termination。source、built 与 Python packaged 冒烟测试进入同一 runner core。
+- 公共 seam 类型、local 与 E2B provider、LSP 与 subagent 消费方、shell fixture、README、Cordis catalog 与 keyless subprocess API snapshot 都不包含普通 PID；terminal PID 保留。
+
+## Alternatives considered
+
+**保留 PID、把它改成可选值，或增加公共 `started` Promise。**不予采用，因为这些表示都会暴露不表示 managed range 的异步 provider identity，并诱导消费方从错误事实推断 startup 或完全停稳。
+
+**扩展进程组、session、parent tree 或 PID 扫描。**不予采用，因为进程可以离开每一种观察关系；当 PTY helper 与 launcher 共用 session 时，更宽的 SID signalling 还可能命中无关进程。native OS membership 会持续存在，并且可以独立查询。
+
+**让 parent 拥有或重新打开 Windows Job。**不予采用，因为复制 handle、named Job、`OpenJobObject`、process-handle handoff 与 completion port 会制造多个 lifecycle owner，却不能改善 direct-result 约定。一个 runner 可以统一拥有 target creation、Job membership、result production 与最终 handle closure。
+
+**在 Windows 上通过 target stdio 或文件传递 control／result。**不予采用，因为用户字节与 EOF 必须继续以既有 Node stream 为权威，而 result file 或 polling 会制造第二个 result owner。一条 IPC channel 可以把 control 与 target stdio 分开。
+
+**在公共 CLI 中解析隐藏 runner 参数，或发布另一个 Node 可执行程序。**不予采用，因为前者扩张公共应用语法，后者扩张分发面。packaging-only bootstrap 保留一个物理可执行程序与两个私有逻辑入口。
+
+**缓存完整的 Linux native 选择结果且不再检查可达性。**不予采用，因为 user-manager availability 可以在两次 spawn 之间变化。runtime 只缓存第一次成功的 bootstrap／scope 深度 probe，并在之后每次符合条件的调用中轻量复查 manager reachability。
+
+**在 native launch 失败后重放命令。**不予采用，因为含糊 failure 可能发生在 target 已经执行之后，重放因此可能把命令执行两次。
+
+## Consequences
+
+受支持的 Linux 普通与 PTY 启动、Windows 普通启动会在后代逃离进程组或 direct parent 退出后继续拥有它们，同时 direct target result 与 range 完全停稳保持独立。代价是每次 spawn 都需要一次 Linux manager 检查与 scope／request，或一套 Windows runner／IPC／Job 生命周期，而且所选 owner 无法证明 settlement 时会显式失败。
+
+fallback 宿主继续运行命令，但携带可见的较弱保证。Windows ConPTY、macOS native containment、active breakaway 后代、旧版或缺失的 user-systemd 环境、target replay、持久 runner recovery，以及 JavaScript 无法执行的终止路径均不属于本决策。

.agents/notes/implemented/architecture/2026-08-31-released-session-format-migrations.i18n.yaml

@@ -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/architecture/2026-08-31-released-session-format-migrations.md
-2026-08-31-released-session-format-migrations.md: 2c75d0b57a0b513c218b6a67b8c2b31c7cae4d0f
-2026-08-31-released-session-format-migrations.zh.md: d88c643cbfaf7f3d4f52ca6e5fa244a26917eb99
+2026-08-31-released-session-format-migrations.md: 592322c0e4c1b2fa52dcf71652f3878f43a8c8ca
+2026-08-31-released-session-format-migrations.zh.md: ba2317903845739cda8da1c01c2f959c7a2ccd50

.agents/notes/implemented/architecture/2026-08-31-released-session-format-migrations.md

@@ -1,4 +1,4 @@
-# Agent Note: Released Session formats migrate on body read through adjacent pure edges
+# Agent Note: Released Session formats migrate through stateful streaming stages
 
 Status: implemented
 
@@ -6,52 +6,176 @@ English | [中文](2026-08-31-released-session-format-migrations.zh.md)
 
 ## Problem
 
-Session format v0 shipped in an alpha release, so a structural writer change can no longer treat existing JSONL as disposable pre-release state. Stored event bodies reach consumers through read or write `SessionHandle` instances used by resume, query, export, fork, and continuation paths. Migrating only one consumer would let callers observe different logical generations or fail only when a later writer reaches the old file.
+Session format v0 shipped in an alpha release, so a structural writer change can no longer treat existing JSONL as disposable pre-release state. The first whole-artifact migration implementation made those logs convertible, but its data model turned a 116 MB real Session into an operation that exhausted a 16 GB Node process before returning a handle.
 
-Migration must retain the exact source path, bytes, and inode, including a torn physical tail, while giving every published format one unambiguous canonical filename. Plain JSONL and Zstandard are encoding choices for the same logical format and must not create parallel migration implementations.
+### Whole-artifact performance failure
+
+- Zstandard input was split into 317,540 frames and each frame used a separate asynchronous decompression call. The implementation retained every plaintext frame and then concatenated them before JSON parsing, creating the same number of Promise, thread-pool, and native decode transitions.
+- Physical Decode materialized a complete plaintext Buffer, one complete string, every JSONL row, expanded source events, migrated target events, encoded target rows, a joined target string, and target physical Buffers at overlapping points in the same request.
+- Every codec and migration edge called `snapshotSessionFormatJson()` or `snapshotSessionFormatArtifact()`. These operations detached, recursively copied, and deeply froze whole headers, rows, payloads, and event arrays before and after adjacent migrations.
+- Released packed Assistant chunks expanded into about 9.14 million logical v0/v1 events before v1-to-v2 folded them into 72,784 current events. The whole-artifact API required both representations and the old-to-new sequence map to coexist.
+- Encoding built the complete JSONL and compressed output in memory. The successful path then decoded the staged target, decoded the committed target, and decoded it again in persistence to construct the business object; it also reread the source for a full fingerprint comparison.
+- Per-frame `await` calls did not provide useful bounded scheduling. The pre-migration reader instead reused one synchronous decoder and yielded from the outer loop about every 500 ms, avoiding hundreds of thousands of asynchronous transitions.
+
+### The interfaces prevented local fixes from composing
+
+`SessionFormatCodec` decoded and encoded complete arrays, each adjacent `SessionFormatMigration` accepted and returned a complete `SessionFormatArtifact`, and the compiled chain could only hand one materialized artifact to the next edge. A faster physical decoder therefore still encountered source-row arrays, expanded-event arrays, per-edge snapshots, and target-row arrays downstream.
+
+The migrations are stateful even though the API presented them as one-shot functions. v0-to-v1 tracks message and retry identity. v1-to-v2 buffers one unsettled Assistant attempt, tracks events blocked behind it, and maintains old-to-new sequence references. Wrapping that state in closures or push/finish helper objects made the runtime structure different from the static declarations and made production, Worker verification, fixtures, and replay use different entry paths.
 
 ## Decision
 
-`SESSION_FORMAT_VERSION` is a monotonic current-writer integer. One profile-independent pure package owns each adjacent `vN -> vN+1` conversion. `@deepseek-ai/dsh-session-format` supplies only lossless snapshots, unique gap-free planning, header-only conversion, and whole-artifact composition; `@deepseek-ai/dsh-session-format-catalog` statically imports the complete chain independently of mounted Cordis plugins. Historical codecs and normalizers live in the named edge package, while current Session and persistence code accept only the latest logical types.
+The Session format packages use a stateful synchronous Stage API. Static migration declarations describe one adjacent version edge and create a new stage for each restored artifact. A stage owns that artifact's mutable state; no stage instance is shared across Sessions.
 
-Each edge freezes strict source and target semantics, while its target physical codec remains vocabulary-neutral so ordinary event growth can stay within one format version. The catalog restores the final generation through the installed peer `@deepseek-ai/dsh-session` and its current `KNOWN_SESSION_EVENT_TYPES`, preventing a frozen historical edge from becoming the current vocabulary owner.
+### Stage and Context protocol
 
-The JSONL provider completes ensure-current work before `open` returns a handle for a stored Session. It selects the highest canonical generation, migrates a supported historical body, and decodes the current result from one physical snapshot; the public `SessionPersistence` and `SessionHandle` interfaces contain no migration operations. Header-only `stat` and `list` rescan Session directories, translate supported historical headers in memory, and never publish a successor. `create` checks canonical filenames independently of header readability, so every existing generation reserves its Session id.
+```text
+interface SessionFormatMigrationContext {
+  emitEvent(event: SessionFormatEvent): void
+  emitRun(run: SessionFormatEventRun): void
+}
 
-Cancellation belongs to the `open`, `stat`, or `list` call that supplied it. Discovery, stable reads, decoding, and pre-publication checks observe that signal; once an immutable successor is published and its directory entry is synced, later cancellation does not delete the committed generation.
+interface SessionFormatMigrationStage {
+  readonly headerInheritedEventCount?: number
+  transformEvent(
+    event: SessionFormatEvent,
+    context: SessionFormatMigrationContext,
+  ): void
+  transformRun(
+    run: SessionFormatEventRun,
+    context: SessionFormatMigrationContext,
+  ): void
+  finish(context: SessionFormatMigrationContext): number
+}
+```
 
-The configured JSONL encoding owns one full suffix, `.jsonl` or `.jsonl.zstd`. Migration reads a stable exact source, decodes the recoverable logical prefix, composes every required edge in memory, validates and syncs a same-directory temporary stage for only the final target, rechecks the source fingerprint, publishes that previously absent target without overwrite, syncs the namespace, and reopens it through current validation before returning a handle. The source never moves or changes; only disposable temporary stages may be moved, linked, or removed. Migration does not synthesize interrupted-turn events: agent-loop appends those repairs through the write handle, while read-only query paths balance them in memory.
+`SessionFormatMigrationContext.emitEvent()` and `emitRun()` are synchronous. The producer declares whether it emits a scalar event or a compact run, so the hot path never infers the category from properties on a parsed file object. The caller owns scheduling and supplies the context to each operation instead of injecting a callback into the stage constructor. One input may emit zero, one, or many outputs without allocating a temporary return array or retaining an internal output queue.
 
-Canonical filenames encode the physical format generation: v0 is `session.jsonl` or `session.jsonl.zstd`; every positive generation is lowercase `session.vN.jsonl` or `session.vN.jsonl.zstd`. `dsh-session-format` owns the raw basename rule (`sessionFormatLogFilename`, `parseSessionFormatLogFilename`); the JSONL provider, the session-log export archive, and recorded-session fixtures append only the compression suffix. Publication never renames, replaces, or deletes a committed generation path. If the target already exists, it is accepted only as a regular current-format file with exactly the expected bytes; any other target refuses. Lower generations remain for operator inspection or explicit copying, but normal runtime operations select the numerically highest canonical name and never use retained predecessors as automatic fallback, restore, or downgrade support.
+`SessionFormatMigration` remains an immutable declaration: version numbers, header migration, target-header validation, and `createStage()`. `CompiledSessionFormatChain` validates a unique gap-free edge sequence once, creates per-artifact stages in source-to-target order, and connects them with context objects in reverse order. `finish()` settles stages in source-to-target order so each stage can emit its tail before the downstream stage closes.
 
-The current-format fast path classifies the header from one stable source snapshot, invokes no historical converter or generation write, and passes that snapshot to current decoding without another file read. The decoded log enters the existing bounded revision-keyed memo for an immediate observe-to-resume handoff, while `stat` and `list` deliberately rescan. Multiple edges leave the original generation unchanged and publish only the final target; intermediate versions exist only in memory. A source fingerprint recheck restarts migration when content changes, and exclusive target publication accepts a racing winner only when its bytes match exactly. Cross-process append fencing remains outside this guarantee.
+```text
+JSONL record
+  → released physical row decoder
+  → v0-to-v1 stage
+  → v1-to-v2 stage
+  → current event collector
+```
 
-The first edge, `@deepseek-ai/dsh-session-format-v0-to-v1`, is intentionally identity-shaped: aside from the version and bounded historical normalizations already accepted by v0, it preserves logical headers, events, sequence numbers, references, timestamps, payloads, and the configured compression choice. The exact `session.jsonl[.zstd]` source remains byte- and inode-identical, while the current writer encodes the new `session.v1.jsonl[.zstd]` successor. This exercises the complete publication lifecycle before a cardinality-changing format needs it.
+The chain contains no `flatMap`, spread expansion, intermediate event array, or scheduler. The final event collector expands a compact run only after every migration stage has had the opportunity to consume it directly.
 
-Projection-cache records bind their fold to the Session header's `formatVersion`. The `session_projcache` v7 reader may load predecessor domain records structurally, but a record without the format generation cannot seed a current Session; the authoritative log refolds it and the next checkpoint writes the complete current identity. This prevents a cache row produced before a bounded normalizer or cardinality-changing edge from bypassing that migration.
+### Physical codecs and packed runs
 
-## Consequences
+Each released codec creates a row decoder with explicit `strict` or `recoverable` recovery. The decoder validates and emits one event or one codec-owned `SessionFormatEventRun` at a time through separate context methods. v0-to-v1 and v1-to-v2 implement both `transformEvent()` and `transformRun()`, so packed Assistant chunks can reach the folding edge without first becoming millions of ordinary events.
+
+The v0-to-v1 edge preserves logical headers, sequence numbers, references, timestamps, and payloads except for bounded released-v0 normalizations. It translates the retired `steering/message` and `compact/*` event names, accepts a released `llm/retry` after its matching `step/end`, deterministically supplies a missing `llm/retry.retryId` per turn/step/provider/policy chain, and supplies one deterministic `compactionId` across a legacy compaction group that omitted it. The v1-to-v2 edge owns attempt folding and reference remapping, and emits only settled current events. It splits a legacy goal-sourced user message into `goal/change` plus the original model-visible message. It also inserts an interrupted `turn/end` for the bounded released restart in which an open turn with no open step is followed by a non-empty `next-turn` inbox splice and the next numbered `turn/start`.
+
+The catalog exposes one `createRestore()` operation for production, Worker, fixture, and replay callers. Recovery policy and final validation policy are chosen once at restore creation. Historical production uses recoverable source parsing with transformed-current validation; this validates the released current result after migration, while input that is already current receives only codec validation. Worker and fixture verification use strict parsing with full installed current restoration. A migration-stage or transformed-current validation refusal remains `SessionFormatUnsupportedMigrationError`; physical decoding failures remain corruption. Test support keeps only fixture-specific token and envelope materialization.
+
+### JSONL integration
+
+The JSONL provider scans frame boundaries once, reuses one Zstandard decoder, parses complete JSONL records incrementally, and feeds rows directly into the catalog restore. The outer loop yields at a bounded cadence; there is no per-frame `await` and no complete plaintext or source-row array.
 
-Reading event bodies with a newer build may durably add a higher generation. The exact old generation remains available, but the runtime thereafter selects the highest canonical filename; retention does not promise that an older build can safely downgrade or that the newer build will fall back when the successor is corrupt. A read-only filesystem reports an actionable migration failure instead of returning an in-memory current view that differs from disk.
+Current encoding is record based. The provider serializes about 1 MiB of plaintext per main-thread slice, streams it through one Zstandard context with source-error propagation, writes compressed output in 4 MiB batches to an exclusively created same-directory temporary file, and syncs it before publication. A process-wide scheduler admits at most two full verification Workers and hands a released permit directly to the oldest waiter.
 
-JSONL publication uses POSIX hard-link creation plus directory sync, and Windows uses no-overwrite `MoveFileExW` with write-through. A competing writer that wins target creation is accepted only when the committed bytes exactly match. One process-local writer per Session is the supported concurrency model. A future per-Session cross-process lock can close the remaining source-check-to-publication race without changing the format edge interface.
+Preparation forwards cancellation through source reads and observes it at the existing approximately 500 ms Decode yield boundary. Once `publish()` starts, encode, Worker verification, and publication do not receive caller cancellation and run to settlement; write open checks its caller signal again afterward. A published generation is never rolled back.
 
-Retained generations are not a live-stream write-ahead log. A future optional WAL sidecar may preserve unfinished assistant streams across a hard crash. Explicit generation inspection or copying, retention tooling, compression conversion, and streamed whole-artifact transformation are separate features; automatic fallback and downgrade compatibility are not implied future work.
+The Stage pipeline ends at one prepared current artifact. [Historical Session read preparation](2026-09-05-read-only-session-migration-preparation.md) defines how read open consumes that artifact immediately while write open performs encode, verification, and publication before returning append access.
 
-This note supersedes the continue-only persistence rule and the deferred-chain status in [Session log versioning](2026-08-10-session-log-version-mechanism.md). That note remains the authority for when to bump the version and for ordinary equal-version `ignorable` event behavior.
+### Durable format and publication rules
+
+Canonical filenames encode physical format generation: v0 is `session.jsonl[.zstd]` and positive generations use `session.vN.jsonl[.zstd]`. Migration never moves, replaces, or deletes a committed generation and writes only the final current target; intermediate versions exist only as stage state.
+
+POSIX publication uses hard-link creation plus directory sync. Windows uses no-overwrite, write-through `MoveFileExW`. An existing target is accepted only when its verified migration prefix equals the staged bytes; any append tail belongs to current-generation reading rather than migration winner verification.
+
+Existing write handles retain the process-local claim and kernel-backed cross-process `SessionWriteLease`. Header-only `stat` and `list` translate supported historical headers without opening the body or publishing a generation. Projection-cache records bind their fold to the Session header's format version so a cache row cannot bypass a cardinality-changing migration.
+
+## Problem-to-solution mapping
+
+| Whole-artifact problem | Implemented mechanism | Result |
+|---|---|---|
+| One asynchronous decode call per Zstandard frame | One reusable decoder; outer 500 ms scheduling cadence | Removes 317,540 async transitions |
+| Complete plaintext, string, and row arrays | Incremental JSONL parser and row decoder | Retains only one cross-chunk record fragment |
+| Complete event array between every edge | Context-connected stateful stages | No intermediate version event arrays |
+| Packed chunks expand before folding | `SessionFormatEventRun` plus `transformRun()` | 9.14 million source events need not materialize |
+| Whole-artifact snapshot and deep freeze at every edge | Stage-owned exclusive values and final validation | Removes repeated recursive copy/freeze |
+| One-shot migration functions hide state | Per-artifact stage classes from immutable declarations | State ownership and concurrency are explicit |
+| Bulk current encode builds whole strings and Buffers | Record encoder, 1 MiB input slices, 4 MiB write batches | Bounds allocation and main-thread slices |
+| Verification repeats on the main thread | At most two complete-generation Workers | Keeps verification CPU off the main thread |
+| Production and fixture migration use different APIs | Catalog `createRestore()` with explicit policies | One decoder/chain implementation |
 
 ## Verification
 
-Release verification runs the committed Session-format corpus gate over every versioned persisted-or-projected `session*.jsonl` fixture under `snapshots/`, `packages/`, and `scripts/snapshots/python-sdk-single-exe/`. Fixture-only omitted envelopes and request-header tokens are materialized before the real static catalog; every fixture reaches the current v1 view through current restoration or historical migration. Released-v0 replay inputs remain suffixless, while fresh v1 writer outputs use `session.v1.jsonl` for a parent and `session.<ordinal>.v1.jsonl` for children. Record and refresh preserve every completed generation, including generations of a child role absent from a later run. Malformed historical fixtures are repaired at their source rather than admitted through path-dependent replay policy. The continuing gate discovers the corpus dynamically and fails every restoration refusal; separate assembled JSONL tests own exact physical-byte migration.
+### Benchmark input and meanings
+
+The benchmark uses Node v24.18.0 and one 116,228,655-byte v0 Zstandard log containing 317,540 frames and 454,151 physical rows. The old reader restores 9,143,111 expanded v0 events. Migration produces 72,784 current v2 events with artifact SHA-256 `fa16ff9472ca350595a3112c20a3db79655bc2673973469987ecaf2a57ebd17c`.
+
+Runs use built artifacts under plain Node, one process per sample, and a 16 GB V8 heap limit. “Retained heap” is measured after forced GC while the restored Session remains live. Values below are three-run medians except the whole-artifact failure, which consistently cannot reach a handle.
+
+### Physical Decode
+
+| Data path | Decode time | Peak RSS | Scheduling |
+|---|---:|---:|---|
+| Pre-migration optimized reader | 1.553s | 916MB | One decoder; 2–3 outer yields |
+| Whole-artifact migration | 7.527s | 7,219MB | 317,540 async decoder calls |
+| Streaming Stage path | 1.467s | 908MB | One decoder; 2 outer yields |
+
+### Historical-file cold open
+
+| Version | Time to restored Session | CPU time | Peak RSS | Retained heap | Restored events | Outcome |
+|---|---:|---:|---:|---:|---:|---|
+| Pre-migration high-performance v0 reader | 4.594s | 6.048s | 2.720GB | 2.016GB | 9,143,111 | Reads v0; does not migrate |
+| Whole-artifact migration | >72.8s | — | ≥7.219GB during Decode | — | — | OOM before a handle |
+| Streaming Stage migration with serial publication | 6.241s | 8.493s | 2.107GB | 477MB | 72,784 | Publishes and opens v2 |
+
+The old reader has lower one-time wall time because it performs no format conversion or durable publication. It also keeps the 9.14-million-event representation live. The Stage path pays encode and verification once, then retains the folded v2 state.
+
+### Current-format cold open
+
+| Version reading its current format | Time to restored Session | Peak RSS | Retained heap |
+|---|---:|---:|---:|
+| Old reader on v0 | 4.594s | 2.720GB | 2.016GB |
+| Whole-artifact-era reader on v2 | 1.273s | 1.107GB | 476MB |
+| Streaming Stage reader on v2 | 1.284s | 1.109GB | 476MB |
+
+The current-v2 fast path remains performance-equivalent. The architectural change does not route current data through historical stages.
+
+### Streaming serial migration breakdown
+
+This table records the serial open flow measured for this Stage decision. The current preparation-first scheduling and its measurements are owned by [Historical Session read preparation](2026-09-05-read-only-session-migration-preparation.md).
+
+| Phase | Median |
+|---|---:|
+| Source Decode and migration | 2.784s |
+| Encode, write, and sync | 0.956s |
+| Full staged-file Worker verification | 1.415s |
+| Source recheck and no-overwrite publication | 0.106s |
+| Committed-prefix verification and header reopen | 0.046s |
+| Generation ensure-current total | 5.318s |
+| Final current decode observed by persistence | 0.620s |
+| Session restoration | 0.594s |
+| End-to-end restored Session | 6.241s |
+
+The generation breakdown and end-to-end table come from separate instrumented runs, so rounded rows are not expected to sum exactly.
+
+Format, catalog, edge, JSONL, fixture, replay, and built-Worker tests cover both encodings, packed runs, header-only classification, torn tails, migration refusal, deterministic legacy normalization, source changes, target collisions, write leases, and Worker failure.
+
+## Consequences
+
+At least one final current-event array remains necessary because Session restoration and Agent execution retain complete history. The Stage architecture removes full source and intermediate target arrays; it does not promise memory proportional to a page window.
+
+Decoded scalar `assistant/chunk` rows receive envelope validation and final target validation, but their complete frozen-v1 source payload-member validation is deferred because that per-event check materially affects Decode and migration time on released logs. Packed Assistant runs remain strictly decoded. The scalar check must be restored only with performance evidence that preserves this migration path's measured behavior.
 
-Handle-integration verification runs the pure format, catalog, persistence-seam, and JSONL provider suites together: 420 tests cover both encodings, immutable publication races, header-only observation, read and write handles, migration refusal, append after migration, cancellation, and crash-tail behavior with per-file 100% statement, branch, function, and line coverage. Repository typecheck and lint, 113 keyless recorded-session replays with two declared skips, and 28 owner-local expected-output cases also pass on the merged master checkpoint.
+Read-only access consumes the Stage result before durable publication, while write open reuses the same result and waits for publication before append. The persistence scheduling remains independent from the format pipeline.
 
-The assembled headless profile test stages `session.jsonl`, resumes it through the shipped composition, observes v1 before Session construction, verifies that the exact v0 bytes and inode remain while `session.v1.jsonl` appears, and proves the next append targets v1. JSONL contract tests exercise raw and Zstandard exclusive publication, torn-tail preservation, source changes, target collisions, future-highest refusal, revision-keyed parsed-log reuse, listing rescans, temporary cleanup, committed reopen, and current-format bypass.
+Lower generations remain for operator inspection. Retention does not promise downgrade compatibility, automatic fallback, or that an older runtime can safely interpret a newer generation.
 
 ## Alternatives considered
 
-- **Migrate only on continuation** — leaves query, export, fork, and suffix consumers on old generations and duplicates restoration policy.
-- **Return a migrated in-memory view without persisting** — lets one process observe state that does not match the highest committed generation and postpones failure until a later writer.
-- **Persist every intermediate version** — consumes space and creates recovery states with no runtime consumer; only the source and final generation are durable.
-- **Let mounted event-owner plugins register migrations** — makes historical readability deployment-dependent; the static catalog must work before feature plugins mount.
-- **Reuse one filename for every current format and relocate its predecessor** — rejected because migration would move or overwrite committed evidence, require collision and retention rules, and make the filename disagree with the stored format. Canonical immutable generation names let discovery select the highest version directly.
+- **Optimize only Zstandard Decode** — restores physical Decode speed but leaves source rows, expanded events, snapshots, intermediate artifacts, and bulk encode in memory.
+- **Synchronous Generator stages** — retain execution frames and batches at each yield. Real-log measurements increased migration time and migrate-complete RSS from about 1.0 GB to about 1.2 GB.
+- **Return arrays from each stage** — preserves the old allocation, traversal, and flattening costs under a new name.
+- **Give each stage an internal output queue** — adds drain, EOF, and error ownership while still retaining intermediate values.
+- **Inject an emit callback through constructors** — forces reverse construction or a partially connected lifecycle. Passing a context to operations keeps stage construction independent of downstream wiring.
+- **Share stateful codec instances globally** — would mix pending attempts, mappings, and counters across concurrent Session restores.
+- **Persist every intermediate format version** — creates durable states with no runtime consumer; only the exact source and final current generation are needed.
+- **Let mounted plugins register migrations** — makes historical readability deployment dependent. The static catalog must restore released formats before feature plugins mount.

.agents/notes/implemented/architecture/2026-08-31-released-session-format-migrations.zh.md

@@ -1,4 +1,4 @@
-# Agent Note: 已发布 Session 格式在读取正文时通过相邻纯迁移边升级
+# Agent Note: 已发布 Session 格式通过有状态流式 Stage 迁移
 
 Status: implemented
 
@@ -6,52 +6,176 @@ Status: implemented
 
 ## 问题
 
-Session 格式 v0 已随 alpha 版本发布，因此结构化 writer 变更不能再把已有 JSONL 当作可丢弃的预发布状态。已存储事件正文通过读或写 `SessionHandle` 到达恢复、查询、导出、分叉与继续路径。只迁移一个消费方会让调用方看到不同的逻辑 generation，或只在后续 writer 到达旧文件时失败。
+Session 格式 v0 已随 alpha 版本发布，因此结构化 writer 变更不能再把已有 JSONL 当作可丢弃的预发布状态。第一版 whole-artifact migration 让这些日志在语义上可迁移，但它的数据模型会让一份 116 MB 真实 Session 在返回 handle 前耗尽 16 GB Node 进程。
 
-迁移必须保留精确源路径、字节与 inode，包括撕裂的物理尾部，同时为每个已发布格式提供一个无歧义的规范文件名。普通 JSONL 与 Zstandard 是同一逻辑格式的编码选择，不能产生两套并行迁移实现。
+### Whole-artifact 性能问题
+
+- Zstandard 输入包含 317,540 个 frame，每个 frame 都单独执行一次异步解压。实现先保留全部 plaintext frame，再在 JSON 解析前统一拼接，因此创建了同等数量的 Promise、线程池与 native Decode 调度。
+- Physical Decode 会在同一请求的重叠阶段物化完整 plaintext Buffer、完整字符串、全部 JSONL row、展开后的 source events、迁移后的 target events、编码后的 target rows、拼接后的目标字符串与目标 physical Buffer。
+- 每个 codec 与 migration edge 都会调用 `snapshotSessionFormatJson()` 或 `snapshotSessionFormatArtifact()`，在相邻迁移前后递归复制并 deep freeze 完整 header、row、payload 与 event array。
+- 已发布的 packed Assistant chunk 会先展开成约 914 万个 v0/v1 逻辑事件，再由 v1-to-v2 折叠成 72,784 个 current events。Whole-artifact API 要求两种表示和 old-to-new seq map 同时存活。
+- Encode 会在内存中构造完整 JSONL 与压缩输出。成功路径随后 Decode staged target、Decode committed target，并由 persistence 再 Decode 一次以创建业务对象；它还会完整重读 source 以比较 fingerprint。
+- 逐 frame `await` 没有形成有意义的有界调度。迁移前的高性能 reader 会复用一个同步 decoder，只由外层循环约每 500 ms yield 一次，从而避免数十万次异步切换。
+
+### 既有接口使单点优化无法组合
+
+`SessionFormatCodec` 以完整数组 Decode 与 Encode；每条相邻 `SessionFormatMigration` 接收并返回完整 `SessionFormatArtifact`；compiled chain 只能把已经物化的 artifact 交给下一条 edge。因此即使 physical decoder 单点变快，下游仍会重新创建 source-row array、expanded-event array、逐 edge snapshot 与 target-row array。
+
+Migration 实际有状态，但 API 把它们表现为一次性函数。v0-to-v1 需要跟踪 message 与 retry identity；v1-to-v2 需要暂存一个尚未结算的 Assistant attempt、被它阻塞的后续事件，并维护 old-to-new seq 引用。把这些状态隐藏在 closure 或 push/finish helper object 中，会让运行结构与静态声明分离，也让 production、Worker verify、fixture 与 replay 使用不同入口。
 
 ## 决策
 
-`SESSION_FORMAT_VERSION` 是单调递增的当前 writer 整数。每个相邻 `vN -> vN+1` 转换由一个与 profile 无关的纯包负责。`@deepseek-ai/dsh-session-format` 只提供无损快照、唯一且无缺口的规划、仅 header 转换与整产物组合；`@deepseek-ai/dsh-session-format-catalog` 静态导入完整链，不依赖已挂载的 Cordis 插件。历史 codec 和归一化器位于具名迁移边包中，而当前 Session 与持久化代码只接纳最新逻辑类型。
+Session format 包采用有状态同步 Stage API。静态 migration declaration 描述一条相邻版本边，并为每次 artifact restore 创建新的 stage。Stage 拥有该 artifact 的可变状态；不同 Session 之间绝不共享 stage instance。
 
-每条迁移边都会冻结严格的源与目标语义，其目标物理 codec 则保持词汇中立，使普通事件增长可以留在同一格式版本内。目录通过已安装的 peer `@deepseek-ai/dsh-session` 及其当前 `KNOWN_SESSION_EVENT_TYPES` 还原最终代，避免冻结的历史迁移边反过来成为当前词汇 owner。
+### Stage 与 Context 协议
 
-JSONL provider 在 `open` 为已存储 Session 返回句柄前完成 ensure-current 工作。它选择最高规范 generation、迁移受支持的历史正文，并从同一物理快照解码当前结果；公开 `SessionPersistence` 与 `SessionHandle` 接口不包含迁移操作。仅 header 的 `stat` 与 `list` 会重新扫描 Session 目录，在内存中转换受支持的历史 header，且绝不发布后继。`create` 独立于 header 可读性检查规范文件名，因此每个现有 generation 都会占用其 Session id。
+```text
+interface SessionFormatMigrationContext {
+  emitEvent(event: SessionFormatEvent): void
+  emitRun(run: SessionFormatEventRun): void
+}
 
-取消属于提供信号的 `open`、`stat` 或 `list` 调用。发现、稳定读取、解码与发布前检查都会观察该信号；不可变后继一旦发布且其目录项已经同步，后续取消不会删除已提交 generation。
+interface SessionFormatMigrationStage {
+  readonly headerInheritedEventCount?: number
+  transformEvent(
+    event: SessionFormatEvent,
+    context: SessionFormatMigrationContext,
+  ): void
+  transformRun(
+    run: SessionFormatEventRun,
+    context: SessionFormatMigrationContext,
+  ): void
+  finish(context: SessionFormatMigrationContext): number
+}
+```
 
-配置的 JSONL 编码拥有一个完整后缀：`.jsonl` 或 `.jsonl.zstd`。迁移读取稳定的精确源，解码可恢复逻辑前缀，在内存中组合全部必需迁移边，只为最终目标校验并同步同目录临时 stage，重新检查源 fingerprint，以不覆盖方式发布此前不存在的目标，同步 namespace，并在返回句柄前通过当前格式校验重新打开。源永不移动或改变；只有可丢弃临时 stage 可以被移动、链接或移除。迁移不会合成中断轮次事件：agent-loop 通过写句柄追加这些修复，而只读查询路径在内存中补齐它们。
+`SessionFormatMigrationContext.emitEvent()` 与 `emitRun()` 都是同步操作。Producer 会声明其发出单个事件还是紧凑 run，因此热路径不会根据已解析文件对象的属性推断类别。调度归 caller 所有，context 在每次调用时传入，而不是把 callback 注入 stage constructor。一个输入可以输出零个、一个或多个值，不需要分配临时返回数组，也不需要 stage 内部保留输出队列。
 
-规范文件名编码物理格式 generation：v0 是 `session.jsonl` 或 `session.jsonl.zstd`；每个正 generation 都是小写 `session.vN.jsonl` 或 `session.vN.jsonl.zstd`。`dsh-session-format` 拥有原始 basename 规则（`sessionFormatLogFilename`、`parseSessionFormatLogFilename`）；JSONL provider、session-log 导出归档与 recorded-session fixture 只追加压缩后缀。发布绝不重命名、替换或删除已提交 generation 路径。目标已经存在时，只有它是普通当前格式文件且字节与预期完全相同时才接受；其他目标都会拒绝。低 generation 为 operator 检查或显式复制而保留，但普通 runtime 操作选择数值最高的规范名称，绝不把保留的前任当作自动 fallback、restore 或 downgrade 支持。
+`SessionFormatMigration` 继续作为 immutable declaration，声明版本号、header migration、target-header validation 与 `createStage()`。`CompiledSessionFormatChain` 只校验一次唯一、无缺口的 edge 序列，按 source-to-target 顺序创建每次 artifact 独占的 stage，再按反方向用 context 连接它们。`finish()` 按 source-to-target 顺序关闭 stage，使每一级都能在下游关闭前发出尾部数据。
 
-当前格式快速路径从一个稳定源快照分类 header，不调用历史 converter，不写 generation，并把该快照交给当前格式解码，而不再次读取文件。解码日志进入现有按 revision 为键的有界 memo，供紧接的观察到恢复交接复用，而 `stat` 与 `list` 会有意重新扫描。多条迁移边保持原 generation 不变，并只发布最终目标；中间版本只存在于内存。源 fingerprint 重新检查会在内容变化时重启迁移，排他目标发布只在竞争胜者字节完全相同时接受它。跨进程 append 隔离不在此保证内。
+```text
+JSONL record
+  → released physical row decoder
+  → v0-to-v1 stage
+  → v1-to-v2 stage
+  → current event collector
+```
 
-第一条迁移边 `@deepseek-ai/dsh-session-format-v0-to-v1` 有意保持恒等形态：除版本和 v0 已接纳的有限历史归一化外，它保留逻辑 header、事件、序号、引用、时间戳、payload 与已配置的压缩选择。精确的 `session.jsonl[.zstd]` 源保持字节与 inode 相同，当前 writer 则编码新的 `session.v1.jsonl[.zstd]` 后继。这样可在出现改变基数的格式前先验证完整发布生命周期。
+Chain 中不存在 `flatMap`、spread expansion、中间 event array 或 scheduler。只有在每个 migration stage 都已获得直接消费 compact run 的机会后，最终 event collector 才会展开它。
 
-投影缓存记录把自己的折叠结果绑定到 Session header 的 `formatVersion`。`session_projcache` v7 reader 可以在结构上载入前代 domain 记录，但缺少格式代的记录不能播种当前 Session；权威日志会重新折叠它，下一次检查点写入完整的当前 identity。这样，任何在有界规范化或基数变化边之前产生的缓存行都不能绕过该迁移。
+### Physical codec 与 packed run
 
-## 后果
+每个 released codec 会用显式 `strict` 或 `recoverable` 策略创建 row decoder。Decoder 每次通过不同的 context 方法校验并 emit 一个 event 或 codec-owned `SessionFormatEventRun`。v0-to-v1 与 v1-to-v2 都实现 `transformEvent()` 和 `transformRun()`，因此 packed Assistant chunk 可以直接到达 folding edge，无需先变成数百万个普通事件。
+
+v0-to-v1 除了有限的 released-v0 归一化外，会保留逻辑 header、seq、引用、时间戳与 payload。它转换已移除的 `steering/message` 与 `compact/*` 事件名称，接受出现在对应 `step/end` 之后的已发布 `llm/retry`，按 turn／step／provider／policy chain 为缺失的 `llm/retry.retryId` 确定性补值，并为省略 id 的旧 compaction group 确定性补充同一个 `compactionId`。v1-to-v2 负责 attempt folding 与引用重写，并且只 emit 已结算的 current event。它会把旧的 goal 来源 user message 拆成 `goal/change` 与原本的模型可见 message。它还会为一种有限的已发布 restart 插入 interrupted `turn/end`：一个没有 open step 的 open turn 后出现非空 `next-turn` inbox splice，随后直接开始编号连续的下一轮。
+
+Catalog 为 production、Worker、fixture 与 replay 暴露同一个 `createRestore()`。Recovery policy 与最终 validation policy 在 restore 创建时一次确定。Historical production 使用 recoverable source parsing 与 transformed-current validation；这种策略会在迁移后校验已发布 current 结果，而已经是 current 的输入只接受 codec 校验。Worker 与 fixture verification 使用 strict parsing 与已安装 current 格式的完整 restoration。Migration stage 或 transformed-current validation 的拒绝会保持为 `SessionFormatUnsupportedMigrationError`；物理解码失败仍是 corruption。Test support 只保留 fixture 自身需要的 token 和 envelope materialization。
+
+### JSONL 串联
+
+JSONL provider 只扫描一次 frame boundary，复用一个 Zstandard decoder，增量解析完整 JSONL record，并把 row 直接送入 catalog restore。外层循环按有界 cadence yield；不存在逐 frame `await`、完整 plaintext 或 source-row array。
 
-较新 build 读取事件正文时可能持久增加一个更高 generation。精确旧 generation 仍然可用，但 runtime 此后选择最高规范文件名；保留不承诺旧 build 能安全 downgrade，也不保证新 build 在后继损坏时 fallback。只读文件系统会报告可操作的迁移失败，而不会返回与磁盘不一致的内存当前视图。
+Current encode 以单条 record 为单位。Provider 在主线程每个 slice 序列化约 1 MiB plaintext，通过一个会传播 source error 的 Zstandard context 流式压缩，以 4 MiB batch 写入同目录排他创建的临时文件，并在 publication 前 sync。进程级 scheduler 最多允许两个完整 verification Worker 并行，并把释放的 permit 直接交给最早的 waiter。
 
-JSONL 发布在 POSIX 上使用硬链接创建与目录同步，在 Windows 上使用 write-through 且不覆盖的 `MoveFileExW`。竞争 writer 已先创建目标时，只有已提交字节完全匹配才接受。每个 Session 只支持一个进程内 writer。未来逐 Session 跨进程锁可以关闭剩余的源检查到发布竞态，而无需改变格式迁移边接口。
+Preparation 会把 cancellation 传给 source read，并在现有的约 500 ms Decode yield 边界观察它。`publish()` 一旦开始，encode、Worker verification 与 publication 不接收 caller cancellation，并运行到终态；write open 会在之后再次检查 caller signal。已经发布的 generation 绝不会回滚。
 
-保留的 generation 不是实时流 WAL。未来可选 WAL sidecar 可以在硬崩溃间保留未完成 assistant 流。显式 generation 检查或复制、保留策略工具、压缩转换与流式整产物转换都是独立功能；自动 fallback 与 downgrade compatibility 并非隐含 future work。
+Stage pipeline 终止于一份 prepared current artifact。[历史 Session 只读迁移准备](2026-09-05-read-only-session-migration-preparation.zh.md)定义 read open 如何立即消费该 artifact，以及 write open 如何在返回 append 权限前完成 encode、verification 与 publication。
 
-本记录取代 [Session 日志版本机制](2026-08-10-session-log-version-mechanism.zh.md) 中仅在继续时持久化和迁移链仍推迟的规则。原记录继续负责何时递增版本，以及普通同版本 `ignorable` 事件行为。
+### Durable format 与 publication 规则
+
+规范文件名编码 physical format generation：v0 使用 `session.jsonl[.zstd]`，正 generation 使用 `session.vN.jsonl[.zstd]`。Migration 不会移动、覆盖或删除任何 committed generation，并且只写最终 current target；中间版本只存在于 stage state。
+
+POSIX publication 使用 hard-link creation 加目录 sync；Windows 使用 no-overwrite、write-through 的 `MoveFileExW`。已有 target 只有在其已校验 migration prefix 等于 staged bytes 时才会被接受；任何 append tail 都属于 current-generation reader，而不是 migration winner verification。
+
+既有 write handle 继续使用进程内 claim 与内核支持的跨进程 `SessionWriteLease`。仅 header 的 `stat` 与 `list` 可以转换受支持的历史 header，但不打开 body，也不发布 generation。Projection-cache record 会把 fold 绑定到 Session header 的 format version，使 cache row 不能绕过改变 event 基数的 migration。
+
+## 问题与方案对照
+
+| Whole-artifact 问题 | 实现机制 | 结果 |
+|---|---|---|
+| 每个 Zstandard frame 单独异步 Decode | 一个可复用 decoder；外层 500 ms 调度 cadence | 删除 317,540 次异步切换 |
+| 完整 plaintext、string 与 row array | 增量 JSONL parser 与 row decoder | 只保留一条跨 chunk 残行 |
+| 每条 edge 之间都形成完整 event array | Context 直连的有状态 stage | 不保留中间版本 event array |
+| Packed chunk 在 folding 前完整展开 | `SessionFormatEventRun` 与 `transformRun()` | 无需物化 914 万 source events |
+| 每条 edge 都 whole-artifact snapshot/deep freeze | Stage-owned 独占值与最终 validation | 删除重复递归复制与冻结 |
+| One-shot migration function 隐藏状态 | Immutable declaration 创建每次 artifact 独占的 stage class | 状态 ownership 与并发关系显式化 |
+| Bulk current encode 构造完整 string/Buffer | 单条 record encoder、1 MiB input slice、4 MiB write batch | 限制分配与主线程 slice |
+| 主线程重复执行完整 verification | 最多两个 complete-generation Worker | Verification CPU 不占用主线程 |
+| Production 与 fixture 使用不同 migration API | Catalog `createRestore()` 加显式 policy | 只保留一套 decoder/chain 实现 |
 
 ## 验证
 
-发布验证针对 `snapshots/`、`packages/` 与 `scripts/snapshots/python-sdk-single-exe/` 下每个带版本、来自持久化或投影的 `session*.jsonl` fixture 运行已提交 Session 格式语料门禁。fixture 专用的缺失信封与 request-header token 会先被实体化，再进入真实静态 catalog；每个 fixture 都会通过当前格式 restore 或历史迁移得到当前 v1 视图。Released-v0 replay 输入保持无后缀，而新鲜 v1 writer 输出对 parent 使用 `session.v1.jsonl`、对 child 使用 `session.<ordinal>.v1.jsonl`。Record 与 refresh 会保留每个已完成 generation，包括后续运行不再产生的 child role generation。Malformed 历史 fixture 在来源处修复，不通过依赖路径的 replay 策略准入。持续运行的门禁会动态发现语料，并拒绝每个 restore failure；独立组装式 JSONL 测试负责精确物理字节迁移。
+### Benchmark 输入与口径
+
+Benchmark 使用 Node v24.18.0 和一份 116,228,655-byte 的 v0 Zstandard 日志，其中包含 317,540 个 frame 与 454,151 个 physical row。老 reader 会恢复 9,143,111 个展开后的 v0 event；migration 会生成 72,784 个 current v2 event，artifact SHA-256 为 `fa16ff9472ca350595a3112c20a3db79655bc2673973469987ecaf2a57ebd17c`。
+
+所有样本均通过 plain Node 运行 build artifact，每个样本使用独立进程，V8 heap limit 为 16 GB。“Retained heap”表示 restored Session 仍存活时强制 GC 后的 heap。除无法得到 handle 的 whole-artifact 失败外，下表使用三次运行中位数。
+
+### Physical Decode
+
+| 数据路径 | Decode 耗时 | 峰值 RSS | 调度 |
+|---|---:|---:|---|
+| Migration 前的高性能 reader | 1.553s | 916MB | 一个 decoder；外层 yield 2–3 次 |
+| Whole-artifact migration | 7.527s | 7,219MB | 317,540 次 async decoder 调用 |
+| Streaming Stage 路径 | 1.467s | 908MB | 一个 decoder；外层 yield 2 次 |
+
+### 历史文件首次冷打开
+
+| 版本 | Session restore 完成 | CPU 时间 | 峰值 RSS | Retained heap | Restore event 数 | 结果 |
+|---|---:|---:|---:|---:|---:|---|
+| Migration 前的高性能 v0 reader | 4.594s | 6.048s | 2.720GB | 2.016GB | 9,143,111 | 读取 v0，不迁移 |
+| Whole-artifact migration | >72.8s | — | Decode 阶段已 ≥7.219GB | — | — | 返回 handle 前 OOM |
+| Streaming Stage migration + 串行 publication | 6.241s | 8.493s | 2.107GB | 477MB | 72,784 | 发布并打开 v2 |
+
+老 reader 的一次性 wall time 更低，因为它不做格式转换和 durable publication；同时它会常驻 914 万 event 的表示。Stage 路径只多支付一次 encode 与 verification，随后保留折叠后的 v2 state。
+
+### Current-format 冷打开
+
+| 版本读取自己的 current format | Session restore 完成 | 峰值 RSS | Retained heap |
+|---|---:|---:|---:|
+| 老 reader 读取 v0 | 4.594s | 2.720GB | 2.016GB |
+| Whole-artifact 时代 reader 读取 v2 | 1.273s | 1.107GB | 476MB |
+| Streaming Stage reader 读取 v2 | 1.284s | 1.109GB | 476MB |
+
+Current-v2 快路径保持性能等价。架构改造不会让 current data 进入 historical stage。
+
+### Streaming 串行 migration 分段
+
+下表记录该 Stage 决策测量的串行 open 流程。当前 preparation-first 调度及其测量由[历史 Session 只读迁移准备](2026-09-05-read-only-session-migration-preparation.zh.md)记录。
+
+| 阶段 | 中位耗时 |
+|---|---:|
+| Source Decode + migration | 2.784s |
+| Encode + write + sync | 0.956s |
+| staged 文件完整 Worker verify | 1.415s |
+| Source recheck + no-overwrite publication | 0.106s |
+| committed-prefix verify + header reopen | 0.046s |
+| Generation ensure-current 总计 | 5.318s |
+| Persistence 观察到的最终 current Decode | 0.620s |
+| Session restore | 0.594s |
+| 端到端 Session restore 完成 | 6.241s |
+
+Generation 分段与端到端数据来自不同 instrumented run，因此四舍五入后的各行不要求精确相加。
+
+Format、catalog、edge、JSONL、fixture、replay 与 built-Worker 测试覆盖两种编码、packed run、仅 header 分类、torn tail、migration refusal、确定性 legacy normalization、source change、target collision、write lease 与 Worker failure。
+
+## 后果
+
+最终 current-event array 仍然不可消除，因为 Session restore 与 Agent 执行需要完整历史。Stage 架构删除完整 source 与中间 target array，但不承诺内存与分页窗口大小成正比。
+
+解码后的单条 `assistant/chunk` 会接受 envelope 校验与最终 target 校验，但其完整冻结 v1 source payload 成员校验仍处于延期状态，因为这项逐事件检查会显著影响已发布日志的 Decode 与 migration 耗时。Packed Assistant run 仍接受严格解码。只有性能证据表明不会破坏该迁移路径的已测表现时，才能恢复单条 chunk 校验。
 
-句柄集成验证会一起运行纯格式、catalog、持久化 seam 与 JSONL provider 测试套件：420 个测试覆盖两种编码、不可变发布竞态、仅 header 观察、读写句柄、迁移拒绝、迁移后 append、取消与崩溃尾部行为，并达到逐文件 100% statement、branch、function 与 line coverage。仓库 typecheck 与 lint、含两个已声明 skip 的 113 个无密钥 recorded-session replay，以及 28 个 owner-local expected-output case 也都在合并 master 的 checkpoint 上通过。
+Read-only access 会在 durable publication 前消费 Stage 结果；write open 则复用同一结果，并在 append 前等待 publication。Persistence 调度仍与 format pipeline 相互独立。
 
-组装后的 headless profile 测试会暂存 `session.jsonl`，通过随附组合恢复它，在构造 Session 前观察到 v1，验证精确 v0 字节与 inode 保持不变而 `session.v1.jsonl` 出现，并证明下一次 append 以 v1 为目标。JSONL 约定测试覆盖 raw 与 Zstandard 排他发布、撕裂尾部保留、源变化、目标冲突、最高未来版本拒绝、按 revision 复用已解析日志、列表重新扫描、临时文件清理、已提交重开与当前格式直通。
+低 generation 为 operator 检查而保留。Retention 不承诺 downgrade compatibility、automatic fallback，也不保证旧 runtime 能安全理解新 generation。
 
 ## 考虑过的替代方案
 
-- **只在继续时迁移**——让查询、导出、分叉与后缀消费者停留在旧代际，并重复恢复策略。
-- **返回迁移后的内存视图但不持久化**——让进程观察到与最高已提交 generation 不一致的状态，并把失败推迟到后续 writer。
-- **持久化每个中间版本**——消耗空间并产生没有 runtime 消费者的恢复状态；只有源与最终代际应持久。
-- **让已挂载事件 owner 插件注册迁移**——使历史可读性依赖部署；静态 catalog 必须在功能插件挂载前工作。
-- **让每个当前格式复用同一个文件名并迁走前任**——不予采用，因为迁移会移动或覆盖已提交证据，需要冲突与保留规则，并让文件名与存储格式不一致。规范不可变 generation 名让发现流程直接选择最高版本。
+- **只优化 Zstandard Decode**——可以恢复 physical Decode 速度，但 source rows、expanded events、snapshot、intermediate artifact 与 bulk encode 仍会留在内存中。
+- **同步 Generator stage**——每个 yield 都会保留执行帧与 batch。真实日志测量使 migration 更慢，并让 migrate-complete RSS 从约 1.0 GB 增长到约 1.2 GB。
+- **每个 stage 返回数组**——只是给旧的 allocation、遍历与 flattening cost 换了名字。
+- **Stage 内部输出队列**——增加 drain、EOF 与 error ownership，同时仍会保留中间值。
+- **通过 constructor 注入 emit callback**——迫使 chain 反向构建或引入 partially connected lifecycle。操作时传 context 可以让 stage construction 不依赖下游 wiring。
+- **全局复用有状态 codec instance**——会让不同 Session 的 pending attempt、mapping 与 counter 相互污染。
+- **持久化每个中间格式版本**——产生没有 runtime consumer 的 durable state；只需要精确 source 与最终 current generation。
+- **让 mounted plugin 注册 migration**——使历史可读性依赖部署。Static catalog 必须在 feature plugin 挂载前恢复已发布格式。

.agents/notes/implemented/architecture/2026-09-01-v2-embedded-assistant-streams.i18n.yaml

@@ -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/architecture/2026-09-01-v2-embedded-assistant-streams.md
-2026-09-01-v2-embedded-assistant-streams.md: c9d7a66481d4258de8f9f7abdaa12d25f5217510
-2026-09-01-v2-embedded-assistant-streams.zh.md: ff4caceb68c34e470d2dd81fc9cf214555ad676e
+2026-09-01-v2-embedded-assistant-streams.md: bee4d50fb830caa277bb700f3415e6d7f98ff64b
+2026-09-01-v2-embedded-assistant-streams.zh.md: 9116e94111b78af68d33f6f01dc289ee9f349b7e

.agents/notes/implemented/architecture/2026-09-01-v2-embedded-assistant-streams.md

@@ -21,17 +21,19 @@ Session format v2 has no top-level `assistant/chunk` event. Each model attempt c
 
 `AssistantStreamAccumulator` snapshots each chunk once. Consecutive text, reasoning, or tool-argument deltas for the same block become one compact run with its first timestamp, exact timestamp gaps, and one array member per original delta. Every other chunk remains a timestamped raw record. `expandAssistantStream()` strictly validates and reconstructs the exact timed sequence; compaction never joins delta boundaries.
 
-The current v2 validator requires the embedded stream to reproduce a non-empty `assistant/message`'s content, usage, and replay state. An empty stream remains valid for a migrated legacy message that had no source chunks. `assistant/message` cannot carry obsolete chunk `sourceEventSeqs`; ordinary user and tool surface provenance remains available.
+The migration publication verifier and frozen v2 fixture validator require the embedded stream to reproduce a non-empty `assistant/message`'s content, usage, and replay state. An empty stream remains valid for a migrated legacy message that had no source chunks. Ordinary Session restoration validates the settlement fields needed by the runtime without expanding every historical stream; consumers that expand a compact stream validate its records when they read it. `assistant/message` cannot carry obsolete chunk `sourceEventSeqs`; ordinary user and tool surface provenance remains available.
 
 ### Live presentation and durable replay
 
 `agent/assistant-stream` publishes process-local start, transient chunk, and end frames. The loop appends the complete `assistant/message` or `assistant/attempt` before a committed end frame names its type and sequence. An abandoned end has no settlement.
 
-The Web follow adapter opts into these process-local frames and adds the last durable sequence observed at each start. It presents chunks as Client-only `assistant/live-chunk` updates between durable cursors, stages only a later matching settlement until the committed end, and reopens follow on a revision gap. A committed end publishes a named settlement delta that removes the attempt's transient matches, adds the durable entry, and replays only affected Conversation Contexts; an abandoned end publishes the same delta without an entry. A reconnect baseline carries the active attempt's durable start cursor and compact prefix. Paged history, replay, telemetry, token accounting, and cold UI assembly read the durable embedded stream rather than the live frames.
+The Web follow adapter opts into these process-local frames and adds the last durable sequence observed at each start. It presents chunks as Client-only `assistant/live-chunk` updates between durable cursors, stages only a later matching settlement until the committed end, and reopens follow on a revision gap. A committed end publishes a named settlement delta that removes the attempt's transient matches, adds the durable entry, and replays only affected Conversation Contexts; an abandoned end publishes the same delta without an entry. A reconnect baseline carries the active attempt's durable start cursor and compact prefix.
+
+The Client event source passes durable settlements through unchanged. The Chat and Trajectory Assistant nodes fold `assistant/live-chunk` while an attempt is active, build settled output directly from `assistant/message`, and do not replay an `assistant/attempt` stream for presentation. Cold settled presentation therefore does not reconstruct per-token timing; other consumers may expand the durable stream when they require its exact evidence.
 
 ### Released v1 to v2 migration
 
-The adjacent migration validates the complete frozen v1 artifact, groups chunks by turn, step, terminal boundary, and exact message provenance, and then substitutes one settlement per attempt. A successful group's chunks move into its message. An unclaimed group becomes `assistant/attempt` at the last consumed chunk's position. Unrelated interleaved events retain their relative order, and survivors receive dense v2 sequence numbers. The edge compacts, expands, and re-assembles embedded streams through the runtime `AssistantStreamAccumulator`, `expandAssistantStream`, and `BlockAssembler` from `dsh-llm` instead of frozen copies, because that package owns the v2 stream encoding. Target validation re-checks agreement between each migrated `assistant/message` and its embedded stream itself, so a disagreeing v1 log is refused as an unsupported migration with its source artifact retained instead of surfacing as corruption from the installed Session restoration. A later format that changes the stream encoding must freeze copies of these helpers into this edge.
+The adjacent migration validates the complete frozen v1 artifact, groups chunks by turn, step, terminal boundary, and exact message provenance, and then substitutes one settlement per attempt. A successful group's chunks move into its message. An unclaimed group becomes `assistant/attempt` at the last consumed chunk's position. Unrelated interleaved events retain their relative order, and survivors receive dense v2 sequence numbers. The edge compacts embedded streams through the runtime `AssistantStreamAccumulator` from `dsh-llm` instead of a frozen copy, because that package owns the v2 stream encoding. The isolated publication verifier expands and re-assembles the written stream through `expandAssistantStream()` and `BlockAssembler`, then checks each migrated `assistant/message` against it before publication. A later format that changes the stream encoding must freeze copies of these helpers into this edge.
 
 The edge remaps the finite declared reference inventory: envelope provenance, surface replacement endpoints, command source events, compaction ranges and shadowed lists, and title message lists. The model-visible text of a validated `session/title-llm-request` remains byte-identical in the source sequence namespace while its `messageSeqs` field moves to the v2 namespace; target validation therefore does not reconstruct that text from remapped sequences. A reference to a consumed chunk refuses migration; it is never redirected to a settlement with different meaning. The edge also refuses an inherited cut that splits an attempt.
 
@@ -49,7 +51,7 @@ The compact-stream tests pin exact accumulation and expansion for text, reasonin
 
 The pre-merge performance acceptance measured static catalog-routing overhead against direct released-v2 restoration of the same already parsed physical rows across three runs, 100 warmup pairs, and 600 measured pairs; it did not compare v1 with v2 or time backend I/O. Every pooled median and p95 regression stayed within the 5% budget, with a worst p95 regression of 3.150%.
 
-Agent-loop tests pin durable-before-end ordering, interrupted visible prefixes, failed and retry attempts, abandonment, usage, and replay metadata. Session Controller and Conversation tests pin live transient display, reconnect baselines, committed settlement release, history replay, Chat and Trajectory parity, while TypeScript and Python SDK snapshots pin the external event representation.
+Agent-loop tests pin durable-before-end ordering, interrupted visible prefixes, failed and retry attempts, abandonment, usage, and replay metadata. Session Controller and Conversation tests pin live transient display, reconnect baselines, committed settlement release, and history replay. Chat and Trajectory tests pin live partial presentation and direct final-message projection, while TypeScript and Python SDK snapshots pin the external event representation.
 
 ## Alternatives considered
 
@@ -59,13 +61,15 @@ Agent-loop tests pin durable-before-end ordering, interrupted visible prefixes,
 
 **Carry packed chunk rows through the history API.** This reduces wire and Client work for v1 but gives the Client a second event vocabulary and keeps transport coupled to token-row cardinality. The current API carries scalar durable settlements plus a separate live transient stream.
 
+**Strip embedded streams in Session Controller.** This reduces retained Client memory but creates a second durable event type and makes a transport-facing owner decide which evidence presentation consumers need. The measured bottleneck is repeated expansion, so each UI consumer decides whether to inspect the unchanged settlement.
+
 **Store the stream in a sidecar or replay-only fixture.** This splits one attempt's message and evidence across durability owners and cannot give ordinary resumed sessions the same failed-output and timing facts. The settlement is the atomic owner.
 
 **Redirect references from consumed chunks to their settlement.** A chunk and an attempt settlement are not interchangeable facts. Refusal prevents a migration from silently changing the meaning of plugin-owned references.
 
 ## Consequences
 
-Current logs, telemetry, history pages, and cold Client assembly scale by model attempts rather than token chunks while retaining exact stream evidence inside each settlement. Live presentation remains incremental and intentionally process-local.
+Current logs, telemetry, and history pages scale by model attempts rather than token chunks while retaining exact stream evidence inside each settlement. The Client event window retains that compact evidence, but the Chat and Trajectory Assistant nodes do not expand settled streams into per-delta objects. Live presentation remains incremental and intentionally process-local.
 
 Unlike v1 top-level chunks, which the buffered persistence writer could flush before an attempt ended, v2 has no durable attempt evidence until settlement. A hard process or host loss before settlement discards the complete in-flight stream; `agent/assistant-stream` is not a write-ahead log. This tradeoff avoids a second durability owner for live output.
 

.agents/notes/implemented/architecture/2026-09-01-v2-embedded-assistant-streams.zh.md

@@ -21,17 +21,19 @@ Session format v2 没有顶层 `assistant/chunk` 事件。每个模型 attempt
 
 `AssistantStreamAccumulator` 对每个 chunk 只快照一次。同一 block 的连续 text、reasoning 或 tool argument delta 会变成一个紧凑 run，包含首个时间戳、精确时间戳间隔和每个原始 delta 对应的一个数组成员。其他 chunk 保留为带时间戳的 raw record。`expandAssistantStream()` 会严格校验并重建精确的带时间序列；压缩绝不会合并 delta 边界。
 
-当前 v2 校验器要求嵌入式 stream 能复现非空 `assistant/message` 的 content、usage 与 replay state。对于没有源 chunk 的已迁移旧 message，空 stream 仍然有效。`assistant/message` 不能携带已停用的 chunk `sourceEventSeqs`；普通 user 与 tool surface provenance 保持可用。
+Migration publication verifier 与冻结的 v2 fixture validator 要求嵌入式 stream 能复现非空 `assistant/message` 的 content、usage 与 replay state。对于没有源 chunk 的已迁移旧 message，空 stream 仍然有效。普通 Session restore 只校验 runtime 直接依赖的 settlement 字段，不展开全部历史 stream；需要展开 compact stream 的 consumer 会在读取时校验 record。`assistant/message` 不能携带已停用的 chunk `sourceEventSeqs`；普通 user 与 tool surface provenance 保持可用。
 
 ### 实时呈现与持久回放
 
 `agent/assistant-stream` 发布进程本地 start、瞬态 chunk 与 end frame。loop 会在 committed end frame 命名其类型和序号前追加完整的 `assistant/message` 或 `assistant/attempt`。abandoned end 没有 settlement。
 
-Web follow adapter 显式选择接收这些进程本地 frame，并为每个 start 补充当时观察到的最后一个持久序号。它把 chunk 呈现为持久 cursor 之间的 Client-only `assistant/live-chunk` update，只暂存 start 之后匹配的 settlement，并在 revision 缺口时重新打开 follow。committed end 会发布具名 settlement delta，删除该 attempt 的 transient match、加入持久 entry，并只重放受影响的 Conversation Context；abandoned end 会发布不含 entry 的同类 delta。重连 baseline 携带活跃 attempt 的持久起始 cursor 与紧凑前缀。分页历史、replay、遥测、token 记账与冷 UI 组装读取持久嵌入式 stream，而不是 live frame。
+Web follow adapter 显式选择接收这些进程本地 frame，并为每个 start 补充当时观察到的最后一个持久序号。它把 chunk 呈现为持久 cursor 之间的 Client-only `assistant/live-chunk` update，只暂存 start 之后匹配的 settlement，并在 revision 缺口时重新打开 follow。committed end 会发布具名 settlement delta，删除该 attempt 的 transient match、加入持久 entry，并只重放受影响的 Conversation Context；abandoned end 会发布不含 entry 的同类 delta。重连 baseline 携带活跃 attempt 的持久起始 cursor 与紧凑前缀。
+
+Client event source 原样传递持久 settlement。Chat 与 Trajectory 的 Assistant node 在 attempt 活跃期间折叠 `assistant/live-chunk`，直接从 `assistant/message` 构建 settled output，并且不为展示重放 `assistant/attempt` stream。因此冷恢复的 settled presentation 不会重建逐 token timing；其他消费方需要精确证据时仍可展开持久 stream。
 
 ### 已发布 v1 到 v2 迁移
 
-相邻迁移会校验完整的冻结 v1 产物，按 turn、step、terminal boundary 与精确 message provenance 对 chunk 分组，再为每个 attempt 替换一个 settlement。成功分组的 chunk 移入其 message。未被认领的分组会在最后一个被消费 chunk 的位置变成 `assistant/attempt`。无关的交错事件保持相对顺序，存活事件获得密集 v2 序号。该迁移边通过 `dsh-llm` 运行时的 `AssistantStreamAccumulator`、`expandAssistantStream` 与 `BlockAssembler` 压缩、展开并重组嵌入 stream，而不持有冻结副本，因为该包拥有 v2 stream 编码。目标校验会自行复核每个迁移后的 `assistant/message` 与其嵌入 stream 是否一致，因此不一致的 v1 日志会作为 unsupported migration 被拒绝并保留源产物，而不是由 installed Session restoration 报告为损坏。日后若某个格式改变 stream 编码，必须把这些 helper 的冻结副本纳入本迁移边。
+相邻迁移会校验完整的冻结 v1 产物，按 turn、step、terminal boundary 与精确 message provenance 对 chunk 分组，再为每个 attempt 替换一个 settlement。成功分组的 chunk 移入其 message。未被认领的分组会在最后一个被消费 chunk 的位置变成 `assistant/attempt`。无关的交错事件保持相对顺序，存活事件获得密集 v2 序号。该迁移边通过 `dsh-llm` 运行时的 `AssistantStreamAccumulator` 压缩嵌入 stream，而不持有冻结副本，因为该包拥有 v2 stream 编码。隔离的 publication verifier 通过 `expandAssistantStream()` 与 `BlockAssembler` 展开并重组写入后的 stream，并在发布前检查每个迁移后的 `assistant/message` 是否与其一致。日后若某个格式改变 stream 编码，必须把这些 helper 的冻结副本纳入本迁移边。
 
 该迁移边会重映射有限的已声明引用清单：信封 provenance、surface replacement 端点、command source event、compaction range 与 shadowed list，以及 title message list。经过校验的 `session/title-llm-request` 模型可见文本会在源序号命名空间中保持逐字节不变，而它的 `messageSeqs` 字段会迁移到 v2 命名空间；因此目标校验不会根据重映射后的序号重建该文本。指向被消费 chunk 的引用会使迁移失败；它绝不会被重定向到含义不同的 settlement。该迁移边也会拒绝切开 attempt 的继承切点。
 
@@ -49,7 +51,7 @@ Generation 选择与发布遵循[已发布 Session 迁移决策](2026-08-31-rele
 
 合并前的 performance acceptance 在三轮、100 组 warmup pair 与 600 组 measured pair 下，针对同一批已经解析的物理 row，把静态 catalog routing 与直接 released-v2 restoration 比较；它不比较 v1 与 v2，也不计入 backend I/O。每个 pooled median 与 p95 regression 都保持在 5% 预算以内，最差 p95 regression 为 3.150%。
 
-Agent-loop 测试固定先持久后 end 的顺序、中断的可见前缀、失败与重试 attempt、abandonment、usage 与 replay metadata。Session Controller 与 Conversation 测试固定实时瞬态显示、重连 baseline、committed settlement 发布、历史回放以及 Chat 与 Trajectory 一致性；TypeScript 与 Python SDK snapshot 固定外部事件表示。
+Agent-loop 测试固定先持久后 end 的顺序、中断的可见前缀、失败与重试 attempt、abandonment、usage 与 replay metadata。Session Controller 与 Conversation 测试固定实时瞬态显示、重连 baseline、committed settlement 发布与历史回放。Chat 与 Trajectory 测试固定实时 partial 展示和最终 message 的直接投影；TypeScript 与 Python SDK snapshot 固定外部事件表示。
 
 ## 备选方案
 
@@ -59,13 +61,15 @@ Agent-loop 测试固定先持久后 end 的顺序、中断的可见前缀、失
 
 **通过历史 API 传递 packed chunk row。** 这会减少 v1 的 wire 与 Client 工作，却让 Client 拥有第二套事件词汇，并让传输继续与 token-row 基数耦合。当前 API 携带标量持久 settlement，并使用独立的实时瞬态 stream。
 
+**在 Session Controller 中删除嵌入式 stream。** 这会减少 Client 保留的内存，却会引入第二种持久事件类型，并让面向传输的 owner 决定展示消费方需要哪些证据。实测瓶颈来自重复展开，因此由各 UI 消费方决定是否检查原样传递的 settlement。
+
 **把 stream 存在 sidecar 或 replay-only fixture 中。** 这会把一个 attempt 的 message 与证据拆给不同持久性 owner，也无法让普通恢复 Session 获得相同的失败输出与时间事实。settlement 是原子 owner。
 
 **把被消费 chunk 的引用重定向到其 settlement。** Chunk 与 attempt settlement 不是可互换事实。拒绝可以防止迁移悄然改变插件自有引用的含义。
 
 ## 后果
 
-当前日志、遥测、历史页与冷 Client 组装按模型 attempt 而非 token chunk 扩展，同时在每个 settlement 内保留精确 stream 证据。实时呈现保持增量，并且有意仅存在于进程内。
+当前日志、遥测与历史页按模型 attempt 而非 token chunk 扩展，同时在每个 settlement 内保留精确 stream 证据。Client event window 保留这份紧凑证据，但 Chat 与 Trajectory 的 Assistant node 不会把 settled stream 展开成逐 delta 对象。实时呈现保持增量，并且有意仅存在于进程内。
 
 v1 的顶层 chunk 可能在 attempt 结束前由带缓冲的持久化 writer 刷盘；与之不同，v2 在 settlement 之前没有持久 attempt 证据。如果进程或主机在 settlement 前硬中断，完整的 in-flight stream 都会丢失；`agent/assistant-stream` 不是 write-ahead log。这项取舍避免为实时输出增加第二个持久性 owner。
 

.agents/notes/implemented/architecture/2026-09-05-read-only-session-migration-preparation.i18n.yaml

@@ -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/architecture/2026-09-05-read-only-session-migration-preparation.md
+2026-09-05-read-only-session-migration-preparation.md: ab23e7e61dcdf0762cae6185de5fd16c4070fcbf
+2026-09-05-read-only-session-migration-preparation.zh.md: 89377d4d84776bebbc6d2ca6acea92ac15f74df3

.agents/notes/implemented/architecture/2026-09-05-read-only-session-migration-preparation.md

@@ -0,0 +1,170 @@
+# Agent Note: Historical Session reads prepare before write publication
+
+Status: implemented
+
+English | [中文](2026-09-05-read-only-session-migration-preparation.zh.md)
+
+## Problem
+
+The stateful Stage pipeline makes historical Decode and migration bounded and fast, but a serial persistence open still performs encode, sync, Worker verification, publication, and committed reopen before returning either handle kind. A read-only consumer therefore waits for about 2.2 seconds of work that it does not need and mutates storage merely to display history.
+
+### Serial readiness cost
+
+- History pagination, projection preparation, export, and the opening `session.follow` snapshot need only the validated current logical artifact.
+- A historical read open nevertheless creates and syncs a temporary v2 generation, starts a full verification Worker, rechecks the source, publishes v2, and reopens the target.
+- The migration result already exists in memory before encode, but the serial API returns only the committed physical snapshot. Persistence must Decode current bytes again to reconstruct the same logical events.
+- `session.follow` cannot deliver its opening snapshot until publication completes, even though Agent resume is the first operation that requires append access.
+- Read-only storage cannot serve a logically valid historical Session because read open requires generation publication.
+
+### A naive split would break lifecycle guarantees
+
+- Returning a write handle before verification would route append into an unpublished temporary file and create a second durability state for accepted events.
+- Starting publication automatically after every read would require backend ownership for task failure, shutdown, cleanup, and a later writer joining work it did not request.
+- A shared preparation cannot inherit the first caller's AbortSignal. One cancelled reader must not terminate work still awaited by another.
+- A read handle must initially serve prepared memory but later observe a current file and its appended tail after another caller publishes.
+- Once readers have observed one prepared artifact, source drift cannot silently rerun migration and substitute a different logical history.
+
+## Decision
+
+The JSONL backend separates logical preparation from durable publication. Read open waits only for preparation. Write open reuses a matching preparation and waits for publication before returning a writable handle.
+
+### Prepared generation API
+
+```text
+interface PreparedJsonlMigration {
+  readonly sourceIdentity: JsonlPhysicalIdentity
+  readonly artifact: SessionFormatArtifact
+  publish(): Promise<JsonlPhysicalIdentity>
+}
+```
+
+`prepareJsonlMigration()` reads one stable historical revision, runs the complete Stage chain once, and returns the current artifact without encoding or writing. `publish()` is idempotent: concurrent and later calls share one terminal Promise, including its rejection, and cannot encode the same prepared artifact twice.
+
+`publish()` streams current records into an exclusively created same-directory temporary file, syncs it, awaits the bounded Worker verifier, compares the source identity captured by preparation, and publishes the canonical path without overwrite. The successful publisher reuses the prepared logical artifact instead of decoding its target. A losing publisher verifies that the winner begins with the exact staged migration prefix; append tail validation remains a current-reader responsibility.
+
+Publication runs to settlement after invocation and is not cancelled midway by the write caller. Write open checks its caller signal before and after publication, so an abort can reject the open after the successor commits without leaking the write lease. A source identity change throws `JsonlGenerationSourceChangedError`, removes the temporary file, and does not repeat Decode or migration.
+
+### Preparation ownership and cancellation
+
+The persistence backend keeps one in-flight entry per Session id, selected source path, and stat-derived revision:
+
+```text
+interface MigrationPreparation {
+  sourcePath: string
+  sourceRevision: SessionPersistenceRevision
+  controller: AbortController
+  promise: Promise<PreparedStoredLog>
+  settled: boolean
+  waiters: number
+}
+```
+
+A new read or write open joins the existing entry only when its source path and revision still match. `waitWithAbort()` races each caller's AbortSignal against the shared Promise without forwarding that signal to shared work. The backend-owned controller is aborted only when the last waiter leaves while preparation is still running. The cancellation test pauses the physical read and observes two registered waiters before aborting one caller; an event-loop yield alone cannot establish admission after asynchronous path and revision lookup.
+
+Completed results enter the existing bounded `coldLogMemo`. The `StoredLog` discriminant separates published current state from `PreparedStoredLog`, whose `publication` field binds current logical events to their matching publication operation. A query followed by Agent resume therefore reuses the same Decode and migration result. The in-flight map owns only running work; it is not a second completed-result cache.
+
+`SessionHandle.read()` reports whether its event values are detached or shared-frozen. The JSONL backend deep-freezes each decoded event graph once before memoization and creates the `shared-frozen` result there; later reads and slices preserve that producer-established state even when the slice is empty. `readColdSessionLog()` combines those values with locally owned interrupted-turn closers and passes the `eventState` through `SessionObservationReader`; `Session.fromRestore()` validates and adopts the seed without copying or freezing. Ordinary create and fork seeds keep their defensive snapshot path.
+
+Read-only restoration validates the event and settlement fields required by Session runtime behavior but does not expand every embedded Assistant stream. The publication Worker retains complete stream replay and checks content, usage, and replay-state agreement before a migrated successor is committed. Existing current-v2 files rely on their writer; consumers that expand a compact stream validate its records when they read it.
+
+### Read handle transition
+
+A read open adopts a handle with prepared events in `state.primed` while no current generation exists. Each later `read()` resolves the current path:
+
+```text
+if current generation is absent:
+  return slice of primed events
+else:
+  clear primed events
+  read current generation and enforce non-shrinking history
+```
+
+`resolveCurrentLog()` may therefore return `undefined` for an existing historical Session: it answers whether a current canonical file exists, not whether the Session can be read. Public `stat` and `list` continue to discover the historical header.
+
+### Write-open publication
+
+Write open acquires the process-local claim and kernel-backed cross-process lease before re-resolving the selected generation. If it remains historical, it obtains or reuses the prepared `StoredLog` and awaits `publish()`. Only then does it return a write handle primed with the prepared events.
+
+```text
+write open
+  → claim process-local ownership
+  → acquire SessionWriteLease
+  → re-resolve generation
+  → join or create preparation
+  → encode + sync temp
+  → Worker verify
+  → source identity check
+  → no-overwrite publish
+  → return writable handle
+```
+
+No external caller can append before the handle exists. `append`, `flush`, and `close` therefore retain their ordinary current-generation behavior and never need a “publishing” branch. Service `flush()` continues to flush only already adopted writers; it does not turn a read-only preparation into a write.
+
+### Follow and Agent promotion
+
+`session.follow` opens history through the read path, restores the Session and projections, emits the opening snapshot, and then starts Agent promotion. Agent resume uses write open, so it waits for publication before the Agent accepts a new turn. History visibility and write readiness are separate timing points without introducing an unpublished append state.
+
+## Problem-to-solution mapping
+
+| Serial-flow problem | Implemented mechanism | Guarantee |
+|---|---|---|
+| Read-only callers wait for encode and verify | Read open returns prepared events | First content waits only for Decode and migration |
+| Concurrent historical opens repeat work | Session/source-revision keyed single-flight | One migration per selected revision |
+| First caller owns shared cancellation | Caller-local `waitWithAbort()` plus backend controller | One cancellation does not kill other waiters |
+| Preparation is lost between query and resume | `PreparedStoredLog.publication` in bounded memo | Write open reuses the same artifact |
+| No current path exists for a read handle | Primed in-memory read | Historical data is readable before publication |
+| Read handle must observe later append | Re-resolve and switch from primed data to current file | Existing handles converge after publication |
+| Append before verification is unsafe | Publish inside write open before returning the handle | Returned writer is immediately durable-ready |
+| Automatic background publication has no owner | Only write open invokes `publish()` | No orphan write task from read-only access |
+| Source changes after readers saw the artifact | Fail publication without rerunning migration | Exposed logical history is never silently replaced |
+
+## Verification
+
+The benchmark uses the same 116,228,655-byte v0 Zstandard Session as the Stage decision. The first table compares every relevant implementation; the detailed scheduling comparison then holds the Codec/Stage chain constant between #3585 and preparation-first scheduling.
+
+### First opening of historical data
+
+| Implementation | Session restored | CPU | Peak RSS | Retained heap | Result |
+|---|---:|---:|---:|---:|---|
+| Original high-performance v0 reader | 4.594s | 6.048s | 2.720GB | 2.016GB | Reads about 9.14 million v0 events without migration |
+| Master whole-artifact v0-to-v2 migration | >72.8s | — | Decode stage reached at least 7.219GB | — | OOM before returning a handle |
+| #3585 streaming migration with serial publication | 6.241s | 8.493s | 2.107GB | 477MB | Produces and publishes a 72,784-event v2 Session |
+| #3586 preparation-first scheduling | 2.954s | — | 1.026GB | 463MB | Produces the same v2 Session and defers publication until write open |
+
+Preparation-first restoration is 53% faster than #3585 and 36% faster than the original high-performance reader even though it also migrates the artifact to v2.
+
+### Scheduling observation points
+
+| User-visible point | Serial publication | Preparation-first | Change |
+|---|---:|---:|---:|
+| Read open plus Session restoration | 6.241s | 2.954s | -53% |
+| `session.follow` opening snapshot | 7.587s | 2.912s | -62% |
+| Agent receives writable Session | 6.246s | 5.161s | -17% |
+| Reopen an already-current v2 Session | 1.284s | 0.964s | -25% |
+| Follow opening-snapshot peak RSS | 2.353GB | 1.059GB | -55% |
+
+Preparation spends about 2.61 seconds in Decode and migration. Deferred publication takes about 2.56 seconds: 0.83 seconds for encode/write/sync, 1.72 seconds for strict Worker verification, and about 0.005 seconds for source check and atomic publication. A read-only request performs none of that publication work.
+
+The prepared artifact and Session restoration peak near 1.03 GB RSS. Preparation and Worker verification together peak near 2.19 GB because the parent retains the logical artifact while the Worker independently validates the physical generation.
+
+Tests cover shared-waiter cancellation, all-waiters cancellation, memo handoff, read-handle switching, source drift, winner collision, publication idempotence, write-open ordering, Worker failure, and the plain-Node bundled Worker entry.
+
+## Consequences
+
+Read-only body access does not publish a generation. The first writer pays publication once before append. A configured JSONL root must still be readable and structurally valid, but historical body migration itself does not require a successor write.
+
+The bounded memo retains one migrated event array to bridge read and write opens. This is intentional: avoiding that retained artifact would require a second Decode and migration or would prevent early read availability.
+
+Publication failure rejects Agent resume and other write opens but does not invalidate read results already delivered from the unchanged historical source. Source drift is terminal for that write attempt rather than a trigger to recompute hidden state.
+
+The backend still has a broader pre-existing lifecycle gap: dispose does not own every `create()` or `open()` operation that has not yet returned a handle. This decision does not add migration-specific tracking to `flush()` or solve that general pending-operation problem.
+
+## Alternatives considered
+
+- **Keep serial publication for every open** — is the simplest physical state model but adds about 2.2 seconds to read-only first content and requires writable storage.
+- **Publish automatically in the background after read** — needs backend task ownership, shutdown quiescence, error reporting, and writer joining even when no caller requested a write.
+- **Return a writer before verification** — requires append to an unpublished stage and creates an additional durability and failure state for accepted events.
+- **Give each caller an independent preparation** — repeats the dominant Decode and migration work and multiplies peak memory under concurrent list/follow/resume operations.
+- **Let the first caller's signal cancel shared work** — makes later callers depend on unrelated cancellation timing.
+- **Rerun migration after source drift** — can replace history already shown to readers and makes one logical operation process the same large file more than once.
+- **Always keep read handles on primed memory** — prevents an existing handle from seeing later append and diverges from ordinary persistence refresh behavior.

.agents/notes/implemented/architecture/2026-09-05-read-only-session-migration-preparation.zh.md

@@ -0,0 +1,170 @@
+# Agent Note: 历史 Session 在写入发布前提供只读迁移结果
+
+Status: implemented
+
+[English](2026-09-05-read-only-session-migration-preparation.md) | 中文
+
+## 问题
+
+有状态 Stage pipeline 已经把历史 Decode 与 migration 恢复到有界、高性能的数据流，但串行 persistence open 仍会在返回任一种 handle 前执行 encode、sync、Worker verification、publication 与 committed reopen。只读 consumer 因此需要额外等待约 2.2 秒不需要的工作，而且仅为展示历史就会修改存储。
+
+### 串行 readable 的额外代价
+
+- 历史分页、projection preparation、export 与 `session.follow` 的 opening snapshot 只需要已经校验的 current logical artifact。
+- Historical read open 仍会创建并 sync 临时 v2 generation、启动完整 verification Worker、复查 source、发布 v2 并重新打开 target。
+- Migration 在 encode 前已经得到完整 current artifact，但串行 API 只返回 committed physical snapshot。Persistence 必须再次 Decode current bytes 才能重建相同逻辑事件。
+- `session.follow` 必须等 publication 完成后才能发出 opening snapshot，而 Agent resume 才是第一个真正要求 append 权限的操作。
+- Read-only storage 无法提供逻辑上有效的历史 Session，因为 read open 强制发布 generation。
+
+### 直接拆分会破坏 lifecycle 保证
+
+- Verify 前返回 write handle 会让 append 写入 unpublished temporary file，并为已接纳事件引入第二种 durability state。
+- 每次 read 后自动启动 publication，需要 backend 负责 task failure、shutdown、cleanup，以及后续 writer 加入一个自己没有请求的任务。
+- Shared preparation 不能继承第一个 caller 的 AbortSignal；一个 reader 取消不能终止其他 waiter 仍依赖的工作。
+- Read handle 必须先提供 prepared memory，并在其他 caller 发布后切换到 current file 与其 append tail。
+- Reader 已经观察一个 prepared artifact 后，source drift 不能悄悄重跑 migration 并替换成另一份逻辑历史。
+
+## 决策
+
+JSONL backend 将 logical preparation 与 durable publication 分开。Read open 只等待 preparation；write open 复用 matching preparation，并在返回 writable handle 前等待 publication。
+
+### Prepared generation API
+
+```text
+interface PreparedJsonlMigration {
+  readonly sourceIdentity: JsonlPhysicalIdentity
+  readonly artifact: SessionFormatArtifact
+  publish(): Promise<JsonlPhysicalIdentity>
+}
+```
+
+`prepareJsonlMigration()` 读取一个稳定 historical revision，只执行一次完整 Stage chain，并在不 encode、不写文件的情况下返回 current artifact。`publish()` 是幂等操作：并发与后续调用共享同一个终态 Promise（包括拒绝结果），不会对同一 prepared artifact 重复 encode。
+
+`publish()` 把 current records 流式写入同目录排他创建的 temporary file，执行 sync，等待 bounded Worker verifier，比较 preparation 捕获的 source identity，再通过 no-overwrite 操作发布 canonical path。成功 publisher 复用 prepared logical artifact，不重新 Decode 自己的 target。竞争失败者只验证 winner 以精确 staged migration prefix 开头；append tail validation 仍属于 current reader。
+
+Publication 调用后会运行到 settlement，不会被 write caller 中途取消。Write open 会在 publication 前后检查 caller signal，因此取消可能在后继已经提交后拒绝 open，但不会泄漏 write lease。Source identity 变化会抛出 `JsonlGenerationSourceChangedError`、删除临时文件，并且不会重复 Decode 或 migration。
+
+### Preparation ownership 与取消
+
+Persistence backend 按 Session id、selected source path 与 stat-derived revision 保存一个 in-flight entry：
+
+```text
+interface MigrationPreparation {
+  sourcePath: string
+  sourceRevision: SessionPersistenceRevision
+  controller: AbortController
+  promise: Promise<PreparedStoredLog>
+  settled: boolean
+  waiters: number
+}
+```
+
+新的 read/write open 只有在 source path 与 revision 仍匹配时才加入已有 entry。`waitWithAbort()` 让每个 caller 的 AbortSignal 与 shared Promise 竞争，但不会把 caller signal 传给共享工作。只有最后一个 waiter 在 preparation 仍运行时离开，backend-owned controller 才会 abort。取消测试暂停物理读取，并在取消一个 caller 前观察到两个已注册的 waiter；仅让出一次事件循环不能证明异步路径与 revision 查找后的加入已经完成。
+
+完成结果进入既有 bounded `coldLogMemo`。`StoredLog` 判别字段把已发布 current state 与 `PreparedStoredLog` 分开，后者的 `publication` 字段把 current logical events 与匹配的 publication operation 绑定，使 query 后紧接的 Agent resume 复用同一次 Decode 与 migration。In-flight map 只拥有运行中的工作，不是第二个 completed-result cache。
+
+`SessionHandle.read()` 会报告 event value 是 detached 还是 shared-frozen。JSONL backend 在 memo 化前只对每个已解码 event graph 深度冻结一次，并在该处构造 `shared-frozen` 结果；后续读取和 slice 即使为空也会保留生产者建立的状态。`readColdSessionLog()` 将这些 event 与本地独占的 interrupted-turn closer 组合，并通过 `SessionObservationReader` 继续传递 `eventState`；`Session.fromRestore()` 只校验和接管 seed，不再复制或冻结。普通 create 与 fork seed 继续使用 defensive snapshot 路径。
+
+Read-only restoration 会校验 Session runtime 直接依赖的 event 与 settlement 字段，但不会展开每一段嵌入式 Assistant stream。Publication Worker 继续执行完整 stream replay，并在提交 migrated successor 前校验 content、usage 与 replay state 一致性。已有 current-v2 文件信任其 writer；需要展开 compact stream 的 consumer 会在读取时校验 record。
+
+### Read handle 切换
+
+Current generation 不存在时，read open 会采用在 `state.primed` 中保存 prepared events 的 handle。后续每次 `read()` 都重新解析 current path：
+
+```text
+if current generation is absent:
+  return slice of primed events
+else:
+  clear primed events
+  read current generation and enforce non-shrinking history
+```
+
+因此，一个已有 historical Session 也可能让 `resolveCurrentLog()` 返回 `undefined`：它回答的是 current canonical file 是否存在，而不是 Session 是否可读。公开 `stat` 与 `list` 继续发现 historical header。
+
+### Write-open publication
+
+Write open 先取得进程内 claim 与内核支持的跨进程 lease，再重新解析 selected generation。如果它仍是 historical，就取得或复用 prepared `StoredLog` 并等待 `publish()`。之后才返回以 prepared events 为 primed state 的 write handle。
+
+```text
+write open
+  → claim process-local ownership
+  → acquire SessionWriteLease
+  → re-resolve generation
+  → join or create preparation
+  → encode + sync temp
+  → Worker verify
+  → source identity check
+  → no-overwrite publish
+  → return writable handle
+```
+
+Handle 返回前，外部 caller 无法 append。因此 `append`、`flush` 与 `close` 保持普通 current-generation 行为，不需要“publishing”分支。Service `flush()` 继续只 flush 已经 adopt 的 writer；它不会把 read-only preparation 转成 write。
+
+### Follow 与 Agent promotion
+
+`session.follow` 通过 read path 打开历史、恢复 Session 与 projections、发出 opening snapshot，然后启动 Agent promotion。Agent resume 使用 write open，因此会在 Agent 接收新一轮对话前等待 publication。历史可见与写入就绪成为两个明确时间点，同时不引入 unpublished append state。
+
+## 问题与方案对照
+
+| 串行流程问题 | 实现机制 | 保证 |
+|---|---|---|
+| Read-only caller 等待 encode 与 verify | Read open 返回 prepared events | 首屏只等待 Decode + migration |
+| 并发 historical open 重复工作 | Session/source-revision keyed single-flight | 每个 selected revision 只迁移一次 |
+| 第一个 caller 拥有共享取消 | Caller-local `waitWithAbort()` + backend controller | 单个取消不终止其他 waiter |
+| Query 与 resume 之间丢失 preparation | Bounded memo 中的 `PreparedStoredLog.publication` | Write open 复用相同 artifact |
+| Read handle 没有 current path | Primed in-memory read | Publication 前 historical data 可读 |
+| Read handle 需要观察后续 append | 重新 resolve，并从 primed data 切到 current file | Publication 后已有 handle 收敛 |
+| Verify 前 append 不安全 | Write open 返回前完成 publication | 返回 writer 立即具备普通 durability |
+| 自动后台 publication 无 owner | 只有 write open 调用 `publish()` | Read-only access 不产生 orphan write task |
+| Reader 已看到 artifact 后 source 改变 | Publication 失败且不重跑 migration | 已暴露逻辑历史不被静默替换 |
+
+## 验证
+
+Benchmark 使用 Stage 决策中的同一份 116,228,655-byte v0 Zstandard Session。第一张表比较 migration 工作涉及的全部实现；后续调度明细则保持 #3585 与 preparation-first 使用同一条 Codec/Stage chain，仅改变 persistence 调度。
+
+### 用户首次打开历史数据
+
+| 实现 | Session restore | CPU | Peak RSS | Retained heap | 结果 |
+|---|---:|---:|---:|---:|---|
+| 原高性能 v0 reader | 4.594s | 6.048s | 2.720GB | 2.016GB | 不迁移，读取约 914 万个 v0 event |
+| Master whole-artifact v0-to-v2 migration | >72.8s | — | Decode 阶段达到至少 7.219GB | — | 返回 handle 前 OOM |
+| #3585 streaming migration + 串行 publication | 6.241s | 8.493s | 2.107GB | 477MB | 生成并发布包含 72,784 个 event 的 v2 Session |
+| #3586 preparation-first 调度 | 2.954s | — | 1.026GB | 463MB | 生成相同 v2 Session，并把 publication 延迟到 write open |
+
+Preparation-first restore 比 #3585 快 53%，也比原高性能 reader 快 36%，同时仍然完成 artifact 到 v2 的 migration。
+
+### 调度观测点
+
+| 用户观测点 | 串行 publication | Preparation-first | 变化 |
+|---|---:|---:|---:|
+| Read open + Session restore | 6.241s | 2.954s | -53% |
+| `session.follow` opening snapshot | 7.587s | 2.912s | -62% |
+| Agent 得到 writable Session | 6.246s | 5.161s | -17% |
+| 已是 current v2 的再次打开 | 1.284s | 0.964s | -25% |
+| Follow opening-snapshot peak RSS | 2.353GB | 1.059GB | -55% |
+
+Preparation 中约 2.61 秒用于 Decode 与 migration。延后的 publication 约为 2.56 秒：encode/write/sync 0.83 秒、严格 Worker verification 1.72 秒、source check 与 atomic publication 约 0.005 秒。Read-only 请求完全不执行这段 publication。
+
+Prepared artifact 与 Session restore 的 peak RSS 约为 1.03 GB。Preparation 与 Worker verification 同时存在时峰值约 2.19 GB，因为 parent 保留 logical artifact，而 Worker 独立校验 physical generation。
+
+测试覆盖 shared-waiter cancellation、all-waiter cancellation、memo handoff、read-handle switching、source drift、winner collision、publication idempotence、write-open ordering、Worker failure 与 plain-Node bundled Worker entry。
+
+## 后果
+
+Read-only body access 不发布 generation。第一个 writer 会在 append 前支付一次 publication。已配置的 JSONL root 仍必须可读且结构有效，但 historical body migration 本身不要求写 successor。
+
+Bounded memo 会保留一份 migrated event array，用于连接 read 与 write open。这是有意的取舍：不保留该 artifact 就必须重复 Decode 与 migration，或者无法提前提供 read。
+
+Publication failure 会拒绝 Agent resume 和其他 write open，但不会使已经从 unchanged historical source 交付的 read result 失效。Source drift 对该 write attempt 是 terminal failure，不会触发 hidden state 重算。
+
+Backend 仍存在一个更广泛的既有 lifecycle 缺口：dispose 不拥有每个尚未返回 handle 的 `create()` 或 `open()` operation。本决策不会向 `flush()` 增加 migration-specific tracking，也不解决通用 pending-operation 问题。
+
+## 考虑过的替代方案
+
+- **每个 open 都保持串行 publication**——physical state 最简单，但让 read-only 首屏多等待约 2.2 秒并要求存储可写。
+- **Read 后自动后台 publish**——需要 backend task ownership、shutdown quiescence、error reporting，以及 writer 加入一个没有 caller 请求的任务。
+- **Verify 前返回 writer**——要求 append 写入 unpublished stage，并为已接纳事件增加一种 durability 与 failure state。
+- **每个 caller 独立 preparation**——重复最重的 Decode 与 migration，并在 list/follow/resume 并发时放大峰值内存。
+- **让第一个 caller signal 取消共享工作**——使后续 caller 依赖无关的 cancellation timing。
+- **Source drift 后重跑 migration**——可能替换已经展示给 reader 的历史，也会让一次逻辑 operation 重复处理同一大文件。
+- **Read handle 永远停留在 primed memory**——无法观察后续 append，并偏离普通 persistence refresh 行为。

.agents/notes/implemented/architecture/2026-09-06-embedded-stream-record-readers.i18n.yaml

@@ -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/architecture/2026-09-06-embedded-stream-record-readers.md
+2026-09-06-embedded-stream-record-readers.md: 972fee634833cef5fd7b0a54f69780b9370f0cc3
+2026-09-06-embedded-stream-record-readers.zh.md: faba6e179887a9943926f2c73aa8e42a10cee300

.agents/notes/implemented/architecture/2026-09-06-embedded-stream-record-readers.md

@@ -0,0 +1,50 @@
+# Agent Note: Embedded Assistant stream consumers read compact records
+
+Status: implemented
+
+English | [中文](2026-09-06-embedded-stream-record-readers.zh.md)
+
+## Problem
+
+Session format v2 embeds each model attempt's compact stream (`AssistantStreamRecord[]`: packed `text-chunks`, `reasoning-chunks`, and `tool-call-chunks` runs plus timestamped raw `chunk` records) in `assistant/message` and `assistant/attempt`. Consumers that folded those settlements called `expandAssistantStream()` first; it materializes the complete per-member array, so a consumer that needs one fact (`find` on the first token, the last usage chunk, a joined text, one block-end) paid O(members) allocation and time: about two objects per member on top of the compact form.
+
+After v2 embedded streams settlement widened with the message content and Chat and Trajectory sections settled directly from it, the remaining expand consumers are the Host and client folds: Session Stats reads the first-token time per `assistant/attempt` and `assistant/message` (the projection phase of every Session open), the token meter rebuilds provider content and scans every stream for its last usage chunk (the projection unit still scans to the end), the subagent output fold joins plain text, and the Session Controller image lookup scans for block-end chunks.
+
+## Decision
+
+`@deepseek-ai/dsh-llm` answers consumer questions directly from compact records; every remaining consumer folds records once with early exit.
+
+`packages/llm/llm/src/assistant-stream.ts` exports record-level readers beside the accumulator and `expandAssistantStream`:
+
+- Chunk rules: `isTokenDelta` (non-empty text, reasoning, or Tool-call arguments fragment, or any name-bearing Tool-call delta), `isVisibleChunk` (non-whitespace text or reasoning, or a block start or end of any kind other than text, reasoning, or Tool call), and `chunkHasVisibleText` (non-whitespace text delta or completed text block).
+- Run readers: `runFirstTokenTime` and `runFirstVisibleTime` reconstruct the first qualifying member's time from `time0` and the `dt` gaps and stop scanning there; a name-bearing Tool-call run yields `time0` without reading a fragment.
+- Stream readers: `assistantStreamFirstTokenTime`, `assistantStreamHasVisibleContent`, `assistantStreamHasVisibleText`, `lastAssistantStreamChunk(stream, type)` (backward scan), `assistantStreamChunks(stream, type)`, `joinAssistantStreamText`, and `assembleAssistantStream`, which feeds a `BlockAssembler` one joined delta per run (assembly only concatenates, so blocks, usage, finish, and replay state equal the per-member result). `RawStreamChunkType` excludes the delta types, so a raw-chunk lookup can never silently skip packed members.
+
+Session Stats reads `assistantStreamFirstTokenTime`; the token meter reads `lastAssistantStreamChunk(stream, 'usage')` and assembles provider output through `assembleAssistantStream`; the subagent output fold appends `joinAssistantStreamText`; the Session Controller scans `assistantStreamChunks(stream, 'block-end')` for images.
+
+`expandAssistantStream` keeps its strict validation and its remaining callers, which need every member or validate the stream at a durable boundary: Session restore validation, the v1-to-v2 migration validator and publication Worker replay, the reconnect baseline, and test support.
+
+### Measurements
+
+The repo's synthetic first-open benchmark (200 turns, 127,400 released-v0 events, 500,000 streamed deltas in 1,600 compact records; five samples, median):
+
+| Phase | Before | After |
+|---|---|---|
+| first-open projection | 28.0 ms | 5.9 ms |
+| first-open total | 76.9 ms | 53.8 ms |
+| first-open peak RSS | 137.2 MB | 94.6 MB |
+| reopen projection | 17.8 ms | 6.5 ms |
+
+Open, read, and restore phases are unchanged; the reader keeps the same first-token time by construction (the first qualifying member is the first record's first qualifying fragment, and the deltas stay ordered).
+
+## Alternatives considered
+
+**Memoize `expandAssistantStream` per input array.** Expanding all streams once costs tens of milliseconds, but retaining the expansions costs about ten times the compact stream for the event's lifetime — a permanent version of the transient allocation the change removes. The readers remove the need for retained expansions entirely.
+
+**Keep the per-member fold.** Early-exit `.find` still materializes the whole array first, so the allocation and O(members) time remain.
+
+## Consequences
+
+Host and Client folds of an embedded settlement cost O(records) plus one join per run, and no consumer materializes members unless it validates at a durable boundary or needs every member. The token, visibility, and visible-text rules have one home in `dsh-llm`, so a record reader and the accumulator's packing rules cannot drift apart.
+
+Publication verification (`assertCurrentAssistantStreams`) still replays every settlement at publish time; because it must prove content-by-chunk agreement, converting it to run-aware assembly without member materialization remains open work.

.agents/notes/implemented/architecture/2026-09-06-embedded-stream-record-readers.zh.md

@@ -0,0 +1,50 @@
+# Agent Note: 内嵌 Assistant 流的消费方直接读取紧凑记录
+
+Status: implemented
+
+[English](2026-09-06-embedded-stream-record-readers.md) | 中文
+
+## 问题
+
+Session 格式 v2 将每次模型尝试的紧凑流（`AssistantStreamRecord[]`：打包的 `text-chunks`、`reasoning-chunks`、`tool-call-chunks` run 加上带时间戳的原始 `chunk` 记录）嵌入 `assistant/message` 与 `assistant/attempt`。折叠这些 settlement 的消费方会先调用 `expandAssistantStream()`；它会物化完整的逐成员数组，因此只需一个事实的消费方（find 首个 token、最后一个 usage chunk、拼接文本、一个 block-end）也要付出 O(members) 的分配与时间：在紧凑形式之上每个成员约两个对象。
+
+在 v2 内嵌流 settlement 随消息内容扩展、Chat 与 Trajectory 区块直接由内容结算之后，剩余的 expand 消费方是 Host 与客户端折叠：Session Stats 读取每个 `assistant/attempt` 与 `assistant/message` 的首 token 时间（每次打开 Session 的 projection 阶段）、token 计量重建提供商内容并扫描每个流到最后一个 usage chunk（projection 单元仍扫描到末尾）、子代理输出折叠拼接纯文本、Session Controller 镜像查找扫描 block-end chunk。
+
+## 决策
+
+`@deepseek-ai/dsh-llm` 直接从紧凑记录回答消费方问题；剩余消费方对记录做一次带提前退出的折叠。
+
+`packages/llm/llm/src/assistant-stream.ts` 在累加器与 `expandAssistantStream` 之外导出记录级读取器：
+
+- Chunk 规则：`isTokenDelta`（非空文本、reasoning 或 Tool-call 参数片段，或任何带名称的 Tool-call delta）、`isVisibleChunk`（非空白文本或 reasoning，或 text/reasoning/Tool call 之外的任意块开始或结束）、`chunkHasVisibleText`（非空白文本 delta 或完成的文本块）。
+- Run 读取器：`runFirstTokenTime` 与 `runFirstVisibleTime` 从 `time0` 与 `dt` 间隔重建首个合格成员的时间并停止扫描；带名称的 Tool-call run 直接产出 `time0`，不读片段。
+- 流读取器：`assistantStreamFirstTokenTime`、`assistantStreamHasVisibleContent`、`assistantStreamHasVisibleText`、`lastAssistantStreamChunk(stream, type)`（逆向扫描）、`assistantStreamChunks(stream, type)`、`joinAssistantStreamText` 与 `assembleAssistantStream`（每个 run 向 `BlockAssembler` 喂入一个拼接后的 delta；组装只做拼接，因此 blocks、usage、finish 与 replay state 与逐成员结果一致）。`RawStreamChunkType` 排除 delta 类型，因此原始 chunk 查找不可能静默跳过打包成员。
+
+Session Stats 读取 `assistantStreamFirstTokenTime`；token 计量读取 `lastAssistantStreamChunk(stream, 'usage')` 并通过 `assembleAssistantStream` 组装提供商输出；子代理输出折叠追加 `joinAssistantStreamText`；Session Controller 用 `assistantStreamChunks(stream, 'block-end')` 扫描镜像。
+
+`expandAssistantStream` 保留其严格校验与其余调用方（需要每个成员或在持久边界校验流）：Session 恢复校验、v1-to-v2 迁移校验器与发布 Worker 重放、重连基线、测试支撑。
+
+### 测量
+
+仓库的合成 first-open 基准（200 循环、127,400 个 released-v0 事件、1,600 条紧凑记录中的 500,000 个流式 delta；五次采样取中位数）：
+
+| 阶段 | 之前 | 之后 |
+|---|---|---|
+| first-open projection | 28.0 ms | 5.9 ms |
+| first-open 总计 | 76.9 ms | 53.8 ms |
+| first-open 峰值 RSS | 137.2 MB | 94.6 MB |
+| reopen projection | 17.8 ms | 6.5 ms |
+
+Open、read、restore 阶段不变；读取器按构造保持相同的首 token 时间（首个合格成员即首条记录的首个合格片段，且 delta 保持有序）。
+
+## 备选方案
+
+**按输入数组记忆化 `expandAssistantStream`。** 展开全部流只需几十毫秒，但保留展开结果在事件生命周期内约花费紧凑流的十倍内存——这是本变更移除的瞬时分配的永久版本。读取器完全消除了对保留展开的需求。
+
+**保留逐成员折叠。** 提前退出的 `.find` 仍然先物化整个数组，因此分配与 O(members) 时间仍在。
+
+## 后果
+
+Host 与客户端折叠一次内嵌结算的代价为 O(records) 加每个 run 一次拼接，且除非在持久边界校验或需要每个成员，消费方不再物化成员。token、可见性与可见文本规则在 `dsh-llm` 中只有一处，因此记录读取器与累加器的打包规则不可能漂移。
+
+发布校验（`assertCurrentAssistantStreams`）仍在发布时重放每个 settlement；因为它必须按 chunk 证明内容一致，将其转为不入成员的 run 感知组装仍是未完成工作。

.agents/notes/implemented/bug-fix/2026-08-17-subagent-message-settlement-ordering.i18n.yaml

@@ -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/bug-fix/2026-08-17-subagent-message-settlement-ordering.md
-2026-08-17-subagent-message-settlement-ordering.md: cdc996643c84c5f50a3bd1836e82645660dc8c57
-2026-08-17-subagent-message-settlement-ordering.zh.md: 1143da1560e4969dcc4f6a0c6d5ca18060b56191
+2026-08-17-subagent-message-settlement-ordering.md: 7462a670766664745e46204dcb01e578b4f86219
+2026-08-17-subagent-message-settlement-ordering.zh.md: bba8fe0e7c543982176def4dbcc6f94dda204339

.agents/notes/implemented/bug-fix/2026-08-17-subagent-message-settlement-ordering.md

@@ -14,7 +14,7 @@ The child instruction says to send a finding whenever it changes what the parent
 
 Every model-authored adjacent-Agent message uses fixed Steer delivery through `SubagentRuntime.sendMessage()`. A running parent reads the child message at its nearest safe step boundary and an idle parent starts a turn. There is no quiet or next-turn model delivery option.
 
-The continuation manager retains `sendWaking()` and `admitWaking()` around messages delivered to resident continuable parents. Their purpose is waking-send admission accounting: the receiving Activation remains live between synchronous inbox insertion and the microtask that observes the wake.
+The continuation manager retains `sendWaking()` around messages delivered to resident continuable parents and routes the synchronous send through the parent's private `SubagentInbox`. The wrapper accepts the send before its closing promise is installed or rejects it afterwards, and an accepted attempt renews the Activation's wake generation before returning. The receiving Activation therefore cannot settle over an accepted waking send.
 
 ### Ordering across parent states
 

.agents/notes/implemented/bug-fix/2026-08-17-subagent-message-settlement-ordering.zh.md

@@ -14,7 +14,7 @@ child 指令要求在发现会改变 parent 下一步动作时发送该发现。
 
 每条模型编写的相邻 Agent 消息都通过 `SubagentRuntime.sendMessage()` 使用固定 Steer 投递。运行中的 parent 在最近安全 step 边界读取 child 消息，空闲 parent 则启动一个轮次。模型没有静默或 next-turn 投递选项。
 
-继续执行管理器在投递到驻留可继续 parent 的消息周围保留 `sendWaking()` 与 `admitWaking()`。它们负责唤醒发送准入记账：接收方 Activation 会在同步 inbox 插入与观察到唤醒的微任务之间保持在线。
+继续执行管理器会在投递到驻留可继续 parent 的消息周围保留 `sendWaking()`，并通过 parent 的私有 `SubagentInbox` 执行同步发送。包装层会在安装 closing promise 前接受发送，并在安装后拒绝发送；被接受的尝试会在返回前更新 Activation 的 wake generation。因此，接收方 Activation 不会越过一条已接受的唤醒发送完成结算。
 
 ### 不同 parent 状态下的顺序
 

.agents/notes/implemented/bug-fix/2026-09-05-session-reference-model-budget.i18n.yaml

@@ -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/bug-fix/2026-09-05-session-reference-model-budget.md
+2026-09-05-session-reference-model-budget.md: 0654f2b89187727668fddb97ab6e982bacd82dc1
+2026-09-05-session-reference-model-budget.zh.md: 84f4fee76c7fb54f3ec37f0ae3417cc56d772af7

.agents/notes/implemented/bug-fix/2026-09-05-session-reference-model-budget.md

@@ -0,0 +1,25 @@
+# Agent Note: Model-relative session-reference budgets
+
+Status: implemented
+
+English | [中文](2026-09-05-session-reference-model-budget.zh.md)
+
+## Problem
+
+A fixed 64 KiB reference budget discards useful source context on large-context models. The target session header describes a prior request, while agent options seed routing; neither necessarily identifies the model selected for the entering step.
+
+## Decision
+
+[Session-reference](../../../../packages/context/session-reference/README.md) observes the completed `system-prompt/assemble` waterfall with a local prepend listener and stores its provider/model pair in a WeakMap keyed by Agent. Preparation resolves that route through the optional LLM service; direct preparation before any assembly uses agent options. Diagnostics without an Agent do not update the map.
+
+Each source receives `max(65536, floor(contextWindow × 4 × referenceContextFraction))` bytes, with a default fraction of `0.2`. Four bytes per token is a sizing heuristic. Explicit `maxReferenceBytes` bypasses model lookup and remains exact. Missing route, service, adapter, or capacity retains the floor; other lookup failures and cancellation propagate. An absent adapter does not prevent stream middleware from serving the route.
+
+## Alternatives considered
+
+**Read the header or options for every step.** Either can select a stale model after a live switch. The completed assembly exposes the route captured by model selection.
+
+**Reassemble or redispatch request routing during pre-step.** These operations repeat plugin effects and can capture a different selection. A local observer needs neither loop changes nor another public routing API.
+
+## Consequences
+
+The budget grows with model capacity without changing projection, retention, or preview policy. It remains per source, not an aggregate token reservation. The listener is effect-owned and disposable; the map does not retain agents. Focused tests cover the floor, fractional conversion, explicit overrides, live selection, absent metadata, cancellation, lookup errors, and listener removal.

.agents/notes/implemented/bug-fix/2026-09-05-session-reference-model-budget.zh.md

@@ -0,0 +1,25 @@
+# Agent Note: 模型相对会话引用预算
+
+Status: implemented
+
+[English](2026-09-05-session-reference-model-budget.md) | 中文
+
+## Problem
+
+固定的 64 KiB 引用预算会在大上下文模型上丢弃有用的来源上下文。目标会话头描述上一次请求，而 agent options 为路由提供初始值；两者都不一定标识当前进入步骤所选的模型。
+
+## Decision
+
+[Session-reference](../../../../packages/context/session-reference/README.zh.md) 通过本地 prepend 监听器观察已完成的 `system-prompt/assemble` 瀑布，并把 provider/model 对存入以 Agent 为键的 WeakMap。准备阶段通过可选 LLM 服务解析该路由；首次组装前直接准备则使用 agent options。不带 Agent 的诊断不会更新映射。
+
+每个来源获得 `max(65536, floor(contextWindow × 4 × referenceContextFraction))` 字节，默认比例为 `0.2`。每个 token 四字节是容量估算。显式 `maxReferenceBytes` 跳过模型查询并保持精确值。缺少路由、服务、适配器或容量时保留下限；其他查询失败和取消会传播。缺少适配器不妨碍流中间件处理该路由。
+
+## Alternatives considered
+
+**每步读取会话头或 options。** 实时切换后，两者都可能选中旧模型。完成的组装公开模型选择所捕获的路由。
+
+**在 pre-step 中重新组装或重新分派请求路由。** 这些操作会重复插件效果，并可能捕获不同的选择。本地观察器不需要修改循环或增加公共路由 API。
+
+## Consequences
+
+预算随模型容量增长，不改变投影、保留或预览策略。它仍按来源计算，而不是聚合 token 预留。监听器由 effect 持有并可释放；映射不会保留 agent。聚焦测试覆盖下限、比例换算、显式覆盖、实时选择、元数据缺失、取消、查询错误与监听器移除。

.agents/notes/implemented/bug-fix/2026-09-05-session-reference-spill-reuse.i18n.yaml

@@ -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/bug-fix/2026-09-05-session-reference-spill-reuse.md
+2026-09-05-session-reference-spill-reuse.md: d9f9e3702075417a9b5c742831737e412e75bd00
+2026-09-05-session-reference-spill-reuse.zh.md: 346ed1a39ee65a560cf9e74e519ae45c2ce133e1

.agents/notes/implemented/bug-fix/2026-09-05-session-reference-spill-reuse.md

@@ -0,0 +1,41 @@
+# Agent Note: Reuse spill storage for truncated session references
+
+Status: implemented
+
+English | [中文](2026-09-05-session-reference-spill-reuse.zh.md)
+
+## Problem
+
+A bounded cross-session preview can omit whole messages or most of a retained message. A model that sees only the preview needs an accurate account of the omission and a way to inspect the captured text, without treating another session's instructions as current authority. Rereading the source later would not recover the same observation when the source advances or compacts.
+
+## Decision
+
+[Session-reference preparation](../../../../packages/context/session-reference/README.md) retains its existing preview policy and per-reference JSON byte budget. Each truncated reference attempts `saveText` through optional `ctx.get("spillStore")`; an untruncated reference writes no artifact. The full transcript and bounded preview derive from the same captured user/assistant text projection, including compaction checkpoints but excluding tools, reasoning, and other injected context. No second source read occurs.
+
+The artifact belongs to the target session receiving the context. Its descriptive source is `{ kind: "session-reference", sessionId, label }`, where `sessionId` identifies the referenced session. [Spill storage](../../../../packages/spill/spill/README.md) accepts this minimal alternative alongside the existing tool source; it requires no fabricated tool name or call id. Storage ownership does not authorize retrieval.
+
+A separate omission notice outside the bounded preview JSON records exact `omittedMessages` and `omittedBytes`. It carries the saved locator and backend `retrievalHint`, or an unavailable outcome distinguishing missing storage from a failed save. This notice is model-visible content in the same durable reference message, not metadata-only UI decoration. A tiny preview budget cannot remove it. The saved transcript carries capture metadata, including `capturedFormatVersion`, and the same untrusted-background warning as the preview. Per-message JSON string fragments contain at most 64 Unicode code points per line; decoding and concatenating them restores exact text, including original newlines. This fixed artifact format keeps long single-line middles retrievable with ordinary paged file reads without changing preview retention.
+
+Cancellation after an asynchronous save prevents context publication, even if storage already created the artifact. The consumer does not add rollback or deletion APIs; the existing backend expiry policy governs that artifact. Replay uses the logged preview and notice and never repeats the save or source read.
+
+## Alternatives considered
+
+**Write a separate session-reference file store.** Rejected because private naming, session-scoped ownership, locator guidance, and artifact lifetime already belong to spill storage. A second store would duplicate those policies.
+
+**Reread the source when saving or retrieving.** Rejected because source mutation could make the artifact disagree with the preview and its captured sequence. Saving the original projection preserves the observation.
+
+**Put omission and retrieval data inside the bounded preview JSON.** Rejected because that spends the conversation budget on metadata and can hide the notice precisely when the budget is smallest. Separate durable model-visible text preserves both obligations.
+
+**Use tool provenance for every spill.** Rejected because a session reference has no model-issued tool call. Invented tool ids would misattribute the artifact rather than describe its producer.
+
+## Consequences
+
+The model can inspect text omitted from a preview without increasing the preview budget. Notices add request tokens outside that budget, and retrieval adds the requested transcript text later. Storage is best-effort: an unavailable notice is honest about loss of retrieval while the bounded preview remains usable. A saved locator can expire even while its notice remains in durable history; this feature does not promise permanent archival or recover content already removed by source compaction.
+
+## Verification
+
+The [unit suite](../../../../packages/context/session-reference/tests/session-reference.spec.ts) pins omission counts, full Unicode and control-character recovery, whole-message drops, three-reference isolation, missing and failed storage, source exclusions and mutation isolation, and cancellation before publication. The [Loader composition test](../../../../packages/context/session-reference/tests/loader-composition.spec.ts) exercises the real local store and paged `read` tool against the middle of a giant single-line message, with target-session storage ownership. The [keyless recorded-session scenario](../../../../snapshots/session/session-reference-spill/snapshot.yml) pins the durable model-visible reference context. Nested Windows-locator regressions cover both serialized extraction and normalization without rewriting unrelated backslashes. Replay [normalizes known quoted spill locators](../../../../packages/test-support/session-snapshot/README.md) while preserving saved byte lengths and omission counts.
+
+## Related decisions
+
+The [tool-output spill decision](../architecture/2026-07-08-tool-output-spill-files.md) remains active: its storage/policy separation, failure degradation, provider caps, and retrieval alternatives still constrain tool consumers. This note extends its producer vocabulary without replacing that rationale. [Separate context injection from turn execution](../architecture/2026-07-24-separate-context-injection-from-turn-execution.md) remains the authority for durable message admission, and [producer-declared context forms](../feature/2026-08-05-context-form-vocabulary.md) remains the authority for recall presentation.

.agents/notes/implemented/bug-fix/2026-09-05-session-reference-spill-reuse.zh.md

@@ -0,0 +1,41 @@
+# Agent Note: 为截断的会话引用复用 spill 存储
+
+Status: implemented
+
+[English](2026-09-05-session-reference-spill-reuse.md) | 中文
+
+## 问题
+
+有界的跨会话预览可能省略整条消息，也可能省略保留消息中的大部分文本。只看到预览的模型需要准确了解省略情况，并能检查已捕获的文本，同时不能把其他会话的指令视为当前授权。源会话继续推进或发生压缩后，再次读取无法恢复同一次观察。
+
+## 决策
+
+[会话引用准备](../../../../packages/context/session-reference/README.zh.md)保留既有预览策略和逐引用 JSON 字节预算。每个被截断的引用通过可选的 `ctx.get("spillStore")` 尝试 `saveText`；未截断的引用不写入产物。完整转录与有界预览来自同一份已捕获的 user／assistant 文本投影，包含压缩检查点，但排除工具、推理与其他注入上下文。不发生第二次源读取。
+
+产物归接收上下文的目标会话所有。其描述性来源是 `{ kind: "session-reference", sessionId, label }`，其中 `sessionId` 标识被引用的会话。[spill 存储](../../../../packages/spill/spill/README.zh.md)在工具来源之外接受这一最小分支；不需要伪造工具名称或调用 id。存储归属不授权取回。
+
+有界预览 JSON 之外的独立省略通知记录精确的 `omittedMessages` 与 `omittedBytes`。通知携带保存后的定位信息和后端 `retrievalHint`，或区分未配置存储与保存失败的不可用结果。该通知是同一条持久引用消息中的模型可见内容，而不是只供 UI 使用的元数据装饰。极小的预览预算无法移除它。保存的转录携带包括 `capturedFormatVersion` 在内的捕获元数据，以及与预览相同的不受信任背景警告。每条消息的 JSON 字符串片段每行至多包含 64 个 Unicode 码点；解码并拼接后可恢复精确文本，包括原始换行。这种固定产物格式让普通分页文件读取可以取回很长的单行文本中部，而不改变预览保留策略。
+
+异步保存后的取消会阻止上下文发布，即使存储已经创建了产物。消费方不增加回滚或删除 API；该产物遵循后端既有过期策略。回放使用已记录的预览与通知，不会重复保存或源读取。
+
+## 考虑过的替代方案
+
+**另写一个会话引用文件存储。** 不予采纳，因为私有命名、会话级归属、定位指引与产物生命周期已经由 spill 存储负责。第二套存储会重复这些策略。
+
+**保存或取回时重新读取源。** 不予采纳，因为源变更可能使产物与预览及其捕获序列不一致。保存原始投影可以保留该次观察。
+
+**把省略与取回数据放入有界预览 JSON。** 不予采纳，因为这会让元数据占用对话预算，并可能在预算最小时恰好隐藏通知。独立的持久模型可见文本同时保留两项保证。
+
+**所有 spill 都使用工具来源。** 不予采纳，因为会话引用没有模型发出的工具调用。虚构工具 id 会错误归属产物，而不是描述其生产者。
+
+## 后果
+
+模型可以检查预览省略的文本，而无需增加预览预算。通知在该预算之外增加请求 token，之后的取回再添加所请求的转录文本。存储采用尽力而为策略：不可用通知如实说明无法取回，而有界预览仍可使用。即使通知仍在持久历史中，保存的定位信息也可能过期；此功能不承诺永久归档，也无法恢复源压缩已经移除的内容。
+
+## 验证
+
+[单元测试](../../../../packages/context/session-reference/tests/session-reference.spec.ts)锁定省略计数、完整 Unicode 与控制字符恢复、整条消息丢弃、三个引用的隔离、无存储与保存失败、来源排除与变更隔离，以及发布前取消。[Loader 组合测试](../../../../packages/context/session-reference/tests/loader-composition.spec.ts)使用真实本地存储和分页 `read` 工具，读取巨型单行消息的中部，并检查存储归目标会话所有。[无密钥录制会话场景](../../../../snapshots/session/session-reference-spill/snapshot.yml)锁定持久的模型可见引用上下文。嵌套 Windows 定位信息回归覆盖序列化提取与规范化，且不改写无关反斜杠。回放会[规范化已知的带引号 spill 定位信息](../../../../packages/test-support/session-snapshot/README.zh.md)，同时保留保存字节数与省略计数。
+
+## 相关决策
+
+[工具输出 spill 决策](../architecture/2026-07-08-tool-output-spill-files.zh.md)保持活跃：其存储／策略分离、失败降级、提供方上限与取回替代方案仍约束工具消费方。本说明扩展其生产者词汇，而不替代这些理由。[分离上下文注入与轮次执行](../architecture/2026-07-24-separate-context-injection-from-turn-execution.zh.md)仍负责持久消息准入，[生产者声明的上下文形式](../feature/2026-08-05-context-form-vocabulary.zh.md)仍负责 recall 展示。

.agents/notes/implemented/bug-fix/2026-09-06-environment-prompt-suffix.i18n.yaml

@@ -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/bug-fix/2026-09-06-environment-prompt-suffix.md
+2026-09-06-environment-prompt-suffix.md: 3438df0fec7b3296084c94b7db68d2800c6d6c98
+2026-09-06-environment-prompt-suffix.zh.md: 102d1196639ac7d7a754a24fc5a93bc93c1d05ba

.agents/notes/implemented/bug-fix/2026-09-06-environment-prompt-suffix.md

@@ -0,0 +1,37 @@
+# Agent Note: Environment facts follow reusable prompt instructions
+
+Status: implemented
+
+English | [中文](2026-09-06-environment-prompt-suffix.zh.md)
+
+## Problem
+
+The local Web URL, Harness checkout path, and session cwd differ across users and machines. Placing those facts before reusable tool instructions makes otherwise identical prompts diverge near their beginning, limiting the prefix available for same-model cache reuse. The model-name introduction identifies the agent and can remain early.
+
+## Decision
+
+The [system-prompt registry](../../../../packages/core/system-prompt/README.md) keeps the fixed Harness identity first and `DEPLOYMENT_PERSONA_PREFIX` at `0`. First-party reusable instructions through `STRUCTURED_OUTPUT` precede the environment suffix: `HARNESS_SOURCE` at `10000`, `WEB_SURFACE` at `10100`, and `DEPLOYMENT_PERSONA_SUFFIX` at `10200`.
+
+Global system-prompt config accepts `personaPrefix` and `personaSuffix`, both defaulting to empty. The [scoped persona row](../../../../packages/preset/persona/README.md) requires `prefix` and accepts `suffix`, defaulting to empty. They register `deployment:persona-prefix` and `deployment:persona-suffix` through the exported `PERSONA_PREFIX_SECTION` and `PERSONA_SUFFIX_SECTION` names. An omitted or empty scoped `suffix` shadows the global suffix away. The shipped Web, headless, SDK, and ACP bundles and standard, PTC, and Cordis presets keep the model introduction in the prefix and place only `Your working directory is {{cwd}}.` in the suffix. These names specify placement, not a classification of the text; no persona parsing or OS field is added.
+
+The [prompt-variables and tool-guidance ownership note](../architecture/2026-07-05-prompt-variables-and-tool-guidance-ownership.md) retains its identity-first persona placement, single-owner rule, strict interpolation, and tool-guidance responsibilities.
+
+## Alternatives considered
+
+**Move the entire persona late.** That moves the model-name introduction away from the beginning without helping same-model reuse. Separating cwd preserves the introduction and reusable instructions together.
+
+**Move only the source path and Web URL.** Leaving cwd inside the early persona still breaks the reusable prefix across workspaces.
+
+**Infer environment fragments from persona text.** Parsing deployment-authored prose makes placement depend on wording. Explicit templates give shipped compositions and custom deployments direct control.
+
+**Move these facts into runtime-context messages.** That changes their message role and persistence placement rather than only separating system sections.
+
+## Consequences
+
+Byte-identical prefixes require the same model introduction, persona prefix, tools, configuration, and preceding section text. Arbitrary extension orders and assembly listeners remain authoritative; this is a first-party placement policy, not a universal stable-prefix guarantee. Provider cache sharing and hit-rate improvements are not measured or promised.
+
+Environment and Web/source guidance follow structured-output instructions. A `complete: true` persona uses only the rendered prefix and ignores the suffix, suppressing every other system section without disabling tool schemas or runtime context. Source and Web facts retain their distinction between the Harness checkout, session workspace, and current working directory.
+
+## Testing
+
+[Registry tests](../../../../packages/core/system-prompt/tests/system-prompt.spec.ts) compare reusable prefixes with the same model and changed checkout paths, URLs, and cwd values; they also cover strict interpolation and complete overrides. [Loop tests](../../../../packages/core/agent-loop/tests/loop.spec.ts) pin early model identity and session-cwd interpolation. [Persona tests](../../../../packages/preset/persona/tests/persona.spec.ts) cover scoped suffix replacement, empty shadowing, and complete personas. [Recorded prompt snapshots](../../../../docs/testing.md) cover emitted prompts in native-tool and generated-SDK compositions; they do not measure provider cache hits.

.agents/notes/implemented/bug-fix/2026-09-06-environment-prompt-suffix.zh.md

@@ -0,0 +1,37 @@
+# Agent Note: 环境事实位于可复用提示词指令之后
+
+Status: implemented
+
+[English](2026-09-06-environment-prompt-suffix.md) | 中文
+
+## 问题
+
+本地 Web URL、Harness checkout 路径和会话 cwd 因用户与机器而异。将这些事实放在可复用工具指令之前，会使其余内容相同的提示词在开头附近就出现差异，限制可供同模型缓存复用的前缀。模型名称介绍标识 agent（智能体），可以保留在靠前的位置。
+
+## 决策
+
+[系统提示词注册表](../../../../packages/core/system-prompt/README.zh.md)将固定 Harness 身份保留在最前，并将 `DEPLOYMENT_PERSONA_PREFIX` 保留在 `0`。截至 `STRUCTURED_OUTPUT` 的第一方可复用指令位于环境后缀之前：`HARNESS_SOURCE` 位于 `10000`，`WEB_SURFACE` 位于 `10100`，`DEPLOYMENT_PERSONA_SUFFIX` 位于 `10200`。
+
+全局 system-prompt 配置接受 `personaPrefix` 与 `personaSuffix`，两者均默认为空。[带作用域的 persona 行](../../../../packages/preset/persona/README.zh.md)要求提供 `prefix`，并接受默认为空的 `suffix`。它们通过导出的 `PERSONA_PREFIX_SECTION` 与 `PERSONA_SUFFIX_SECTION` 名称注册 `deployment:persona-prefix` 与 `deployment:persona-suffix`。省略或为空的作用域 `suffix` 会遮蔽掉全局后缀。交付的 Web、headless、SDK、ACP bundle 以及 standard、PTC、Cordis preset 将模型介绍保留在前缀中，仅将 `Your working directory is {{cwd}}.` 放入后缀。这些名称指定位置，而不对文本分类；不添加 persona 解析或 OS 字段。
+
+[提示词变量与工具指导归属记录](../architecture/2026-07-05-prompt-variables-and-tool-guidance-ownership.zh.md)仍保留 identity-first 的 persona 位置、单一归属规则、严格插值和工具指导职责。
+
+## 曾考虑的替代方案
+
+**将整个 persona 后移。** 这会将模型名称介绍移离开头，却无助于同模型复用。分离 cwd 可以将介绍与可复用指令一起保留。
+
+**仅移动源码路径与 Web URL。** 若 cwd 仍位于靠前的 persona 内，不同工作区之间的可复用前缀仍会被打断。
+
+**从 persona 文本推断环境片段。** 解析部署方撰写的行文会使位置依赖措辞。显式模板让交付组合与自定义部署直接控制位置。
+
+**将这些事实移到 runtime-context 消息。** 这会改变其消息角色和持久化位置，而不只是分离系统段落。
+
+## 后果
+
+字节相同的前缀要求模型介绍、persona 前缀、工具、配置和前置段落文本一致。任意扩展顺序与组装监听器仍决定最终结果；这是一项第一方位置策略，而非通用稳定前缀保证。不测量或承诺提供方共享缓存及命中率提升。
+
+环境与 Web／源码指导位于结构化输出指令之后。`complete: true` persona 仅使用渲染后的前缀并忽略后缀，抑制其他所有系统段落，但不禁用工具 schema 或 runtime context。源码与 Web 事实保留 Harness checkout、会话工作区和当前工作目录之间的区分。
+
+## 测试
+
+[注册表测试](../../../../packages/core/system-prompt/tests/system-prompt.spec.ts)在模型相同、checkout 路径、URL 和 cwd 值变化时比较可复用前缀；同时覆盖严格插值与完整覆盖。[循环测试](../../../../packages/core/agent-loop/tests/loop.spec.ts)固定靠前的模型身份和会话 cwd 插值。[Persona 测试](../../../../packages/preset/persona/tests/persona.spec.ts)覆盖作用域后缀替换、空值遮蔽与完整 persona。[录制的提示词快照](../../../../docs/testing.zh.md)覆盖原生工具与生成 SDK 组合发出的提示词；它们不测量提供方缓存命中。

.agents/notes/implemented/bug-fix/2026-09-06-windows-python-console-spawn-wait.i18n.yaml

@@ -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/bug-fix/2026-09-06-windows-python-console-spawn-wait.md
+2026-09-06-windows-python-console-spawn-wait.md: 92443bcf8a6e4e5609dc469efa4ebd1d82ab127f
+2026-09-06-windows-python-console-spawn-wait.zh.md: dba2f324b29955580fc11e7cea7a0525a8bc8c86

.agents/notes/implemented/bug-fix/2026-09-06-windows-python-console-spawn-wait.md

@@ -0,0 +1,27 @@
+# Agent Note: Wait for the Windows Python console runtime
+
+Status: implemented
+
+English | [中文](2026-09-06-windows-python-console-spawn-wait.zh.md)
+
+## Problem
+
+The installed Python `dsh.exe` console command intermittently exits with Windows access violation `0xc0000005` before initializing a profile. Its smoke assertion omitted the process status and reported only empty streams. A [native faulthandler probe](https://github.com/deepseek-harness/deepseek-harness/actions/runs/34030851888) captures the fault in Python 3.10 `os._execvpe`, called by the runtime console entry, rather than in the bundled Node executable. Direct executable controls pass.
+
+## Decision
+
+The [Python console entry](../../../../python/sdk-runtime/src/deepseek_harness_runtime/__init__.py) uses `subprocess.run` on Windows, inherits standard streams and environment, waits for runtime completion, and exits with the runtime status. POSIX retains `os.execvpe` process replacement. Windows CRT exec is not POSIX process replacement; the explicit spawn-and-wait path avoids the observed native exec operation.
+
+The [installed-wheel smoke](../../../../scripts/smoke-python-runtime.py) reports decimal and unsigned 32-bit hexadecimal status alongside captured streams when profile installation fails. This preserves the distinction between ordinary command failure and native process exceptions.
+
+## Alternatives considered
+
+**Disable Node compile caching.** Not selected: cache environment changes correlated with early probes, but cold-cache controls also passed and Python faulthandler locates the actual fault at the native exec call. Cache configuration remains unchanged.
+
+**Retry or bypass the installed console command.** Rejected because either masks the shipped command failure instead of repairing its process launch. The keyless installed-wheel assertion remains required.
+
+## Consequences
+
+Windows keeps a Python parent until the runtime exits; it no longer depends on CRT overlay behavior. The standard synchronous subprocess implementation owns waiting and interruption cleanup. No custom process-tree manager or global host setting is added.
+
+[Runtime-resolution tests](../../../../python/sdk/tests/test_runtime_resolution.py) retain POSIX forwarding and cover Windows argument/environment forwarding, statuses 0/37/513, real child completion, Unicode streams and arguments with spaces. Native Windows owns the wide exit-status case because POSIX truncates process statuses to eight bits. The [native fixed-count comparison](https://github.com/deepseek-harness/deepseek-harness/actions/runs/34031142773) passes all four patched launches with compile caching enabled; all four unpatched controls also pass in that batch, so it is not a same-batch reproduction. Full installed-wheel CI must validate the final artifact separately from local branch-level tests.

.agents/notes/implemented/bug-fix/2026-09-06-windows-python-console-spawn-wait.zh.md

@@ -0,0 +1,27 @@
+# Agent Note: 等待 Windows Python 控制台运行时
+
+Status: implemented
+
+[English](2026-09-06-windows-python-console-spawn-wait.md) | 中文
+
+## 问题
+
+Python 安装的 `dsh.exe` 控制台命令会在初始化 profile 前间歇性地以 Windows 访问冲突 `0xc0000005` 退出。其冒烟断言遗漏进程状态，只报告空标准流。[原生 faulthandler 探测](https://github.com/deepseek-harness/deepseek-harness/actions/runs/34030851888) 将故障定位在运行时控制台入口调用的 Python 3.10 `os._execvpe`，而非打包的 Node 可执行文件。直接启动可执行文件的对照通过。
+
+## 决策
+
+[Python 控制台入口](../../../../python/sdk-runtime/src/deepseek_harness_runtime/__init__.py) 在 Windows 上使用 `subprocess.run`，继承标准流与环境，等待运行时结束，再以运行时状态退出。POSIX 保留 `os.execvpe` 进程替换。Windows CRT exec 并非 POSIX 进程替换；显式启动并等待的路径避开观测到的原生 exec 操作。
+
+[安装后 wheel 冒烟测试](../../../../scripts/smoke-python-runtime.py) 在 profile 安装失败时，同时报告十进制、无符号 32 位十六进制状态与捕获的标准流。这保留普通命令失败和原生进程异常的区别。
+
+## 已考虑的替代方案
+
+**禁用 Node 编译缓存。** 未采用：早期探测中缓存环境变化与结果相关，但冷缓存对照也能通过，且 Python faulthandler 将实际故障定位在原生 exec 调用。缓存配置保持不变。
+
+**重试或绕过已安装的控制台命令。** 拒绝，因为二者都会掩盖已发布命令的失败，而不是修复进程启动。keyless 安装后 wheel 断言仍为必需检查。
+
+## 后果
+
+Windows 保留 Python 父进程直到运行时退出，不再依赖 CRT overlay 行为。标准同步子进程实现负责等待和中断清理。不添加自定义进程树管理器或全局主机设置。
+
+[运行时解析测试](../../../../python/sdk/tests/test_runtime_resolution.py) 保留 POSIX 转发验证，并覆盖 Windows 参数／环境转发、状态 0/37/513、真实子进程完成、Unicode 标准流和带空格的参数。宽退出状态由原生 Windows 验证，因为 POSIX 会将进程状态截断为八位。[原生固定次数对照](https://github.com/deepseek-harness/deepseek-harness/actions/runs/34031142773) 中，启用编译缓存的四次修复后启动全部通过；该批次四次未修复对照也全部通过，因此它不是同批次复现。完整安装后 wheel CI 必须独立于本地分支级测试，验证最终产物。

.agents/notes/implemented/bug-fix/2026-09-07-pinned-scroll-delivery-before-layout.i18n.yaml

@@ -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/bug-fix/2026-09-07-pinned-scroll-delivery-before-layout.md
+2026-09-07-pinned-scroll-delivery-before-layout.md: f3b428dbe0d2fdc7cbaedf0383d2ab80ea1bfd45
+2026-09-07-pinned-scroll-delivery-before-layout.zh.md: 9b200f45535c67cfbbb75feaf3025c5a3b5c1fa2

.agents/notes/implemented/bug-fix/2026-09-07-pinned-scroll-delivery-before-layout.md

@@ -0,0 +1,23 @@
+# Agent Note: Settle pinned scroll deliveries before layout changes
+
+Status: implemented
+
+English | [中文](2026-09-07-pinned-scroll-delivery-before-layout.zh.md)
+
+## Problem
+
+A delayed scroll sample compares positions from different layouts. While Chat is pinned, a composer or transcript shrink can move the browser floor; subsequent growth can move the browser position again before `scrollend` or the sampling timer. Deferring follow during that interval leaves the observed-top ledger stale and can classify browser layout movement as reader input, disabling follow without a reader gesture.
+
+## Decision
+
+[ChatView](../../../../packages/client/ui-chat/src/client/chat/ChatView.tsx) uses the existing observed-top comparison to sample non-reader pinned scroll deliveries synchronously through the same sample operation that clears pending work. This releases layout follow before further growth. Genuine reader movement remains pending until the existing interval or `scrollend`, even within the follow threshold: growth must not erase small gestures before they accumulate into a scroll-away. Immediate pinned samples use scroll metrics, not semantic-row geometry.
+
+## Alternatives considered
+
+**Defer every delivery.** Coalescing reduces geometry work while reading history, but a pinned browser position and its floor must be attributed in the same layout. A longer timeout or retry cannot recover ownership once the stale comparison disarms it.
+
+**Sample every delivery synchronously.** This restores attribution but also repeats semantic-anchor and reading-line measurements throughout an away-reader scroll burst. Only pinned ownership needs the immediate path.
+
+## Consequences
+
+Pinned deliveries incur immediate scroll-metric reads. History reading retains its bounded sampling cadence, and explicit return-to-bottom deliveries clear any pending away sample. [Focused tests](../../../../packages/client/ui-chat/tests/chat-view.client.spec.tsx) cover shrink/regrowth before scrollend, observer growth without row measurements, repinning with a pending sample, timer and scrollend sampling, and unmount cancellation. The [keyless browser scenario](../../../../apps/web/tests/chat-scroll-contract.e2e.ts) covers pinned Send, real scroll-away input, streaming, and tool disclosure across the long transcript.

.agents/notes/implemented/bug-fix/2026-09-07-pinned-scroll-delivery-before-layout.zh.md

@@ -0,0 +1,23 @@
+# Agent Note: 在布局变化前处理贴底滚动事件
+
+Status: implemented
+
+[English](2026-09-07-pinned-scroll-delivery-before-layout.md) | 中文
+
+## Problem
+
+延迟的滚动采样会比较来自不同布局的位置。Chat 贴底时，输入框或 transcript（文本记录）收缩可能改变浏览器底部位置；随后的增长又可能在 `scrollend` 或采样定时器触发前改变浏览器位置。在此期间推迟跟随会使已观察顶部位置记录过期，把浏览器布局移动误判为读者输入，在没有读者操作时关闭跟随。
+
+## Decision
+
+[ChatView](../../../../packages/client/ui-chat/src/client/chat/ChatView.tsx) 使用现有的已观察顶部位置比较，通过同一个清除待处理工作的采样操作，同步采样非读者引起的贴底滚动事件。这会在后续增长前恢复布局跟随。真实读者移动即使位于跟随阈值内，也保持待处理直到现有周期或 `scrollend`：增长不能在小幅操作累积为离底滚动前将其抵消。立即执行的贴底采样只读取滚动指标，不读取语义行几何。
+
+## Alternatives considered
+
+**延迟所有事件。** 合并采样减少阅读历史时的几何计算，但贴底浏览器位置及其底部必须在同一布局中完成归因。过期比较关闭跟随后，延长超时或重试都无法恢复归属。
+
+**同步采样所有事件。** 这能恢复归因，却也会在离底读者连续滚动时重复测量语义锚点和阅读线。只有贴底归属需要立即处理。
+
+## Consequences
+
+贴底事件会立即读取滚动指标。历史阅读保留有界采样节奏，显式回到底部的滚动事件会清除任何待处理的离底采样。[聚焦测试](../../../../packages/client/ui-chat/tests/chat-view.client.spec.tsx) 覆盖 scrollend 前的收缩与增长、无需行测量的观察器增长、存在待处理采样时重新贴底、定时器与 scrollend 采样，以及卸载取消。[无密钥浏览器场景](../../../../apps/web/tests/chat-scroll-contract.e2e.ts) 覆盖长 transcript 中贴底发送、真实离底输入、流式输出与工具详情展开。

.agents/notes/implemented/feature/2026-07-06-sandbox.i18n.yaml

@@ -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/2026-07-06-sandbox.md
-2026-07-06-sandbox.md: ceffb260163bc21e2ed13275dde38ea0843aac7d
-2026-07-06-sandbox.zh.md: 2b6753585b2778bc0dc898661d9f21779dc37f18
+2026-07-06-sandbox.md: a6e5639e21ca7140cb0314c9918319e99b1495f6
+2026-07-06-sandbox.zh.md: 5144b3fa719465707d7fb2870d45094b8c07661a

.agents/notes/implemented/feature/2026-07-06-sandbox.md

@@ -38,7 +38,7 @@ Four `cordis.yml` entries turn an unconfined coding agent into the sandboxed pro
 
 The swap is invisible to every consumer of `ctx.shell`: the bash tools, hook commands, and background jobs run exactly as before by directly spawning the wrapped argv the provider returns. Deleting the `sandbox` and `permission` entries and replacing `bash` with `@deepseek-ai/dsh-bash-local` is the opt-out — execution is unconfined again and the escalation fields vanish from the tool schema, because they are capability-gated on the mounted executor, not on configuration. Omitting only `approval` keeps confinement but fails every escalation closed with its own error text; `permission` also requires the approval seam and a confining executor, so a partially composed preset layer fails loud at load.
 
-Misconfiguration fails loud: `mode` outside the closed vocabulary is rejected at plugin load, and a host with no usable backend throws the structured `SANDBOX_UNAVAILABLE` at `confine()` rather than degrading to unconfined execution. If the selected runner rejects with attributable `ENOENT` or `EACCES`, the consumer reports the same infrastructure error from the spawn channel before any command starts; other spawn errors retain local command-start semantics while still running nothing. `runnerCommand` on `dsh-sandbox-local` is the operator's explicit assertion of a bwrap-compatible runner (chain and probes skipped); it doubles as the deterministic fake-runner hook for keyless tests.
+Misconfiguration fails loud: `mode` outside the closed vocabulary is rejected at plugin load, and a host with no usable backend throws the structured `SANDBOX_UNAVAILABLE` at `confine()` rather than degrading to unconfined execution. An attributable `ENOENT` or `EACCES` that names the selected runner proves that executable did not start, so the consumer reports the same infrastructure error; other synchronous creation errors propagate unchanged. An asynchronous subprocess-provider rejection exposes no public target stage and retains stage-neutral local semantics. `runnerCommand` on `dsh-sandbox-local` is the operator's explicit assertion of a bwrap-compatible runner (chain and probes skipped); it doubles as the deterministic fake-runner hook for keyless tests.
 
 Denied file effects return a `[sandbox: file access denied under <mode> mode]` marker and instructions not to work around the denial. A confining executor adds paired `sandbox_permissions` and `justification` fields for one approved retry that must be strictly wider than the session's effective mode. A grant widens only that retry; rejection executes nothing, returns `the user rejected escalating this command to "<mode>"`, and permits no re-ask. The owner-derived pending policy context states the current file policy without replacing those enforcement boundaries. When `dsh-permission-presets` is composed with a UI adapter, one preset selects both knob values; unmatched values fold to `custom`. The [ACP application bundle](../../../../packages/bundle/acp-app/README.md) does not mount that UI service and selects its deployment mode explicitly.
 
@@ -70,7 +70,7 @@ Backend profiles share the mode contract but differ in necessary host grants. La
 
 #### The bash consumer
 
-`dsh-bash-sandbox` extends `LocalBashExecutor`, hands `ctx.sandbox` the exact `['bash', '-c', command]` argv, and directly spawns the provider result. This leaves shell semantics and `BASH_ENV` on the inner Bash after the shipped native runner establishes confinement. A provider error propagates unchanged. A pre-process rejection counts as a runner failure only when the caller-owned workdir is independently usable and Node reports `ENOENT` or `EACCES` with either an `error.path` equal to provider argv[0] or, when `error.path` is absent, an exact `syscall: 'spawn <runner>'`; a present path also requires `syscall: 'spawn'` or the exact `spawn <runner>`. Other codes, invalid workdirs, resource failures, unrelated syscalls, and unstructured rejections retain local command-start semantics. Foreground execution converts runner failures to `SANDBOX_UNAVAILABLE` with the original detail; an asynchronous background rejection stamps `runnerFailed: true`, `denied: false`. A `SubprocessRuntime` that synchronously throws the same runner-identifying shape makes background start throw `SANDBOX_UNAVAILABLE`, while other synchronous errors propagate unchanged. After a process starts, foreground and background use one runner-failure classifier that requires the rule's exit-code check and a remaining fatal line after informational exclusions. A match takes priority over denial: foreground execution throws `SANDBOX_UNAVAILABLE` with that fatal line as detail; a settled `ShellProcess` stamps `sandbox.runnerFailed`, and the bash producer renders it through generic `job_output`.
+`dsh-bash-sandbox` extends `LocalBashExecutor`, hands `ctx.sandbox` the exact `['bash', '-c', command]` argv, and directly spawns the provider result. This leaves shell semantics and `BASH_ENV` on the inner Bash after the shipped native runner establishes confinement. A provider error propagates unchanged. An asynchronous rejection counts as a runner failure only when the caller-owned workdir is independently usable and Node reports `ENOENT` or `EACCES` with either an `error.path` equal to provider argv[0] or, when `error.path` is absent, an exact `syscall: 'spawn <runner>'`; a present path also requires `syscall: 'spawn'` or the exact `spawn <runner>`. Other codes, invalid workdirs, resource failures, unrelated syscalls, and unstructured rejections retain stage-neutral provider-failure semantics. Foreground execution converts attributable runner failures to `SANDBOX_UNAVAILABLE` with the original detail; an attributable asynchronous background rejection stamps `runnerFailed: true`, `denied: false`. A `SubprocessRuntime` that synchronously throws the same runner-identifying shape makes background start throw `SANDBOX_UNAVAILABLE`, while other synchronous errors propagate unchanged. After a direct outcome exists, foreground and background use one runner-failure classifier that requires the rule's exit-code check and a remaining fatal line after informational exclusions. A match takes priority over denial: foreground execution throws `SANDBOX_UNAVAILABLE` with that fatal line as detail; a settled `ShellProcess` stamps `sandbox.runnerFailed`, and the bash producer renders it through generic `job_output`.
 
 The model sees the current effective file policy in the owner-derived `sandbox:policy` context, while the static tool description explains the denial marker (`[sandbox: file access denied under <mode> mode]`), encourages attempting commands that may be denied, and forbids retrying around a denial; when the escalation fields are advertised, a denied result additionally carries the escalation hint itself, so the sanctioned same-turn retry is prompted at the decision point rather than depending on the model recalling the description (§ Escalation). [The current-policy decision](../../archived/feature/2026-07-30-current-sandbox-policy-context.md) owns the context's rationale and boundaries.
 
@@ -184,7 +184,7 @@ Costs and accepted limits:
 ## FAQ
 
 - **A command came back with `[sandbox: file access denied under read-only mode]` — did it fail?** It RAN, and the kernel refused a file effect: the denial is a result fact orthogonal to exit code. The teaching forbids retrying around it; the one sanctioned move is the same command retried once with an escalation request.
-- **How is a BROKEN sandbox told apart from a failing command?** Any provider-argv spawn rejection proves the confined launch never started, but it identifies a broken runner only when the caller-owned workdir is usable and Node reports attributable `ENOENT` or `EACCES` for that argv[0]. A bare `syscall: 'spawn'` without an exact error path and all other rejections remain ordinary command-start errors. After a process starts, runner failure outranks denial only when one `runnerFailureRules` entry matches both its optional exit-code gate and a fatal stderr line after exact informational exclusions. Foreground failures throw structured `SANDBOX_UNAVAILABLE` with spawn or matched-line detail; an asynchronously rejected or settled background job stamps `sandbox.runnerFailed` and renders its own marker. A `SubprocessRuntime` that synchronously throws the same `ENOENT`/`EACCES` shape with the runner path makes background start throw the structured error; other synchronous errors propagate unchanged. A Landlock partial-enforcement notice plus an ordinary child failure remains a command result.
+- **How is a BROKEN sandbox told apart from a failing command?** An asynchronous subprocess-provider rejection exposes no public execution stage. It identifies a broken confinement runner only when the caller-owned workdir is usable and Node reports attributable `ENOENT` or `EACCES` for that argv[0]; a bare `syscall: 'spawn'` without an exact error path and all other rejections remain stage-neutral provider failures. After a direct outcome exists, runner failure outranks denial only when one `runnerFailureRules` entry matches both its optional exit-code gate and a fatal stderr line after exact informational exclusions. Foreground runner failures throw structured `SANDBOX_UNAVAILABLE` with executable or matched-line detail; an attributable asynchronous background rejection or a matched settled failure stamps `sandbox.runnerFailed`, while every asynchronous rejection renders the local executor's provider-failure note. A `SubprocessRuntime` that synchronously throws the same `ENOENT`/`EACCES` shape with the runner path makes background start throw the structured error; other synchronous errors propagate unchanged. A Landlock partial-enforcement notice plus an ordinary child failure remains a command result.
 - **What happens on a platform with no backend?** `confine()` throws the fail-closed `SANDBOX_UNAVAILABLE`, and the command never spawns.
 - **`bwrap` is installed on my host but unusable (disabled unprivileged userns, an LSM denying `mount`) — what happens?** The chain probe is functional — it builds and enforces a real profile rather than checking `--version` — so a present-but-unusable `bwrap` fails its probe, selection falls to the packaged Landlock launcher, and the verdict is cached for the provider's lifetime.
 - **Does the sandbox restrict network or process visibility?** `SandboxMode` claims FILE effects only, and no backend claims network. Process visibility is backend-specific: bwrap unshares PID and mounts matching procfs because host `/proc/<pid>` magic links otherwise bypass file confinement, while Landlock and Seatbelt leave process visibility unchanged ([decision](../bug-fix/2026-08-06-bwrap-private-pid-namespace.md)). Whether network restriction becomes its own knob is left open in § The seam.

.agents/notes/implemented/feature/2026-07-06-sandbox.zh.md

@@ -38,7 +38,7 @@ harness 是一个 SDK，因此约束必须是开发者可组合的能力：是
 
 这一替换对 `ctx.shell` 的所有消费方透明：bash 工具、钩子命令和后台任务照常运行，直接使用提供方返回的已包装 argv 启动。删除 `sandbox` 和 `permission` 条目、将 `bash` 替换为 `@deepseek-ai/dsh-bash-local` 即为退出——执行恢复为无约束，升级字段从工具 schema 中消失，因为它们是基于已挂载执行器的能力门控，而非基于配置。仅省略 `approval` 则保留约束但以自身错误文本关闭每次升级；`permission` 还要求 approval seam 和约束执行器同时存在，因此组合不完整的 preset 层会在加载时明确报错。
 
-配置错误会显式导致失败：`mode` 不在封闭词汇中时在插件加载时被拒绝；主机上没有可用后端时在 `confine()` 阶段抛出结构化的 `SANDBOX_UNAVAILABLE`，而非降级为无约束执行。如果所选 runner 以可归因的 `ENOENT` 或 `EACCES` 拒绝，消费方会在任何命令开始前通过 spawn 通道报告同一基础设施错误；其他 spawn 错误仍保留本地命令启动语义，同时也不会运行任何内容。`dsh-sandbox-local` 上的 `runnerCommand` 是运维人员对一个 bwrap 兼容 runner 的显式断言（跳过链和探测）；它同时充当 keyless 测试的确定性 fake-runner 钩子。
+配置错误会显式导致失败：`mode` 不在封闭词汇中时在插件加载时被拒绝；主机上没有可用后端时在 `confine()` 阶段抛出结构化的 `SANDBOX_UNAVAILABLE`，而非降级为无约束执行。若可归因的 `ENOENT` 或 `EACCES` 指明所选 runner，就能证明该 executable 未启动，因此消费方会报告同一基础设施错误；其他同步创建错误原样传播。异步 subprocess-provider rejection 不公开 target 阶段，保留本地不声明阶段的语义。`dsh-sandbox-local` 上的 `runnerCommand` 是运维人员对一个 bwrap 兼容 runner 的显式断言（跳过链和探测）；它同时充当 keyless 测试的确定性 fake-runner 钩子。
 
 被拒绝的文件操作返回 `[sandbox: file access denied under <mode> mode]` 标记，并附带不要绕过拒绝的指令。约束执行器添加配对的 `sandbox_permissions` 和 `justification` 字段，用于一次经批准的重试，该重试必须严格宽于会话的有效模式。授权仅放宽该次重试；拒绝则不执行任何内容，返回 `the user rejected escalating this command to "<mode>"`，且不允许再次请求。由归属方派生的待处理策略上下文会说明当前文件策略，但不会取代这些强制执行边界。当 `dsh-permission-presets` 与某个 UI 适配器一起组合时，一个 preset 同时选定两个旋钮值；不匹配的组合折叠为 `custom`。[ACP（Agent Client Protocol）应用组合包](../../../../packages/bundle/acp-app/README.zh.md)不挂载该 UI 服务，而是显式选定其部署模式。
 
@@ -70,7 +70,7 @@ Landlock launcher 源码和包家族位于 `native/landlock-run`，与 harness
 
 #### bash 消费方
 
-`dsh-bash-sandbox` 扩展 `LocalBashExecutor`，把精确的 `['bash', '-c', command]` argv 交给 `ctx.sandbox`，并直接 spawn 提供方返回的 argv。这样，随附的原生 runner 建立约束后，shell 语义与 `BASH_ENV` 仍由内层 Bash 处理。提供方错误原样传播。进程启动前，只有当调用方拥有的 workdir 经独立验证可用，Node 报告 `ENOENT` 或 `EACCES`，并且错误符合以下一种形态时，才判定为 runner 失败：`error.path` 等于提供方返回的 `argv[0]`，同时 `syscall` 为 `'spawn'` 或精确的 `'spawn <runner>'`；或者 `error.path` 不存在，同时 `syscall` 为精确的 `'spawn <runner>'`。其他错误码、无效 workdir、资源失败、无关 syscall 与无结构拒绝保留本地命令启动语义。前台执行会将 runner 失败转为 `SANDBOX_UNAVAILABLE` 并附上原始详细信息；异步后台拒绝则盖章 `runnerFailed: true`、`denied: false`。如果 `SubprocessRuntime` 同步抛出同样能指明 runner 的形态，后台启动会抛出 `SANDBOX_UNAVAILABLE`；其他同步错误原样传播。进程启动后，前台与后台共用一个 runner 失败分类器：先排除信息性行，再要求规则的退出码检查与余下的一行致命诊断同时匹配。匹配结果优先于拒绝：前台执行抛出 `SANDBOX_UNAVAILABLE`，并以该致命行作为详细信息；结算后的 `ShellProcess` 会盖章 `sandbox.runnerFailed`，bash 生产者再通过通用 `job_output` 渲染它。
+`dsh-bash-sandbox` 扩展 `LocalBashExecutor`，把精确的 `['bash', '-c', command]` argv 交给 `ctx.sandbox`，并直接 spawn 提供方返回的 argv。这样，随附的原生 runner 建立约束后，shell 语义与 `BASH_ENV` 仍由内层 Bash 处理。提供方错误原样传播。异步 rejection 只有在调用方拥有的 workdir 经独立验证可用，Node 报告 `ENOENT` 或 `EACCES`，并且错误符合以下一种形态时，才判定为 runner 失败：`error.path` 等于提供方返回的 `argv[0]`，同时 `syscall` 为 `'spawn'` 或精确的 `'spawn <runner>'`；或者 `error.path` 不存在，同时 `syscall` 为精确的 `'spawn <runner>'`。其他错误码、无效 workdir、资源失败、无关 syscall 与无结构 rejection 保留不声明阶段的 provider-failure 语义。前台执行会将可归因的 runner failure 转为 `SANDBOX_UNAVAILABLE` 并附上原始详情；可归因的异步后台 rejection 则盖章 `runnerFailed: true`、`denied: false`。如果 `SubprocessRuntime` 同步抛出同样能指明 runner 的形态，后台启动会抛出 `SANDBOX_UNAVAILABLE`；其他同步错误原样传播。direct outcome 已存在后，前台与后台共用一个 runner failure 分类器：先排除信息性行，再要求规则的退出码检查与余下的一行致命诊断同时匹配。匹配结果优先于拒绝：前台执行抛出 `SANDBOX_UNAVAILABLE`，并以该致命行作为详情；结算后的 `ShellProcess` 会盖章 `sandbox.runnerFailed`，bash 生产者再通过通用 `job_output` 渲染它。
 
 模型会在归属方派生的 `sandbox:policy` 上下文中看到当前有效的文件策略；静态工具描述则解释拒绝标记（`[sandbox: file access denied under <mode> mode]`），鼓励尝试可能被拒绝的命令，并禁止绕过拒绝重试。当升级字段被公布时，被拒绝的结果还会携带升级提示本身，使被认可的同轮次重试在决策点获得提示，而非依赖模型回忆描述（§ 升级机制）。[当前策略决策](../../archived/feature/2026-07-30-current-sandbox-policy-context.md)负责该上下文的理由与边界。
 
@@ -184,7 +184,7 @@ fs/web/todo 在进程内执行，因此它们的沙箱语义是各自能力边
 ## FAQ
 
 - **一个命令返回了 `[sandbox: file access denied under read-only mode]`——它失败了吗？** 它运行了，内核拒绝了一个文件操作：拒绝是与退出码正交的结果事实。相关指令禁止通过绕过限制来重试；唯一被认可的动作是以升级请求重试同一命令一次。
-- **如何区分损坏的沙箱与失败的命令？** 提供方 argv 的任何 spawn 拒绝都能证明受限启动从未开始，但只有在调用方拥有的 workdir 可用，且 Node 为该 argv[0] 报告可归因的 `ENOENT` 或 `EACCES` 时，才能据此判定 runner 损坏。没有精确错误路径的裸 `syscall: 'spawn'` 和其他所有拒绝仍是普通的命令启动错误。进程启动后，只有当 `runnerFailureRules` 中某一条目同时匹配其可选退出码门控，以及排除整行精确信息性行后的一行致命 stderr 诊断时，runner 失败才会优先于拒绝。前台失败会抛出结构化的 `SANDBOX_UNAVAILABLE`，并附带 spawn 错误或匹配行作为详细信息；遭异步拒绝或已结算的后台任务则盖章 `sandbox.runnerFailed` 并渲染自己的标记。如果 `SubprocessRuntime` 同步抛出同样带有 runner 路径的 `ENOENT`／`EACCES` 形态，后台启动会抛出该结构化错误；其他同步错误原样传播。Landlock 部分强制执行通知加上普通子进程失败时，仍返回命令结果。
+- **如何区分损坏的沙箱与失败的命令？** 异步 subprocess-provider rejection 不公开执行阶段。只有在调用方拥有的 workdir 可用，且 Node 为该 argv[0] 报告可归因的 `ENOENT` 或 `EACCES` 时，才能据此判定 confinement runner 损坏；没有精确错误路径的裸 `syscall: 'spawn'` 和其他所有 rejection 都保持不声明阶段的 provider failure。direct outcome 已存在后，只有当 `runnerFailureRules` 中某一条目同时匹配其可选退出码门控，以及排除整行精确信息性行后的一行致命 stderr 诊断时，runner 失败才会优先于拒绝。前台 runner failure 会抛出结构化的 `SANDBOX_UNAVAILABLE`，并附带 executable 或匹配行详情；可归因的异步后台 rejection 或匹配到的已结算失败会盖章 `sandbox.runnerFailed`，而所有异步 rejection 都会渲染本地执行器的 provider-failure 提示。如果 `SubprocessRuntime` 同步抛出同样带有 runner 路径的 `ENOENT`／`EACCES` 形态，后台启动会抛出该结构化错误；其他同步错误原样传播。Landlock 部分强制执行通知加上普通子进程失败时，仍返回命令结果。
 - **在没有后端的平台上会发生什么？** `confine()` 抛出失败关闭的 `SANDBOX_UNAVAILABLE`，命令永不 spawn。
 - **`bwrap` 已安装在我的主机上但不可用（禁用了非特权 userns、LSM 拒绝 `mount`）——会发生什么？** 链探测是功能性的——它构建并强制一个真实 profile 而非检查 `--version`——因此存在但不可用的 `bwrap` 探测失败，选择落到已打包的 Landlock launcher，结论在提供方生命周期内缓存。
 - **沙箱限制网络或进程可见性吗？** `SandboxMode` 仅声称文件操作，没有后端声称网络。进程可见性取决于后端：bwrap 会 unshare PID 并挂载匹配的 procfs，因为宿主 `/proc/<pid>` 魔法链接会绕过文件约束；Landlock 与 Seatbelt 则保持进程可见性不变（[决策](../bug-fix/2026-08-06-bwrap-private-pid-namespace.zh.md)）。网络限制是否成为自己的旋钮留在 § seam 中开放。

.agents/notes/implemented/feature/2026-07-12-subagent-persona-tool-filter-and-depth.i18n.yaml

@@ -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/2026-07-12-subagent-persona-tool-filter-and-depth.md
-2026-07-12-subagent-persona-tool-filter-and-depth.md: 7ba9768df3679da6b07728cf64237c47d4c73b2f
-2026-07-12-subagent-persona-tool-filter-and-depth.zh.md: d3a8240542d896a27e82b1be1b491c241003e27e
+2026-07-12-subagent-persona-tool-filter-and-depth.md: 511e340c81b811ebcaaea48946c377c99c53da54
+2026-07-12-subagent-persona-tool-filter-and-depth.zh.md: 8a213f33a9b70a9bdec6f23b5bec4db5d94f110f

.agents/notes/implemented/feature/2026-07-12-subagent-persona-tool-filter-and-depth.md

@@ -18,7 +18,7 @@ The controls answer different questions:
 
 | Control | Question | Result |
 |---|---|---|
-| `persona` | What role instructions replace the deployment persona for this child? | A child-local prompt section shadows `deployment:persona` |
+| `persona` | What role instructions replace the deployment persona for this child? | A child-local prompt section shadows `deployment:persona-prefix` |
 | `toolFilter` | Which deployment-global tools enter this child's visible tool view? | A scoped restriction filters globals before child-local tools are added |
 | `maxDepth` | How deep may this delegation tree grow? | A start whose child depth exceeds the absolute cap is rejected |
 
@@ -26,7 +26,7 @@ The controls answer different questions:
 
 ### Persona is a scoped shadow
 
-The persona control changes one child without changing deployment-wide prompt assembly. During unpublished setup, an in-process provider registers a child-scoped section named `deployment:persona`; ordinary most-specific-wins resolution replaces the global section only in that child's assemblies.
+The persona control changes one child without changing deployment-wide prompt assembly. During unpublished setup, an in-process provider registers a child-scoped section named `deployment:persona-prefix`; ordinary most-specific-wins resolution replaces the global section only in that child's assemblies.
 
 The value has the same strict template semantics as the deployment persona. Omitting it inherits the deployment section through the global layer; an explicit empty string shadows the global persona with an empty section. Parent and sibling personas never enter the child's flat scope.
 

.agents/notes/implemented/feature/2026-07-12-subagent-persona-tool-filter-and-depth.zh.md

@@ -18,7 +18,7 @@ subagent 启动有三个独立的组合控制：`persona`、`toolFilter` 和 `ma
 
 | 控制 | 问题 | 结果 |
 |---|---|---|
-| `persona` | 什么角色指令替换该子 agent 的部署人设？ | 一个子 agent 局部的提示词段落遮蔽 `deployment:persona` |
+| `persona` | 什么角色指令替换该子 agent 的部署人设？ | 一个子 agent 局部的提示词段落遮蔽 `deployment:persona-prefix` |
 | `toolFilter` | 部署全局工具中哪些进入该子 agent 的可见工具视图？ | 一个有作用域的限制在添加子 agent 局部工具之前过滤全局工具 |
 | `maxDepth` | 这棵委派树最深可以长到多少层？ | 子 agent 深度超过绝对上限时，启动请求被拒绝 |
 
@@ -26,7 +26,7 @@ subagent 启动有三个独立的组合控制：`persona`、`toolFilter` 和 `ma
 
 ### 人设是有作用域的遮蔽
 
-人设控制改变一个子 agent 的行为，而不改变部署级的提示词组装。在未发布的设置阶段，进程内提供方在子 agent 作用域中注册一个名为 `deployment:persona` 的段落；普通的最具体者优先解析规则仅在该子 agent 的组装中替换全局段落。
+人设控制改变一个子 agent 的行为，而不改变部署级的提示词组装。在未发布的设置阶段，进程内提供方在子 agent 作用域中注册一个名为 `deployment:persona-prefix` 的段落；普通的最具体者优先解析规则仅在该子 agent 的组装中替换全局段落。
 
 其值与部署人设具有相同的严格模板语义。省略时通过全局层继承部署段落；显式空字符串则以空段落遮蔽全局人设。父级和兄弟级的人设永远不会进入子 agent 的扁平作用域。
 

.agents/notes/implemented/feature/2026-07-16-persistent-pty-sessions.i18n.yaml

@@ -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: d0bbb230bdb1cf7ab723a3451f0524cc86996209
-2026-07-16-persistent-pty-sessions.zh.md: 359d8e2b6c7f0ea857d0e73349421fc673d5cc3b
+2026-07-16-persistent-pty-sessions.md: d62587ced78be75de121fe907ca258d07a2830a1
+2026-07-16-persistent-pty-sessions.zh.md: a26c3ed852b3a18842585e19ecb2a6994f8d9117

.agents/notes/implemented/feature/2026-07-16-persistent-pty-sessions.md

@@ -45,7 +45,7 @@ A registered `shell` backend constrains how a terminal starts; it does not const
 
 Sandboxing confines local process effects but does not make arbitrary shell input safe: network calls and other external side effects remain governed by deployment policy. Tool descriptions state that PTY sessions are less auditable than one-shot tools and should be used only when persistence or interactive stdin is necessary.
 
-The local subprocess terminal primitive uses only public `node-pty` capabilities: child PID, `data` and `exit` notifications, `write`, and `kill`. It does not assume access to the native master fd or call `waitpid` from TypeScript. Platform process inspectors below that primitive derive foreground process groups and parent/child identity from `/proc` on Linux and `ps` on macOS. The [portable execution-world decision](../architecture/2026-07-28-portable-execution-world-consumers.md) owns this process/consumer split.
+The local subprocess terminal primitive uses only public `node-pty` capabilities: child PID, `data` and `exit` notifications, `write`, and `kill`. It does not assume access to the native master fd or call `waitpid` from TypeScript. On supported Linux hosts the [native-containment owner](../architecture/2026-08-28-subprocess-native-containment.md) starts that same PTY command inside a user-systemd scope without changing its PID, session, controlling terminal, foreground-group, or readiness semantics. Platform process inspectors below the primitive still derive foreground process groups and fallback parent/child identity from `/proc` on Linux and `ps` on macOS. The [portable execution-world decision](../architecture/2026-07-28-portable-execution-world-consumers.md) owns this process/consumer split.
 
 ### Six model-facing tools
 
@@ -92,7 +92,9 @@ Background sends use the existing task completion notice and `job_output` result
 
 ### Process-tree teardown
 
-The subprocess terminal handle owns the top-level terminal process and its session. On close it snapshots transitive descendants by parent PID in children-first order, sends `SIGTERM`, waits, rescans for children forked during shutdown, sends `SIGKILL` to the union, and verifies every non-zombie descendant left the process table before stopping the top-level process. A matching Linux zombie has no executable work and therefore counts as quiescent. Every captured PID includes process-start identity so reuse cannot redirect escalation.
+On supported Linux hosts, the subprocess terminal handle binds the top-level PTY process to its transient user-systemd scope. Before establishment, close sends `SIGTERM` through the direct PTY fallback so the bootstrap cannot continue; after establishment, it signals the scope alone and uses the direct fallback only if scope signalling fails. It waits for the manager to prove that range empty and escalates to `SIGKILL` after the configured grace. Scope membership continues to include descendants that call `setsid` or reparent, while the PTY's direct exit notification remains the terminal outcome.
+
+Fallback hosts retain observational process-session cleanup. The handle snapshots transitive descendants by parent PID in children-first order, sends `SIGTERM`, waits, rescans for children forked during shutdown, sends `SIGKILL` to the union, and verifies every non-zombie descendant left the process table before stopping the top-level process. A matching Linux zombie has no executable work and therefore counts as quiescent. Every captured PID includes process-start identity so reuse cannot redirect escalation.
 
 Teardown reports top-level exit and survivor cleanup independently. The PTY session does not claim success merely because the shell exited: it calls `SubprocessTerminalHandle.terminate()` and awaits whole-session quiescence, propagating a cleanup failure that names survivors. A failed close is not cached forever: the registry and local session clear the fence only when it still names that failed attempt, so a later explicit or lifecycle close retries without disturbing a newer concurrent attempt. Service disposal still clears its backend, reservation, and owner-detacher registries when a close fails.
 
@@ -158,7 +160,7 @@ The package ships concise tool guidance explaining persistent state, owner isola
 
 - Per-file coverage pins owner fencing, concurrent reservations, cancellation during pre-write inspection, unpublished-spawn cancellation and awaited teardown, sandbox-mode change rejection, retriable lifecycle cleanup, readiness tiers, rejection of pre-write stdin waits and delayed earlier prompts, 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.
 - Subprocess process fixtures cover non-leader and non-main-thread stdin waits, thread-local fd tables, the `/dev/tty` alias, supported kernel ABIs under user-mode emulation, rejection of fd 0 backed by a pipe, zombie quiescence, unreadable process state, unsupported architectures, and other false-positive rejection; macOS inspector logic is injected into the same unit suite.
-- Real `node-pty` and PTY-consumer tests jointly exercise shell state, controlling-terminal input through `/dev/tty`, the exact attribution when process syscalls are readable, its bounded idle fallback when host policy denies them, shared sandbox policy, environment scrubbing, raw-mode foreground `SIGINT`, a TERM-ignoring descendant, and immediate post-disposal quiescence on supported hosts.
+- Real `node-pty` and PTY-consumer tests jointly exercise shell state, controlling-terminal input through `/dev/tty`, the exact attribution when process syscalls are readable, its bounded idle fallback when host policy denies them, shared sandbox policy, environment scrubbing, raw-mode foreground `SIGINT`, a TERM-ignoring descendant, and immediate post-disposal quiescence. The Linux native smoke keeps the PTY PID, session leader, controlling terminal, foreground `inputWaiting`, and readiness while a reparented `setsid` descendant remains owned by the scope; fallback suites retain identity-fenced observational cleanup coverage.
 - A Loader-driven `cordis.yml` test mounts the real three-package composition and verifies that delayed pipeline output returns with the completed command instead of being classified as terminal-input readiness. The SDK minimal snapshot pins that output through the persistent Bash tool; ACP and headless snapshots pin the six terminal schemas, bounded results, and errors through opt-in overlays; TUI snapshots pin terminal and generic card presentation.
 - Package contracts, the architecture map, subsystem pages, generated catalogs, and the website API describe the same shipped surface.
 
@@ -172,7 +174,7 @@ The package ships concise tool guidance explaining persistent state, owner isola
 
 **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 local provider's captured tree.** A process that reparents before teardown is no longer discoverable from the `node-pty` root. The local terminal primitive accepts that cleanup gap instead of risking SID-wide signals to unrelated processes.
+**Native Linux ownership closes the process-tree observation gap; fallback ownership does not.** A supported user-systemd scope retains a daemonized or reparented descendant as a member until the scope becomes empty. On macOS, Windows ConPTY, and Linux hosts that cannot establish the scope, a process that escapes before observational teardown can still evade the captured tree; the fallback accepts that gap instead of risking SID-wide signals to unrelated processes.
 
 **A shell can cause external side effects.** Session sandboxing and environment scrubbing reduce local exposure but do not undo pushes, API calls, or messages. Deployments that cannot tolerate those effects must omit PTY or add network policy.
 

.agents/notes/implemented/feature/2026-07-16-persistent-pty-sessions.zh.md

@@ -45,7 +45,7 @@ agent scope dispose（资源释放）时先撤销注册，再等待全部所属
 
 沙箱限制本地进程副作用，但不会让任意 shell 输入自动安全：网络调用和其他外部副作用仍由部署策略治理。工具描述会说明 PTY 会话比一次性工具更难审计，只应在确实需要持久状态或交互式 stdin 时使用。
 
-本地子进程终端原语只使用 `node-pty` 的公开能力：子进程 PID、`data` 与 `exit` 通知、`write` 和 `kill`。它不假设能访问原生 master fd，也不从 TypeScript 调用 `waitpid`。该原语下的平台进程检查器在 Linux 上通过 `/proc`、在 macOS 上通过 `ps` 推导前台进程组和父子进程身份。[可移植执行环境决策](../architecture/2026-07-28-portable-execution-world-consumers.zh.md)负责定义这种进程／消费方拆分。
+本地子进程终端原语只使用 `node-pty` 的公开能力：子进程 PID、`data` 与 `exit` 通知、`write` 和 `kill`。它不假设能访问原生 master fd，也不从 TypeScript 调用 `waitpid`。在受支持的 Linux 宿主上，[原生收容 owner](../architecture/2026-08-28-subprocess-native-containment.zh.md)会在 user-systemd scope 内启动同一条 PTY 命令，同时保持 PID、session、控制终端、前台进程组与就绪语义。该原语下的平台进程检查器仍在 Linux 上通过 `/proc`、在 macOS 上通过 `ps` 推导前台进程组和 fallback 父子进程身份。[可移植执行环境决策](../architecture/2026-07-28-portable-execution-world-consumers.zh.md)负责定义这种进程／消费方拆分。
 
 ### 6 个面向模型的工具
 
@@ -92,7 +92,9 @@ Tier 2 在持续 `idleSilenceMs` 没有输出后返回 `inferred_idle`，因此
 
 ### 进程树 teardown
 
-子进程终端句柄拥有顶层终端进程及其会话。关闭时，它按父 PID 以子进程优先顺序捕获传递后代、发送 `SIGTERM` 并等待，然后重新扫描关停期间 fork 出的子进程，向二者并集发送 `SIGKILL`，并在停止顶层进程前验证每个非僵尸后代都已离开进程表。身份匹配的 Linux 僵尸进程已无可执行工作，因此视为完全停稳。每个捕获的 PID 都包含进程启动身份，避免 PID 复用把升级信号发给无关进程。
+在受支持的 Linux 宿主上，subprocess 终端句柄会把顶层 PTY 进程绑定到临时 user-systemd scope。建立前，close 通过 direct PTY fallback 发送 `SIGTERM`，阻止 bootstrap 继续；建立后只向 scope 发送信号，并且仅在 scope signalling 失败时使用 direct fallback。随后它等待 manager 证明该 range 为空，并在配置的宽限期后升级到 `SIGKILL`。调用 `setsid` 或发生 reparent 的后代仍属于 scope，而 PTY 的 direct exit 通知继续作为终端结果。
+
+fallback 宿主保留观察式进程 session 清理。句柄会按父 PID 以子进程优先顺序捕获传递后代、发送 `SIGTERM` 并等待，然后重新扫描关停期间 fork 出的子进程，向二者并集发送 `SIGKILL`，并在停止顶层进程前验证每个非僵尸后代都已离开进程表。身份匹配的 Linux 僵尸进程已无可执行工作，因此视为完全停稳。每个捕获的 PID 都包含进程启动身份，避免 PID 复用把升级信号发给无关进程。
 
 teardown 独立报告顶层进程退出与存活进程清理。PTY 会话不会只因 shell 退出就声称成功：它会调用 `SubprocessTerminalHandle.terminate()` 并等待整个会话完全停稳，若清理失败则向外传播并列出存活者。失败的 close 不会永久缓存：注册表与本地会话各自仅在关闭围栏仍指向该次失败尝试时才将其清除，因此后续的显式 close 或生命周期 close 会重试，且不会干扰较新的并发尝试。即使某个 close 失败，服务 dispose 仍会清空其后端、预留与 owner detacher 注册表。
 
@@ -158,7 +160,7 @@ plugins:
 
 - 逐文件覆盖测试锁定了 owner 隔离、并发预留、写入前检查期间的取消、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、对写入前 stdin 等待与延迟到达的先前 prompt 的拒绝、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。
 - 子进程 fixture（测试前置数据）覆盖非 leader 与非主线程的 stdin 等待、线程本地 fd 表、`/dev/tty` 别名、用户态模拟下受支持的内核 ABI、拒绝把指向管道的 fd 0 当作终端输入、僵尸进程完全停稳、不可读进程状态、不支持的架构和其他误报拒绝；同一单元测试套件通过注入覆盖 macOS 检查器逻辑。
-- 真实 `node-pty` 与 PTY 消费方测试共同在受支持宿主上覆盖 shell 状态、通过 `/dev/tty` 读取控制终端输入、进程 syscall 可读时的精确归因、宿主策略拒绝读取时的有界 idle fallback、共享沙箱策略、环境清洗、raw mode 前台 `SIGINT`、忽略 `SIGTERM` 的后代进程，以及 dispose 返回后立即完全停稳。
+- 真实 `node-pty` 与 PTY 消费方测试共同覆盖 shell 状态、通过 `/dev/tty` 读取控制终端输入、进程 syscall 可读时的精确归因、宿主策略拒绝读取时的有界 idle fallback、共享沙箱策略、环境清洗、raw mode 前台 `SIGINT`、忽略 `SIGTERM` 的后代进程，以及 dispose 返回后立即完全停稳。Linux native 冒烟测试会在一个 reparent 的 `setsid` 后代仍由 scope 拥有时，保持 PTY PID、session leader、控制终端、前台 `inputWaiting` 与 readiness；fallback 测试套件继续覆盖带身份围栏的观察式清理。
 - Loader 驱动的 `cordis.yml` 测试挂载真实三包组合，并验证延迟到达的流水线输出随已完成命令返回，而不会被归类为终端输入就绪。SDK minimal 快照通过持久 Bash 工具固定该输出；ACP 与 headless 快照通过 opt-in overlay 固定 6 个终端 schema、有界结果和错误；TUI 快照固定 terminal 与 generic 卡片展示。
 - 包约定、架构图、子系统页面、生成目录和 website API 描述同一个已发布接口。
 
@@ -172,7 +174,7 @@ plugins:
 
 **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，但任何提示词都无法让状态持久化变成确定行为。
 
-**daemonized 后代进程可能离开本地提供方捕获的进程树。**在 teardown 前 reparent 的进程无法再从 `node-pty` 根进程发现。本地终端原语接受这个清理缺口，不冒险按 SID 向无关进程发送信号。
+**Linux native ownership 消除了进程树观察缺口，fallback ownership 没有。**受支持的 user-systemd scope 会持续保有 daemonized 或 reparent 后代，直到 scope 为空。在 macOS、Windows ConPTY，以及无法建立 scope 的 Linux 宿主上，观察式 teardown 开始前已经逃逸的进程仍可能避开捕获树；fallback 接受这个缺口，不冒险按 SID 向无关进程发送信号。
 
 **Shell 可以造成外部副作用。**会话沙箱和环境清洗降低本地暴露，但无法撤销 push、API 调用或消息发送。无法容忍这些副作用的部署必须省略 PTY 或增加网络策略。
 

.agents/notes/implemented/feature/2026-07-27-web-subagent-conversations.i18n.yaml

@@ -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/2026-07-27-web-subagent-conversations.md
-2026-07-27-web-subagent-conversations.md: 28e06d1db9103de53e6e2fb266a03e168428d0d2
-2026-07-27-web-subagent-conversations.zh.md: e93abe89e855d688d822bfd20fd1644f24bdf3e9
+2026-07-27-web-subagent-conversations.md: d0713731de86df639a426ee56c39857c50af6bc7
+2026-07-27-web-subagent-conversations.zh.md: 0f4f6467484ee1caa3cdf28f44e016ff674ac1dc

.agents/notes/implemented/feature/2026-07-27-web-subagent-conversations.md

@@ -16,11 +16,11 @@ The UI must also preserve the membership, modes, and diagnostics of the [durable
 
 The Web product exposes the selected session's direct session-backed subagents from the current-title lineage region in the header. Users can lazily expand descendant catalogs and open either mode in the existing conversation region. A one-shot child is permanently read-only. A continuable child accepts human follow-ups only while its exact direct-parent Agent is live; otherwise its persisted transcript remains readable with a recovery explanation.
 
-Every opened child carries a catalog-derived address `{ parentSessionId, childSessionId, mode }`. The mode-bearing address, not lineage or the coarse origin marker, selects dedicated history and prompt transports. History reads the persisted session without activation. A continuable prompt calls `ctx.subagents.followup()` and succeeds at inbox acceptance with `{ messageId }`; it does not steer an open turn, expose an Activation, wait for completion, or return an outcome.
+Every opened child carries a catalog-derived address `{ parentSessionId, childSessionId, mode }`. The mode-bearing address, not lineage or the coarse origin marker, selects dedicated history and prompt transports. History reads the persisted session without activation. A continuable prompt carries Queue or Steer delivery through `subagent.prompt` and succeeds at inbox acceptance with `{ messageId }`; it does not expose an Activation, wait for completion, or return an outcome. Adjacent-Agent model messages use the separately owned fixed-Steer operation.
 
 The generic Host domain preserves the same ownership boundary. `session.history` and the source side of `session.fork` read an attached Session or inspect persistence without acquiring an Agent; history folds cold projection values from that exact inspected prefix, while a fork publishes an ordinary independent session. Generic Agent-bound session, command, and goal routes return `agent-busy` for session-backed subagents, as do explicit-id `session.create` adoption and attached-only queue controls. The denial classifier accepts the coarse `origin` marker, a `subagent/descriptor` in the session's own suffix, or exact live runtime ownership by the parent; these signals only prevent generic ownership and never replace catalog mode or direct-parent authorization.
 
-Stopping an addressed child never falls through to `session.cancel`. `SubagentRuntime.followup()` owns admission only until inbox acceptance and grants no cancellation handle; a running continuable child is stopped through the dedicated `subagent.interrupt` route under the [current-turn interrupt contract](2026-08-06-continuable-subagent-interrupt.md), which parks pending work instead of discarding it. One-shot children remain uncancellable from the Web.
+Stopping an addressed child never falls through to `session.cancel`. Browser prompt delivery owns admission only until inbox acceptance and grants no cancellation handle; a running continuable child is stopped through the dedicated `subagent.interrupt` route under the [current-turn interrupt contract](2026-08-06-continuable-subagent-interrupt.md), which parks pending work instead of discarding it. One-shot children remain uncancellable from the Web.
 
 This decision covers Web discovery, transcript viewing, and parent-authorized human continuation. It does not make a subagent independently user-owned; that product remains [interactive side sessions](../../proposed/feature/2026-07-08-interactive-side-sessions.md).
 
@@ -45,7 +45,7 @@ Healthy rows reuse the standard session projections retained in the list mirror.
 
 Selecting a row records its exact address before opening the resident client `Session`. History pagination, event folding, tool render intents, titles, and live mux reconciliation reuse the ordinary conversation machinery. Breadcrumbs follow parent links only through `origin: 'subagent'` rows, include the first ordinary owner, and keep ordinary forks single-level. Each subagent breadcrumb gets its direct-parent sibling catalog and uses that catalog's label when available. Forking an addressed subagent creates an ordinary fork with direct source lineage and attaches it to the nearest workspace-owning ancestor. The catalog is an ARIA tree with lazy ArrowRight/ArrowLeft disclosure, linear ArrowUp/ArrowDown navigation, Home/End, Escape, and focus restoration.
 
-A one-shot row always replaces the composer with copy explaining that the execution record is read-only. A continuable row does so only while `parentAvailable` is false and the child is not running; a running parent-offline child keeps the ordinary composer with its input and Send action disabled so independent Stop stays reachable, and the read-only takeover returns once it stops. With a live parent, Enter and Send admit another FIFO turn even while the child runs, while independent Stop routes through `subagent.interrupt` ([interrupt contract](2026-08-06-continuable-subagent-interrupt.md)). Prompt failures retain the draft through the ordinary error behavior.
+A one-shot row always replaces the composer with copy explaining that the execution record is read-only. A continuable row does so only while `parentAvailable` is false and the child is not running; a running parent-offline child keeps the ordinary composer with its input and Send action disabled so independent Stop and live QueueDock controls stay reachable, and the read-only takeover returns once it stops. With a live parent, the ordinary Enter/Cmd+Enter preference selects Queue or best-effort Steer even while the child runs. QueueDock Edit, Remove, and Steer remain available for a live continuable child even when its parent is offline, while independent Stop routes through `subagent.interrupt` ([interrupt contract](2026-08-06-continuable-subagent-interrupt.md)). Prompt failures retain the draft through the ordinary error behavior.
 
 Agent-bound auxiliary controls are unavailable in addressed child views. In particular, the model selector and `/model` contribution do not call ordinary `session.models` or `session.selectModel`; the Host also rejects any accidental call instead of activating persisted child history outside the direct-parent continuation path.
 
@@ -55,13 +55,13 @@ Agent-bound auxiliary controls are unavailable in addressed child views. In part
 
 - `subagent.list` takes `parentSessionId`, calls `ctx.subagents.listChildren(parentSessionId, signal)`, returns the complete ordered entries with each healthy row's boolean `hasChildren` snapshot, replaces each healthy row's corpus activity with whether its exact Agent driver is running, and includes whether the exact parent currently resolves from `ctx.agents`.
 - `subagent.history` takes the full mode-bearing address plus ordinary page arguments. It verifies the child and mode against the direct catalog, reads through `ctx.sessionQuery.readSession()`, rechecks direct lineage, and returns the ordinary raw-event, render-intent, pagination, and host-computed session-projection baseline without publishing an Agent.
-- `subagent.prompt` accepts only a `mode: 'continuable'` address and upload-shaped `PromptContentPart[]`; the Host admits and persists image parts into durable references before delivery ([image delivery](../../archived/bug-fix/2026-08-27-steer-followup-image-delivery.md)). It requires the exact live parent, revalidates the catalog address, calls `ctx.subagents.followup(parent, childId, content, { source, signal })`, and returns the accepted `MessageId`.
+- `subagent.prompt` accepts only a `mode: 'continuable'` address, `delivery: 'queue' | 'steer'`, and upload-shaped `PromptContentPart[]`; the Host admits and persists image parts into durable references before delivery ([image delivery](../../archived/bug-fix/2026-08-27-steer-followup-image-delivery.md)). It requires the exact live parent, revalidates the catalog address, uses the continuation manager's shared human-delivery admission, and returns the accepted `MessageId`.
 
 The gateway maps missing parent, missing or diagnostic catalog entries, not-resumable and unauthorized children, request cancellation, image admission and image-capability refusals (`subagent/attachment-invalid`), and temporarily unavailable continuation admission to typed RPC errors. It does not expose descriptor or provider details. A list/prompt race is normal: the prompt result, not the earlier availability or activity snapshot, is authoritative.
 
 Viewing persisted history creates no mux subscription by itself. When a follow-up materializes a cold child Activation, the existing Host and mux streams publish its lifecycle and events. Reconnect rebuilds the addressed window through `subagent.history`.
 
-The ordinary `session.history` route is likewise observation-only for both ordinary and subagent sessions, but it does not carry the catalog address or grant continuation authority. Every ordinary route that needs an Agent resolves through the shared ownership fence before cold resume; `session.cancel` and `session.updateQueue` apply the same check directly because they intentionally query only attached Agents.
+The ordinary `session.history` route is likewise observation-only for both ordinary and subagent sessions, but it does not carry the catalog address or grant continuation authority. Every ordinary route that needs an Agent resolves through the shared ownership fence before cold resume; `session.cancel` retains that fence. `session.updateQueue` has one target-local exception for a live child whose current projected identity is continuable and comes from its own non-seed suffix; one-shot, missing, unknown, corrupt, seed-only, or cold children remain fenced.
 
 The adapter stays behind the generated Remote namespace; `dsh-host-webserver` remains a carrier. Browser code imports the contract through the existing connection package and never reaches host `ctx`, preserving the [archived GUI RPC layering decision](../../archived/architecture/2026-07-19-gui-layering-and-rpc-protocol.md).
 

.agents/notes/implemented/feature/2026-07-27-web-subagent-conversations.zh.md

@@ -16,11 +16,11 @@ UI 还必须保留[持久化目录](../../archived/feature/2026-07-22-durable-su
 
 Web 产品通过页头的当前 title 谱系区域公开选中会话中由会话支撑的直接 subagent。用户可以懒加载展开后代目录，并在现有对话区域中打开任一 mode。one-shot child 永久只读。可继续 child 只有在其确切直接 parent agent 存活时才接受用户后续消息；否则，其持久化 transcript 仍然可读，并附带恢复说明。
 
-每个打开的 child 都携带目录派生地址 `{ parentSessionId, childSessionId, mode }`。选择专用历史与提示词传输的是包含 mode 的地址，而不是谱系或粗粒度 origin 标记。历史操作会从持久化存储读取会话，而不触发激活。可继续提示词操作会调用 `ctx.subagents.followup()`，并在 inbox 接受消息时以 `{ messageId }` 成功返回；它不会对进行中的轮次执行 steering（中途引导）、公开 Activation、等待完成或返回结果。
+每个打开的 child 都携带目录派生地址 `{ parentSessionId, childSessionId, mode }`。选择专用历史与提示词传输的是包含 mode 的地址，而不是谱系或粗粒度 origin 标记。历史操作会从持久化存储读取会话，而不触发激活。可继续提示词通过 `subagent.prompt` 携带 Queue 或 Steer 投递，并在 inbox 接受消息时以 `{ messageId }` 成功返回；它不会公开 Activation、等待完成或返回结果。相邻 Agent 的模型消息使用单独拥有的固定 Steer 操作。
 
 通用 Host 领域遵守同一所有权边界。`session.history` 与 `session.fork` 的源端会读取已附加 Session 或检查持久化存储，而不获取 Agent；history 从所检查的确切前缀归并冷态投影值，fork 则发布一个普通的独立会话。绑定到 Agent 的通用会话、命令与目标路由会对由会话支撑的 subagent 返回 `agent-busy`；显式 id 的 `session.create` 接纳与仅针对已附加会话的队列控件亦然。拒绝分类器接受粗粒度 `origin` 标记、会话自身后缀中的 `subagent/descriptor`，或 parent 对其确切的存活运行时所有权；这些信号只会阻止通用路径取得所有权，绝不取代目录 mode 或直接 parent 授权。
 
-停止一个已寻址 child 绝不回退到 `session.cancel`。`SubagentRuntime.followup()` 只负责消息被 inbox 接受前的准入，不授予取消句柄；正在运行的可继续 child 通过专用的 `subagent.interrupt` 路由停止，遵循[当前轮次中断约定](2026-08-06-continuable-subagent-interrupt.zh.md)，该约定会停放并保留待处理工作，而不是将其丢弃。one-shot child 在 Web 端仍不可取消。
+停止一个已寻址 child 绝不回退到 `session.cancel`。浏览器 prompt 投递只负责消息被 inbox 接受前的准入，不授予取消句柄；正在运行的可继续 child 通过专用的 `subagent.interrupt` 路由停止，遵循[当前轮次中断约定](2026-08-06-continuable-subagent-interrupt.zh.md)，该约定会停放并保留待处理工作，而不是将其丢弃。one-shot child 在 Web 端仍不可取消。
 
 本决策涵盖 Web 端发现、transcript 查看与经 parent 授权的用户继续交互。它不会让 subagent 成为用户独立所有的对象；这类产品仍然属于[交互式 side session](../../proposed/feature/2026-07-08-interactive-side-sessions.zh.md)。
 
@@ -45,7 +45,7 @@ Figma 中的 [subagent 列表](https://www.figma.com/design/jRBBK7zBgcszdVWQ0Fh5
 
 选择一行后，系统会先记录其确切地址，再打开常驻客户端 `Session`。历史分页、事件 fold、工具渲染意图、title 与实时 mux 归并都会复用普通对话机制。面包屑导航只会沿 `origin: 'subagent'` 行的父链接逐级回溯，包含第一个普通 owner，并让普通 fork 保持单层。每一级 subagent 面包屑都会获得其直接 parent 的 sibling 目录，并在目录可用时采用其中的 label。从已寻址 subagent 创建 fork 时，会生成具有直接源谱系的普通 fork，并将其附加到最近拥有 Workspace 的祖先。目录是一棵 ARIA 树，支持懒加载式 ArrowRight／ArrowLeft 展开与折叠、线性 ArrowUp／ArrowDown 导航、Home／End、Escape 以及焦点恢复。
 
-one-shot 行始终会用文案替代输入框，说明执行记录为只读。可继续行仅在 `parentAvailable` 为 false 且 child 未在运行时如此；parent 离线但仍在运行的 child 保留普通输入框，并禁用其输入区和 Send 操作，让独立的 Stop 保持可达，停止后只读替代恢复。parent 在线时，即使 child 正在运行，Enter 和 Send 也会准入另一个 FIFO 轮次，而独立的 Stop 经由 `subagent.interrupt` 路由（[中断约定](2026-08-06-continuable-subagent-interrupt.zh.md)）。提示词失败会通过普通错误行为保留草稿。
+one-shot 行始终会用文案替代输入框，说明执行记录为只读。可继续行仅在 `parentAvailable` 为 false 且 child 未在运行时如此；parent 离线但仍在运行的 child 保留普通输入框，并禁用其输入区和 Send 操作，让独立的 Stop 与在线 QueueDock 控制保持可达，停止后只读替代恢复。parent 在线时，即使 child 正在运行，普通 Enter／Cmd+Enter 偏好也会选择 Queue 或 best-effort Steer。对在线可继续 child，QueueDock Edit、Remove 与 Steer 在 parent 离线时仍可用；独立 Stop 经由 `subagent.interrupt` 路由（[中断约定](2026-08-06-continuable-subagent-interrupt.zh.md)）。提示词失败会通过普通错误行为保留草稿。
 
 已寻址 child 视图不提供绑定到 agent 的辅助控件。具体而言，模型选择器与 `/model` contribution 不会调用普通 `session.models` 或 `session.selectModel`；Host 也会拒绝任何意外调用，而不是在直接 parent 继续执行路径之外激活持久化 child 历史。
 
@@ -55,13 +55,13 @@ one-shot 行始终会用文案替代输入框，说明执行记录为只读。
 
 - `subagent.list` 接受 `parentSessionId`，调用 `ctx.subagents.listChildren(parentSessionId, signal)`，返回完整有序的条目以及每个健康行的布尔 `hasChildren` 快照，把每个健康行的语料活动状态替换为其确切 Agent driver 是否正在运行，并说明当前能否从 `ctx.agents` 解析出确切 parent。
 - `subagent.history` 接受包含 mode 的完整地址与普通页参数。它对照直接目录校验 child 与 mode，通过 `ctx.sessionQuery.readSession()` 读取，再次检查直接谱系，并在不发布 agent 的情况下返回普通原始事件、渲染意图、分页与由 Host 计算的会话投影基线。
-- `subagent.prompt` 只接受 `mode: 'continuable'` 地址与上传形态的 `PromptContentPart[]`；Host 在投递前把图片部分准入并持久化为持久引用（[图片投递](../../archived/bug-fix/2026-08-27-steer-followup-image-delivery.md)）。它要求确切的存活 parent，重新校验目录地址，调用 `ctx.subagents.followup(parent, childId, content, { source, signal })`，并返回已接受的 `MessageId`。
+- `subagent.prompt` 只接受 `mode: 'continuable'` 地址、`delivery: 'queue' | 'steer'` 与上传形态的 `PromptContentPart[]`；Host 在投递前把图片部分准入并持久化为持久引用（[图片投递](../../archived/bug-fix/2026-08-27-steer-followup-image-delivery.md)）。它要求确切的存活 parent，重新校验目录地址，使用 continuation manager 共享的人类投递准入，并返回已接受的 `MessageId`。
 
 网关会将 parent 缺失、目录条目缺失或为 diagnostic、child 不可恢复或未授权、请求取消、图片准入或图片能力拒绝（`subagent/attachment-invalid`）以及继续执行准入暂时不可用等失败映射为类型化 RPC 错误。它不会公开描述符或提供方细节。list／prompt 竞态属于正常情况：权威依据是提示词操作的结果，而不是更早的可用性或活动快照。
 
 查看持久化历史本身不会创建 mux 订阅。当后续消息物化冷态 child Activation 时，现有 Host 与 mux 流会发布其生命周期与事件。重新连接时，系统通过 `subagent.history` 重建已寻址窗口。
 
-普通 `session.history` 路由对于普通会话和 subagent 会话同样只执行观察，但它既不携带目录地址，也不授予继续执行权限。每条需要 Agent 的普通路由都会在恢复冷会话前经过共享所有权栅栏；`session.cancel` 与 `session.updateQueue` 会直接执行同一检查，因为它们有意只查询已附加的 Agent。
+普通 `session.history` 路由对于普通会话和 subagent 会话同样只执行观察，但它既不携带目录地址，也不授予继续执行权限。每条需要 Agent 的普通路由都会在恢复冷会话前经过共享所有权栅栏；`session.cancel` 保留该栅栏。`session.updateQueue` 只有一个目标本地例外：目标是在线 child，且其当前 projection identity 为 continuable 并来自自身的非 seed suffix；one-shot、缺失、未知、损坏、仅含 seed identity 或冷 child 仍受栅栏阻挡。
 
 适配器仍位于生成的 Remote 命名空间之后；`dsh-host-webserver` 仍作为载体。浏览器代码通过现有连接包导入约定，绝不直接访问宿主 `ctx`，从而保持[已归档的 GUI RPC 分层决策](../../archived/architecture/2026-07-19-gui-layering-and-rpc-protocol.md)。
 

.agents/notes/implemented/feature/2026-07-28-continuable-subagent-conversations.i18n.yaml

@@ -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/2026-07-28-continuable-subagent-conversations.md
-2026-07-28-continuable-subagent-conversations.md: fef7aba2080521253d7a67dd5169880fa5146dc0
-2026-07-28-continuable-subagent-conversations.zh.md: 875d6ab7ea5b8dd78276c3a7a9e789646340bc15
+2026-07-28-continuable-subagent-conversations.md: 8c3f2e1da593157f17528f13fc8842012aad0284
+2026-07-28-continuable-subagent-conversations.zh.md: 886e31fa3d88b78f875d51058bb2ec8c525837fe

.agents/notes/implemented/feature/2026-07-28-continuable-subagent-conversations.md

@@ -62,14 +62,14 @@ The internal residency lifecycle has three conditions and no separate `queued` s
 
 ```text
 running
-  | Agent quiescent with live children
+  | Agent quiescent with pending inbox or live children
   v
 waiting
-  | next-turn
+  | waking delivery
   +--------------------------> running
 
 running or waiting
-  | Agent quiescent and no live children
+  | Agent quiescent, empty inbox, and no live children
   v
 settled
   | AgentHandle.dispose completes
@@ -77,15 +77,15 @@ settled
 no Activation
 ```
 
-`running` means the Agent has an active admission or turn, or its inbox contains waking work. `waiting` means the Agent is quiescent but the Activation still owns at least one child Activation that has not completed disposal. `settled` means the Agent is quiescent and every owned child is disposed; the manager then disposes the `AgentHandle` and removes the Activation.
+`running` means the Agent has an active admission or turn. `waiting` means the Agent is quiescent but its Inbox is nonempty or the Activation still owns at least one child Activation that has not completed disposal. `settled` means the Agent is quiescent, its Inbox is empty, and every owned child is disposed; the manager then disposes the `AgentHandle` and removes the Activation.
 
-The manager derives these states from Agent quiescence and the owned-child set rather than maintaining a second execution state machine. A `next-turn` delivered while `running` joins the Agent inbox. A `next-turn` delivered while `waiting` wakes the same Agent and returns the Activation to `running`. Delivery after disposal cold-resumes a new Activation.
+The manager derives these states from Agent quiescence, the Inbox's pending state, and the owned-child set rather than maintaining a second execution state machine. A `next-turn` delivered while `running` joins the Agent inbox. A waking delivery while `waiting` wakes the same Agent and returns the Activation to `running`. Delivery after disposal cold-resumes a new Activation.
 
-The manager linearizes delivery, child release, and disposal for each durable child. If a delivery races with final disposal, exactly one side wins the admission cutoff: delivery either enters the still-live Agent inbox, or waits for disposal and cold-resumes a new Activation. No caller can send to a handle after its disposal transaction begins.
+The manager linearizes manager-owned delivery, child release, and disposal for each durable child. A private `SubagentInbox` delegates Queue and Steer to the Agent inbox and owns the Activation's existing close transaction. If manager delivery races with final disposal, exactly one side wins this admission cutoff: delivery either enters the still-live Agent inbox, or observes closing and follows its operation-specific rejection or cold-resume path. Direct Agent work does not use this wrapper, so natural settlement uses short maintenance claims to validate the idle phase before the final flush and final disposal decision, then revalidates the Session sequence, Inbox pending state, wake generation, and owned-child set under the child lock. Accepted work that remains active or changes Session, Inbox, or ownership state invalidates that settlement attempt instead of being cancelled by it; maintenance that starts and finishes entirely during the flush has completed before the cutoff.
 
 ### One inbox and follow-up delivery
 
-The Agent inbox is the only queue. Every continuation message uses `Agent.followup()` and becomes one FIFO turn; neither the continuation manager nor the host maintains another message queue. Every accepted waking item keeps the current Activation live until `Agent.whenIdle()` observes the complete waking suffix.
+The Agent inbox is the only queue. Every continuation message uses `Agent.followup()` and becomes one FIFO turn; neither the continuation manager nor the host maintains another message queue. Every pending Inbox occurrence keeps the current Activation live until it is claimed or discarded. This conservative rule also retains injected context: a quiet injection that remains after quiescence can keep the Activation and its live ancestors resident until a waking delivery claims it, a queue mutation removes it, or manager teardown disposes the tree.
 
 Routing depends only on Activation residency:
 
@@ -103,21 +103,21 @@ Every Activation owns its `AgentHandle` and an `ownedChildren: Set<SessionId>`.
 
 When the authenticated parent is itself a continuation-managed Activation, starting a child or submitting parent-originated work adds the child Session id to that parent's `ownedChildren` before the child can run or the message can enter its inbox. That parent cannot settle or dispose while this set is non-empty. A top-level or other non-continuation Agent has no Activation and does not join this waiting graph.
 
-Child release occurs only after the child Agent is quiescent, every child of that child is disposed, the best-effort final session flush settles, and the child's `AgentHandle` completes disposal. The manager awaits `ctx.sessions.flush(child.session)` but does not interpret its participation boolean: an arbitrary listener cannot prove that the selected persistence backend stored the state. A rejection is logged without preventing handle disposal or ownership release, because retaining a child would permanently pin its ancestors in `waiting`. If the child is owned, the manager then resolves the live parent through `SessionHeader.parentSession` and removes the child Session id from its `ownedChildren`. Manager teardown uses the same child-first order.
+Child release occurs only after the child Agent is quiescent, its Inbox is empty, every child of that child is disposed, the best-effort final session flush settles, the same settlement facts survive a child-lock revalidation, and the child's `AgentHandle` completes disposal. The manager awaits `ctx.sessions.flush(child.session)` before closing admission but does not interpret its participation boolean: an arbitrary listener cannot prove that the selected persistence backend stored the state. A rejection is logged without preventing revalidation, handle disposal, or ownership release, because retaining a child would permanently pin its ancestors in `waiting`. If the child is owned, the manager then resolves the live parent through `SessionHeader.parentSession` and removes the child Session id from its `ownedChildren`. Manager teardown uses the same child-first order but closes admission and stops work immediately rather than performing natural-settlement revalidation.
 
 Ownership is retained until the child Activation is disposed. A later refinement may release a request-scoped lease earlier, but it would require an exact turn-completion correlation that this Task-free design deliberately does not add.
 
 Top-level teardown is host-owned rather than represented as another Activation. Manager unload invokes its internal manager-wide drain to close admission synchronously, await every admitted materialization through publication or rollback, stop the stable live forest, and release it child-first. A host that owns selected top-level Agents uses `drainContinuableDescendants(parents)`: exact Agent identities close admission only below those roots until each leaves the registry, while unrelated forests and manager-wide admission remain live; the manager stops their visible descendants before its first await, waits only materializations admitted below those roots, and releases only the selected branches. Every materialized start and live delivery rechecks caller cancellation, the applicable draining scope, Activation disposal, and exact parent authority in the same synchronous span as inbox submission, so teardown or parent replacement that wins before acceptance prevents delivery to the closing handle. Only after the applicable drain settles may the host dispose its top-level Agents; only manager-wide drain precedes manager-scope disposal.
 
-The activation-owner scope exists because ordinary Cordis owner effects unwind in reverse registration order, which cannot express the dynamic child graph. Manager initialization registers the private scope's structural disposer first and its drain disposer afterward, so reverse unwind invokes the drain before releasing that scope; merely registering a cleanup effect on the same scope as later Agent handles would allow structural handle disposal to bypass child-first ordering. Each materialization registers its barrier participant and snapshots its exact live ancestry before starting the inner transaction, then remains tracked until it installs an Activation or fully rolls back. The Activation retains weak membership of that ancestry, so an intermediate Agent may leave the registry without hiding a still-live descendant from its host root. Each Activation installs one memoized disposal promise before cancellation or recursive callbacks, allowing scoped host shutdown, global manager unload, child release, and normal settlement to converge without double release. Cancellation propagates top-down before slow descendant cleanup; handle release remains child-first. Sibling branches drain independently; one disposal failure is recorded but does not prevent the manager from attempting the remaining selected handles, and the aggregate drain reports failure after all selected branches settle. Durable child Sessions survive this process-local teardown.
+The activation-owner scope exists because ordinary Cordis owner effects unwind in reverse registration order, which cannot express the dynamic child graph. Manager initialization registers the private scope's structural disposer first and its drain disposer afterward, so reverse unwind invokes the drain before releasing that scope; merely registering a cleanup effect on the same scope as later Agent handles would allow structural handle disposal to bypass child-first ordering. Each materialization registers its barrier participant and snapshots its exact live ancestry before starting the inner transaction, then remains tracked until it installs an Activation or fully rolls back. The Activation retains weak membership of that ancestry, so an intermediate Agent may leave the registry without hiding a still-live descendant from its host root. Its private `SubagentInbox` installs one memoized closing promise before cancellation or recursive callbacks, allowing scoped host shutdown, global manager unload, child release, and normal settlement to converge without double release. Cancellation propagates top-down before slow descendant cleanup; handle release remains child-first. Sibling branches drain independently; one disposal failure is recorded but does not prevent the manager from attempting the remaining selected handles, and the aggregate drain reports failure after all selected branches settle. Durable child Sessions survive this process-local teardown.
 
 ### Adjacent-Agent messaging
 
 The shared `sendMessage(sender, targetId, content, options)` service operation adds no second queue. It accepts an exact live sender, permits only its direct parent or direct continuable child, and uses fixed Steer scheduling through the Agent inbox. The global `send_message({ agent_id, message })` tool exposes that same operation in both directions; the child's initial task identifies its direct parent when the tool is visible. The [adjacent-Agent messaging Agent Note](../architecture/2026-08-27-adjacent-agent-steer-messaging.md) owns its schema, authority, attribution, and prompt placement.
 
-### Fixed Steer scheduling
+### Agent and human scheduling
 
-Every accepted Agent message uses `Agent.steer()`. A running target claims it at the nearest step boundary; an idle or cold-resumed target starts a turn. The continuation layer does not expose a caller-selectable quiet, next-turn, or follow-up mode.
+Every accepted Agent message uses `Agent.steer()`. A running target claims it at the nearest step boundary; an idle or cold-resumed target starts a turn. Browser-authored human input separately carries `delivery: 'queue' | 'steer'` through `subagent.prompt`: Queue opens a later FIFO turn, while Steer uses the same best-effort nearest-step scheduling without changing the message's human provenance. The public service exposes no caller-selectable scheduling mode for Agent messages.
 
 ### Authority and recorded sender identity
 
@@ -133,7 +133,7 @@ Without Jobs there is no `job_output`, `job_kill`, Task status, or per-message r
 
 Host and manager teardown remains the lifecycle stop path. Manager unload applies it globally; a host applies it only below the exact top-level Agents it owns. Each form closes the applicable admission scope, stops the selected visible Activations, awaits admitted materializations in that scope, releases child-first, and preserves the durable Sessions.
 
-Each turn requests the Session durability checkpoint, while final Activation settlement additionally awaits `ctx.sessions.flush()` as a best-effort barrier. The manager deliberately ignores the boolean result because listener participation cannot identify a persistence backend. A rejection is logged without changing the lifecycle result or host-drain outcome; the manager still disposes the handle and releases ownership, and the persisted child state may be missing or stale on a later resume.
+Each turn requests the Session durability checkpoint, while final Activation settlement additionally awaits `ctx.sessions.flush()` as a best-effort barrier before closing admission. The manager then revalidates that no Agent, Inbox, Session, or owned-child state changed during the await; a changed observation retries settlement and flushes the newer state. The manager deliberately ignores the flush boolean because listener participation cannot identify a persistence backend. A rejection is logged without changing the lifecycle result or host-drain outcome; the manager still performs the final revalidation, disposes the handle when it succeeds, and releases ownership, while the persisted child state may be missing or stale on a later resume.
 
 Only messages written to the child Session log are reconstructable with the source that supplied them; inbox acceptance alone provides no restart guarantee.
 
@@ -189,13 +189,13 @@ The implementation pins these behaviors:
 - An idle Agent with live owned children yields a `waiting` Activation whose `AgentHandle` remains retained.
 - A `next-turn` delivered to `waiting` wakes the same Activation; delivery after completed disposal cold-resumes a new Activation.
 - Every continuation-managed parent Activation disposes only after all directly owned child Activations complete `AgentHandle` disposal; top-level Agents do not join the waiting graph.
-- Final Activation settlement awaits `ctx.sessions.flush(child.session)` as a best-effort barrier, logs rejection without interpreting listener participation as durability proof, then disposes the child handle and releases parent ownership so a flush failure cannot leak a `waiting` Activation.
+- Final Activation settlement awaits `ctx.sessions.flush(child.session)` with admission open, logs rejection without interpreting listener participation as durability proof, revalidates the final state under the child lock, then closes admission, disposes the child handle, and releases parent ownership so a flush failure cannot leak a `waiting` Activation.
 - Manager teardown closes admission globally; a host owning selected top-level Agents instead closes admission only below their exact identities until those roots leave the registry. Both track admitted materializations by exact ancestry, install one memoized disposal cutoff per selected visible Activation, propagate cancellation top-down, release handles child-first, await every selected branch despite individual failures, and only then dispose the corresponding top-level Agents or manager scope.
 - The base lifecycle has no implicit report behavior; the optional report package contributes an explicit child-scoped tool through the setup hook.
 - Session logs reconstruct only messages that were actually written, with the source that supplied each message; inbox-accepted but unlogged messages have no restart guarantee.
 - No continuable-subagent path creates or depends on a Task, `JobId`, Task completion notice, Task cancellation, or intermediate result-bearing execution wrapper.
 - Unit coverage pins the `startContinuable()` inbox-acceptance return boundary, complete rollback for each pre-acceptance and lifecycle-publication failure, global and parent-scoped drain quiescence for materialization caught between Agent publication and Activation registration, sibling-forest isolation, exact ancestry after an intermediate Agent leaves the registry, provider-independent cold resume, final exact-parent reauthorization after cold-resume materialization, caller-signal and teardown ownership on both sides of acceptance, and the absence of automatic replay for accepted-but-unlogged messages.
-- Unit coverage pins the residency-only routing table, single-inbox ordering, `MessageId` correlation through inbox events, follow-up during an open turn, waiting wakeup, cold resume, ownership registration and release, child-first disposal, send-versus-dispose races, best-effort final flush with absent and failing listeners, and the absence of public subagent cancellation and steering.
+- Unit coverage pins the residency-only routing table, single-inbox ordering, `MessageId` correlation through inbox events, follow-up during an open turn, waiting wakeup, cold resume, ownership registration and release, child-first disposal, send-versus-dispose races, direct Agent turns, Session-only work, and maintenance accepted during the final-flush await, best-effort final flush with absent and failing listeners, and the absence of public subagent cancellation and steering.
 - Report-package unit coverage separately pins child-only visibility, setup revocation, authority, delivery modes, stable message identity, and lifecycle races.
 - A keyless assembled-app snapshot covers parent delegation and follow-up queueing, the absence of subagent steering and implicit report delivery, retained waiting `AgentHandle`, and child-first disposal. A separate report snapshot covers the optional explicit return channel.
 

.agents/notes/implemented/feature/2026-07-28-continuable-subagent-conversations.zh.md

@@ -62,14 +62,14 @@ inbox 接受消息前发生任何失败，操作都会在不返回任何 id 的
 
 ```text
 running
-  | Agent quiescent with live children
+  | Agent quiescent with pending inbox or live children
   v
 waiting
-  | next-turn
+  | waking delivery
   +--------------------------> running
 
 running or waiting
-  | Agent quiescent and no live children
+  | Agent quiescent, empty inbox, and no live children
   v
 settled
   | AgentHandle.dispose completes
@@ -77,15 +77,15 @@ settled
 no Activation
 ```
 
-`running` 表示 Agent 正在执行准入或轮次，或者 inbox 中存在会唤醒 Agent 的工作。`waiting` 表示 Agent 已经完全停稳，但激活仍持有至少一个尚未完成 dispose 的 child 激活。`settled` 表示 Agent 已经完全停稳且所有持有的 child 都已 dispose；随后管理器会 dispose `AgentHandle` 并移除激活。
+`running` 表示 Agent 正在执行准入或轮次。`waiting` 表示 Agent 已经完全停稳，但其 Inbox 非空，或激活仍持有至少一个尚未完成 dispose 的 child 激活。`settled` 表示 Agent 已经完全停稳、其 Inbox 为空且所有持有的 child 都已 dispose；随后管理器会 dispose `AgentHandle` 并移除激活。
 
-管理器根据 Agent 是否完全停稳以及所持 child 集合派生这些状态，而不是维护第二套执行状态机。在 `running` 时投递的 `next-turn` 会进入 Agent inbox。在 `waiting` 时投递的 `next-turn` 会唤醒同一个 Agent，并使激活回到 `running`。在 dispose 完成后投递消息则会冷恢复新激活。
+管理器根据 Agent 是否完全停稳、Inbox 的待处理状态以及所持 child 集合派生这些状态，而不是维护第二套执行状态机。在 `running` 时投递的 `next-turn` 会进入 Agent inbox。在 `waiting` 时到达的唤醒投递会唤醒同一个 Agent，并使激活回到 `running`。在 dispose 完成后投递消息则会冷恢复新激活。
 
-管理器会针对每个持久化 child，将投递、child 释放和 dispose 线性化。如果投递与最终 dispose 发生竞争，只有一方能越过准入截止点：投递要么进入仍在线的 Agent inbox，要么等待 dispose 完成后冷恢复新激活。任何调用方都不能向已经开始 dispose 事务的 handle 发送消息。
+管理器会针对每个持久化 child，将 manager 所有的投递、child 释放和 dispose 线性化。私有 `SubagentInbox` 会把 Queue 与 Steer 委托给 Agent inbox，并持有 Activation 既有的关闭事务。如果 manager 投递与最终 dispose 发生竞争，只有一方能越过这条准入截止点：投递要么进入仍在线的 Agent inbox，要么观察到正在关闭，并遵循该操作特有的拒绝或冷恢复路径。直接操作 Agent 的工作不经过这层包装，因此自然结算会通过短暂的 maintenance 占用，在最终 flush 与最终 dispose 决策之前验证 idle 阶段，并在 child lock 内重新验证 Session 序号、Inbox 待处理状态、wake generation 与 owned-child set。仍然活跃或改变 Session、Inbox 或所有权状态的已接受工作会让本次结算尝试失效，而不会被它取消；完全在 flush 期间开始并结束的 maintenance 已在截止点前完成。
 
 ### 一个 inbox 与 follow-up 投递
 
-Agent inbox 是唯一队列。每条继续执行消息都使用 `Agent.followup()`，并成为一个 FIFO 轮次；继续执行管理器和宿主都不维护另一条消息队列。每个已接受且会唤醒 Agent 的条目都会让当前激活保持在线，直至 `Agent.whenIdle()` 观察到完整的唤醒工作后缀已经结束。
+Agent inbox 是唯一队列。每条继续执行消息都使用 `Agent.followup()`，并成为一个 FIFO 轮次；继续执行管理器和宿主都不维护另一条消息队列。每个待处理 Inbox occurrence 都会让当前激活保持在线，直到它被 claim 或 discard。这条保守规则也会保留注入 context：完全停稳后仍存在的静默注入可以让 Activation 及其在线祖先继续驻留，直到唤醒投递将其 claim、queue 变更将其移除，或 manager teardown dispose 整棵树。
 
 路由只取决于激活的驻留状态：
 
@@ -103,21 +103,21 @@ Agent inbox 是唯一队列。每条继续执行消息都使用 `Agent.followup(
 
 当经过身份认证的 parent 自身是由继续执行管理器管理的激活时，启动 child 或提交由 parent 发起的工作，会在 child 可以运行或消息可以进入其 inbox 前，将 child 会话 id 加入该 parent 的 `ownedChildren`。该集合非空时，这个 parent 不能结算或 dispose。顶层 Agent 或其他非继续执行 Agent 没有激活，也不会加入该等待图。
 
-只有在 child Agent 完全停稳、该 child 持有的每个 child 都已 dispose、best-effort 的最终会话 flush 结算且 child 的 `AgentHandle` 完成 dispose 后，系统才释放 child。管理器会等待 `ctx.sessions.flush(child.session)`，但不解释其参与布尔值：任意 listener 都无法证明所选持久化后端已存储该状态。rejection 会被记录，但不会阻止 handle dispose 或释放所有权，因为保留 child 会让其祖先永久固定在 `waiting`。如果 child 归 parent 所有，管理器随后会通过 `SessionHeader.parentSession` 解析在线 parent，并从其 `ownedChildren` 中移除 child 会话 id。管理器拆卸使用相同的 child-first 顺序。
+只有在 child Agent 完全停稳、其 Inbox 为空、该 child 持有的每个 child 都已 dispose、best-effort 的最终会话 flush 结算、相同结算事实通过 child-lock 重验且 child 的 `AgentHandle` 完成 dispose 后，系统才释放 child。管理器会在关闭准入前等待 `ctx.sessions.flush(child.session)`，但不解释其参与布尔值：任意 listener 都无法证明所选持久化后端已存储该状态。系统会记录 rejection，但不会让它阻止重验、handle dispose 或释放所有权，因为保留 child 会让其祖先永久固定在 `waiting`。如果 child 归 parent 所有，管理器随后会通过 `SessionHeader.parentSession` 解析在线 parent，并从其 `ownedChildren` 中移除 child 会话 id。Manager teardown 使用相同的 child-first 顺序，但会立即关闭准入并停止工作，而不执行自然结算重验。
 
 系统会一直保留所有权，直至 child 激活完成 dispose。后续改进可以更早释放限定到请求的 lease，但这需要精确关联轮次完成，而本 Task-free 设计特意不增加该机制。
 
 顶层拆卸由宿主负责，而不表示为另一次激活。管理器卸载会调用其内部的管理器全局 drain，同步关闭准入，等待每个已获准的物化过程完成发布或回滚，停止稳定的在线森林，并按 child-first 顺序释放。拥有选定顶层 Agent 的宿主使用 `drainContinuableDescendants(parents)`：确切的 Agent 身份只关闭这些根之下的准入，直到每个身份离开注册表，而无关森林和管理器全局准入保持在线；管理器会在第一次 await 之前停止其可见后代，只等待这些根之下已获准的物化过程，并且只释放选定分支。每个已物化的 start 和在线投递都会在与 inbox 提交相同的同步区间内重新检查调用方取消、适用的 draining 作用域、Activation dispose 和确切的 parent 权限，因此只要拆卸或 parent 替换先于接受发生，就会阻止向正在关闭的 handle 投递。只有适用的 drain 结算后，宿主才能 dispose 自己的顶层 Agent；只有管理器全局 drain 会先于管理器作用域 dispose。
 
-activation-owner 作用域之所以存在，是因为普通 Cordis owner effect 按注册逆序撤销，无法表达动态 child 图。管理器初始化时先注册私有作用域的结构化 disposer，再注册自身的 drain disposer，使逆序撤销先执行 drain、再释放该作用域；如果只在与后续 Agent handle 相同的作用域上注册 cleanup effect，结构化 handle dispose 就可能绕过 child-first 顺序。每个物化过程都会在启动内部事务前注册其屏障参与项，并对其确切的在线祖先建立快照，然后保持跟踪，直到安装 Activation 或完全回滚。Activation 会保留其在这组祖先中的弱成员关系，因此中间 Agent 即使离开注册表，也不会让仍在线的后代脱离宿主根节点的可见范围。每个 Activation 都会在取消或递归回调前安装一个记忆化的 dispose promise，使限定作用域的宿主关闭、全局管理器卸载、child 释放和正常结算能够汇合，而不会重复释放。取消会在等待缓慢的后代清理之前自顶向下传播；handle 释放仍是 child-first。同级分支独立 drain；系统会记录单次 dispose 失败，但仍会尝试其余选中 handle，聚合 drain 则在所有选中分支结算后报告失败。这次进程内拆卸不会销毁持久化 child 会话。
+activation-owner 作用域之所以存在，是因为普通 Cordis owner effect 按注册逆序撤销，无法表达动态 child 图。管理器初始化时先注册私有作用域的结构化 disposer，再注册自身的 drain disposer，使逆序撤销先执行 drain、再释放该作用域；如果只在与后续 Agent handle 相同的作用域上注册 cleanup effect，结构化 handle dispose 就可能绕过 child-first 顺序。每个物化过程都会在启动内部事务前注册其屏障参与项，并对其确切的在线祖先建立快照，然后保持跟踪，直到安装 Activation 或完全回滚。Activation 会保留其在这组祖先中的弱成员关系，因此中间 Agent 即使离开注册表，也不会让仍在线的后代脱离宿主根节点的可见范围。其私有 `SubagentInbox` 会在取消或递归回调前安装一个记忆化的 closing promise，使限定作用域的宿主关闭、全局管理器卸载、child 释放和正常结算能够汇合，而不会重复释放。取消会在等待缓慢的后代清理之前自顶向下传播；handle 释放仍是 child-first。同级分支独立 drain；系统会记录单次 dispose 失败，但仍会尝试其余选中 handle，聚合 drain 则在所有选中分支结算后报告失败。这次进程内拆卸不会销毁持久化 child 会话。
 
 ### 相邻 Agent 消息
 
 共享的 `sendMessage(sender, targetId, content, options)` 服务操作不会增加第二条队列。它接收确切在线 sender，只允许其直接 parent 或直接可继续 child，并通过 Agent inbox 使用固定 Steer 调度。全局 `send_message({ agent_id, message })` 工具在两个方向暴露同一个操作；当 child 可以看到该工具时，其初始任务会标明直接 parent。[相邻 Agent 消息 Agent Note](../architecture/2026-08-27-adjacent-agent-steer-messaging.zh.md)规定其 schema、权限、来源信息与提示词位置。
 
-### 固定 Steer 调度
+### Agent 与人类调度
 
-每条已接受的 Agent 消息都使用 `Agent.steer()`。运行中的目标会在最近的 step 边界领取消息；空闲或冷恢复的目标会启动一个轮次。继续执行层不暴露由调用方选择的 quiet、next-turn 或 follow-up 模式。
+每条已接受的 Agent 消息都使用 `Agent.steer()`。运行中的目标会在最近的 step 边界领取消息；空闲或冷恢复的目标会启动一个轮次。浏览器编写的人类输入会另行通过 `subagent.prompt` 携带 `delivery: 'queue' | 'steer'`：Queue 开启后续 FIFO 轮次，Steer 使用相同的 best-effort 最近 step 调度，并保留消息的人类来源。公开服务不为 Agent 消息提供调用方可选的调度模式。
 
 ### 权限与已记录的发送方身份
 
@@ -133,7 +133,7 @@ activation-owner 作用域之所以存在，是因为普通 Cordis owner effect
 
 宿主和管理器拆卸仍是生命周期停止路径。管理器卸载会全局应用它；宿主只会在自己确切拥有的顶层 Agent 之下应用它。两种形式都会关闭适用的准入作用域，停止选中的可见 Activation，等待该作用域中已获准的物化过程，按 child-first 顺序释放，并保留持久化 Session。
 
-每个轮次都会请求执行会话持久性检查点，而 Activation 最终结算还会等待 `ctx.sessions.flush()`，将其作为 best-effort 屏障。管理器特意忽略布尔结果，因为 listener 是否参与无法标识持久化后端。rejection 会被记录，但不会改变生命周期结果或宿主 drain 的结果；管理器仍会 dispose handle 并释放所有权，后续恢复时持久化 child 状态可能缺失或陈旧。
+每个轮次都会请求执行会话持久性检查点，而 Activation 最终结算还会在关闭准入前等待 `ctx.sessions.flush()`，将其作为 best-effort 屏障。管理器随后会重新验证 await 期间没有 Agent、Inbox、Session 或 owned-child 状态发生变化；观察发生变化时，系统会重试结算并 flush 更新后的状态。管理器特意忽略 flush 布尔结果，因为 listener 是否参与无法标识持久化后端。系统会记录 rejection，但不会改变生命周期结果或宿主 drain 的结果；管理器仍会执行最终重验，在重验成功时 dispose handle 并释放所有权，后续恢复时持久化 child 状态可能缺失或陈旧。
 
 只有实际写入 child 会话日志的消息，才能在重建时保留提供它的来源；仅被 inbox 接受并不提供重启保证。
 
@@ -189,13 +189,13 @@ activation-owner 作用域之所以存在，是因为普通 Cordis owner effect
 - 带有在线所持 child 的空闲 Agent 会产生 `waiting` 激活，其 `AgentHandle` 继续保留。
 - 向 `waiting` 投递 `next-turn` 会唤醒同一个激活；完成 dispose 后投递消息会冷恢复新激活。
 - 每个由继续执行管理器管理的 parent 激活只会在直接持有的所有 child 激活完成 `AgentHandle` dispose 后进行 dispose；顶层 Agent 不加入等待图。
-- Activation 最终结算会等待 `ctx.sessions.flush(child.session)`，将其作为 best-effort 屏障；它会记录 rejection，但不会把 listener 参与解释为持久性证明，然后 dispose child handle 并释放 parent 所有权，使 flush 失败不会泄漏 `waiting` Activation。
+- Activation 最终结算会在准入开放时等待 `ctx.sessions.flush(child.session)`，将其作为 best-effort 屏障；它会记录 rejection，但不会把 listener 参与解释为持久性证明，随后在 child lock 内重新验证最终状态，再关闭准入、dispose child handle 并释放 parent 所有权，使 flush 失败不会泄漏 `waiting` Activation。
 - 管理器拆卸会全局关闭准入；拥有选定顶层 Agent 的宿主则只关闭这些确切身份之下的准入，直到这些根离开注册表。两者都会按确切祖先关系跟踪已获准的物化过程，为每个选中的可见 Activation 安装一个记忆化 dispose 截止点，自顶向下传播取消，按 child-first 顺序释放 handle，即使个别分支失败也会等待所有选中分支，之后才 dispose 对应的顶层 Agent 或管理器作用域。
 - 基础生命周期不暴露隐式报告行为；可选的 report 包通过 setup 钩子贡献一个显式的 child 作用域工具。
 - 会话日志只会重建实际写入的消息，并保留每条消息的提供来源；已被 inbox 接受但未写入日志的消息没有重启保证。
 - 可继续 subagent 路径不创建或依赖 Task、`JobId`、Task 完成通知、Task 取消或中间的带结果执行包装层。
 - 单元覆盖固定 `startContinuable()` 在 inbox 接受消息时的返回边界、每条接受前和生命周期发布失败路径的完整回滚、全局和限定到 parent 作用域的 drain 都会等待夹在 Agent 发布与 Activation 注册之间的物化过程完全停稳、同级森林隔离、中间 Agent 离开注册表后的确切祖先关系、不依赖提供方的冷恢复、冷恢复物化后的最终确切 parent 再授权、接受前后两个阶段的调用方 signal 与拆卸所有权，以及已接受但未写入日志的消息不会自动回放。
-- 单元覆盖固定仅由驻留状态决定的路由表、单 inbox 顺序、通过 inbox 事件关联 `MessageId`、在开放轮次期间 follow-up、等待唤醒、冷恢复、所有权注册与释放、child-first dispose、发送与 dispose 的竞争、没有 listener 和 listener 失败时的 best-effort 最终 flush，以及不存在公开 subagent 取消和 steering。
+- 单元覆盖固定仅由驻留状态决定的路由表、单 inbox 顺序、通过 inbox 事件关联 `MessageId`、在开放轮次期间 follow-up、等待唤醒、冷恢复、所有权注册与释放、child-first dispose、发送与 dispose 的竞争、在最终 flush await 期间接受的直接 Agent 轮次、仅修改 Session 的工作与 maintenance、没有 listener 和 listener 失败时的 best-effort 最终 flush，以及不存在公开 subagent 取消和 steering。
 - report 包的单元覆盖会分别固定仅 child 可见性、setup 撤销、权限、投递模式、稳定消息身份和生命周期竞争。
 - 一项无密钥整套应用快照覆盖 parent 委派和 follow-up 排队、不存在 subagent steering 和隐式 report 投递、保留 waiting 中的 `AgentHandle` 以及 child-first dispose。另一项 report 快照覆盖可选的显式返回通道。
 

.agents/notes/implemented/feature/2026-08-06-manager-owned-subagent-settlement-delivery.i18n.yaml

@@ -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/2026-08-06-manager-owned-subagent-settlement-delivery.md
-2026-08-06-manager-owned-subagent-settlement-delivery.md: d06245eacd3b7453a031716b1921015a5e38a25c
-2026-08-06-manager-owned-subagent-settlement-delivery.zh.md: e5259e97f185203ed77ae9427e5523ac29d4162f
+2026-08-06-manager-owned-subagent-settlement-delivery.md: f223571dc91300d085b7fcf0a9e3196daa48b760
+2026-08-06-manager-owned-subagent-settlement-delivery.zh.md: 4bbee373aa2b50902af5319f9398a89da9cc3143

.agents/notes/implemented/feature/2026-08-06-manager-owned-subagent-settlement-delivery.md

@@ -26,9 +26,9 @@ The notice carries `{ kind: 'subagent-settled', form: 'notice', summary, senderS
 
 An external `ctx.on('subagent/end')` listener looks more decoupled and is wrong. `SubagentRunEndInfo` names no parent, the child handle is already disposed when the edge fires so the parent cannot be recovered from it, and the ownership release that wakes the parent's own settlement watcher has already run. The manager holds the parent reference throughout disposal, so none of those obstacles exist for it.
 
-**The send happens before `releaseOwnership`.** At that point the parent still counts this child, so `stateOf(parent)` is `waiting` and the parent is structurally unable to be judged settled. Delivering after the release instead races a watcher that resumes one microtask later, finds itself childless and quiet, and disposes an Agent whose `cancel()` clears the very inbox the notice is sitting in. The failure mode is a silently missing message with no error anywhere.
+**The send happens before `releaseOwnership`.** At that point the parent's owned-child set still contains this child, so the settlement predicate cannot succeed. Delivering after the release instead races a watcher that resumes one microtask later, finds itself childless and quiet, and disposes an Agent whose `cancel()` clears the very inbox the notice is sitting in. The failure mode is a silently missing message with no error anywhere.
 
-**A resident parent receives it through `admitWaking`.** Registering the message id before the synchronous send is what keeps the window between `followup()` and the microtask that admits it from being read as quiescence. This is not belt-and-braces over the first rule: `Agent.status` folds context maintenance into `idle`, and a waking send behind maintenance only arms a deferred wake, so a parent compacting its context is judged quiet by both `status` and the owned-child set the moment the release lands.
+**A resident parent receives it through its private `SubagentInbox`.** The wrapper checks the Activation's closing promise immediately before the synchronous waking send, and the manager renews the wake generation before returning. The final settlement decision rechecks that generation, the Session sequence, the pending Inbox, and the owned-child set under the child lock, then claims the Agent's idle phase through `runMaintenance()` before closing admission. This is not redundant with the first rule: `Agent.status` folds context maintenance into `idle`, and a waking send behind maintenance only arms a deferred wake.
 
 Both rules are pinned by tests that fail when the ordering is reversed or the accounting removed.
 

.agents/notes/implemented/feature/2026-08-06-manager-owned-subagent-settlement-delivery.zh.md

@@ -26,9 +26,9 @@ Status: implemented
 
 外部 `ctx.on('subagent/end')` listener 看起来更解耦，但它是错的。`SubagentRunEndInfo` 不指名父级；该边触发时 child handle 已被 dispose，因此无法从中恢复父级；而唤醒父级自身结算 watcher 的所有权释放也已经执行过了。管理器在整个 dispose 过程中都持有父级引用，因此这些障碍对它都不存在。
 
-**发送发生在 `releaseOwnership` 之前。** 此刻父级仍然计入这个 child，因此 `stateOf(parent)` 为 `waiting`，父级在结构上不可能被判定为已结算。改在释放之后投递，则会与一个在下一个 microtask 恢复的 watcher 竞争：它会发现自己没有 child 且处于静止，于是 dispose 一个 Agent，而该 Agent 的 `cancel()` 会清空正装着这条通知的那个 inbox。失效表现是一条静默丢失的消息，任何地方都不会报错。
+**发送发生在 `releaseOwnership` 之前。** 此刻 parent 的 owned-child set 仍然包含这个 child，因此结算判据不可能成立。改在释放之后投递，则会与一个在下一个 microtask 恢复的 watcher 竞争：它会发现自己没有 child 且处于静止，于是 dispose 一个 Agent，而该 Agent 的 `cancel()` 会清空正装着这条通知的那个 inbox。失效表现是一条静默丢失的消息，任何地方都不会报错。
 
-**驻留父级通过 `admitWaking` 接收它。** 在同步发送之前登记消息 id，正是让 `followup()` 与承认它的那个 microtask 之间的窗口不被读作静止的原因。这不是对第一条规则的多余保险：`Agent.status` 会把上下文维护折叠成 `idle`，而维护期间的唤醒发送只会预置一次延后唤醒，因此正在压缩上下文的父级，在所有权释放落地的那一刻会同时被 `status` 与已拥有 child 集合判定为静止。
+**驻留 parent 通过私有 `SubagentInbox` 接收它。** 包装层会在同步唤醒发送前立即检查 Activation 的 closing promise，manager 则会在返回前更新 wake generation。最终结算决策会在 child lock 内重新检查该 generation、Session 序号、待处理 Inbox 与 owned-child set，再通过 `runMaintenance()` 占用 Agent 的 idle 阶段，然后关闭准入。这并非对第一条规则的重复保护：`Agent.status` 会把 context maintenance 折叠成 `idle`，而 maintenance 期间的唤醒发送只会预置一次延后唤醒。
 
 两条规则都有测试固定：把顺序反转或去掉记账，测试就会失败。
 

.agents/notes/implemented/feature/2026-08-25-promote-open-anywhere-plugin.i18n.yaml

@@ -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/2026-08-25-promote-open-anywhere-plugin.md
+2026-08-25-promote-open-anywhere-plugin.md: ee83c424d1454b26c1ce6cf6954105cdbfbb7419
+2026-08-25-promote-open-anywhere-plugin.zh.md: f1696cec10a683d44dcaa3db454d343821fc13c9

.agents/notes/implemented/feature/2026-08-25-promote-open-anywhere-plugin.md

@@ -0,0 +1,57 @@
+# Agent Note: Promote open-anywhere from community plugin to first-party package
+
+Status: implemented
+
+English | [中文](2026-08-25-promote-open-anywhere-plugin.zh.md)
+
+## Problem
+
+The community plugin `@dsh-plugins/open-anywhere` (gitlab.deepseek.com/Ciyou/dsh-open-anywhere) adds a Session-header "Open In..." split button that opens the session's workspace directory in Finder, Cursor, VS Code, Xcode, a Git GUI, or a terminal. It shipped as hand-authored `lib/` JavaScript installed through `dsh plugin add`: untyped, untested, calling `node:child_process` directly, hand-rolling its own dropdown and style tag, carrying a browser-side DSH-version gate against rc6–rc8, and probing `process.argv` to guess the running dsh version. Users wanted the feature as a shipped part of the Web profile, which the bundle-install path cannot give it — and the external form violates nearly every repository convention (locale-owned copy, per-file coverage, wire-boundary validation, capability seams for host commands).
+
+## Decision
+
+The first-party feature is named `open-in-app`: it selects the application that opens a workspace directory on the Harness host, not another machine or destination.
+
+The feature's first-party owners are `@deepseek-ai/dsh-host-open-in-app` at `packages/host/open-in-app/` (the probe, catalog, and launch routes) and `@deepseek-ai/dsh-client-ui-open-in-app` at `packages/client/ui-open-in-app/` (the split button), mounted in the Web profile by the `dsh-web-app` bundle rows `open-in-app` and `ui-open-in-app`. The promotion is a rewrite, not a vendoring:
+
+- **A host/client package pair, following the `directory-picker-browse`/`ui-directory-picker-browse` pairing**: the host package's `src/index.ts` registers the three HTTP routes on `ctx.webServer` (`GET /open-in-app/apps`, `GET /open-in-app/icon/<id>`, `POST /open-in-app/open`); the ui package's `src/client/index.ts` registers the split button into `conversation.session.header.utilities` through the standard slot/inject currency, with copy in a typed `open-in-app` locale namespace and styling in CSS Modules over `--dsw-*` tokens (the original's hand-injected style tag and inline dropdown are replaced by the `Menu` primitive), over an empty-apply node half that keeps the plugin on the host roster. Route paths and wire payload types have one home, the host package's browser-safe `./shared` subpath (constants and types only); the client bundle inlines it through an `INLINE_SAFE` entry in the client tsdown preset, the same channel `dsh-session`'s wire slices use. The host root exports only the Loader-required plugin values and types; catalog, resolver, launcher, and icon helpers remain source-internal.
+- **One resolution pass yields verified launchers; a click never re-detects.** The host resolves the whole catalog lazily once per process into a map of catalog id to `OpenInAppResolvedLaunch` — a launcher this host actually holds, never a bare install record. `GET /apps` serves the map's keys and `POST /open` launches its value directly; a launch whose executable is gone (spawn `ENOENT`) invalidates that one entry, re-resolves it once, and drops it from the list when nothing proves it anymore (so uninstalls self-heal while new installs wait for a restart).
+- **Resolution sources are platform-honest and cheap.** macOS checks the known application directories (`/Applications`, `~/Applications`) for the entry's bundle spellings and launches `open -a <resolved bundle>`; Xcode follows `xcode-select -p`. Windows reads `App Paths`, then Uninstall records kept only when they prove an executable on disk, then known paths and the newest versioned install directory where an application uses one — one batched `reg.exe query` per root per pass. GitHub Desktop resolves its versioned executable together with the packaged `cli.js` and invokes the supported `github open <path>` behavior without a command shell. CLI names resolve in-process through `ctx.subprocess.resolveExecutable()` (PATH/PATHEXT stat, no shell, no `which`/`where.exe` children); Linux GUI entries whose CLI is off PATH fall back to their XDG desktop entry's verified `TryExec`/`Exec`, while `xdg-open` is offered only when the host announces a display server. The remaining host commands (`xcode-select`, `reg.exe`, icon extraction) run through `@deepseek-ai/dsh-native-command` (argv, never a shell).
+- **Three independent deadlines replace the single `commandTimeoutMs`**: `probeTimeoutMs` (resolution commands), `iconTimeoutMs` (icon extraction commands), and `launchWatchMs` (the early-failure watch window), so tuning one operation never changes another's response time — the shipped bundle keeps conservative 10 s command deadlines (timeouts are failure bounds, not latency budgets) with a 1 s watch window, which is what bounds how long the open route and the button's busy dress hold a successful launch. Launches spawn detached with a credential-scrubbed environment (`scrubbedParentEnv` from `dsh-subprocess`) plus explicit adapter entries; Windows GUI launchers remain visible unless an adapter explicitly hides its CLI process because that process opens the GUI separately. A launcher still running when the watch window closes counts as launched and is never killed or awaited (kitty and the JetBrains IDEs stay in the foreground for their window lifetime).
+- **Icons extract on every platform with a host source.** macOS converts the resolved bundle's `.icns` to a 128px PNG (`plutil` + `sips`); Windows extracts the resolved executable's associated icon as a 32px PNG through a generated PowerShell script run with positional `-File` args (no command-line path parsing); Linux follows the spec's desktop entry `Icon=` into the hicolor theme and pixmaps directories (PNG or SVG, filesystem only). Any failure answers 404 and the browser keeps its generic glyph.
+- **The launch catalog is a data table with per-platform entries** (`OPEN_IN_APP_CATALOG`): each application id declares, per platform (`darwin`/`win32`/`linux`), a chain of launcher sources tried in order — `fixed` (ships with the OS), `app` (known-directory bundle spellings), `xcode` (`xcode-select -p`), `cli` (in-process PATH resolution), `file` (first existing candidate under `${VAR}`/`~/` expansion), `scan` (newest versioned install directory), `app-paths` and `install-record` (the Windows registry tiers), `github-desktop` (the versioned executable plus packaged CLI), and `desktop` (Linux XDG desktop entries) — plus a launcher argv template (a `{path}` token carries the directory in place, otherwise it is appended), optional environment and Windows visibility policy, and an optional fallback launcher (Xcode's `open -a <bundle>` behind `xed`). The whitelist follows Codex's "Open In" target list: editors and IDEs (VS Code, VS Code Insiders, Cursor, Windsurf, Zed, Sublime Text, Xcode, Android Studio, seven JetBrains IDEs), the promoted plugin's Git GUIs, terminals (Terminal, iTerm2, Ghostty, Warp, kitty, Windows Terminal, Git Bash, GNOME Terminal, Konsole), and per-platform file managers. File managers and platform terminals are separate ids (`finder`/`explorer`/`filemanager`) rather than one id with per-platform labels, because labels are static browser dictionary entries and only one of them probes as available per host. The file-manager entries launch through `dsh-native-command`'s path opener itself (`shell-open`, the OS shell's open verb with the full parent environment, not a detached scrubbed spawn), because a direct `explorer.exe <dir>` spawn does not reliably raise a window. Closed unions end in `assertNever`.
+- **Every route runs behind the composition connection service's trust fence** (`requestRejection`: the Host/Origin fence defeating DNS rebinding and cross-site calls, plus browser authentication), the same guard the API gateway applies to its WebSocket upgrade; the mechanism's one home is the `src/index.ts` module comment. On top of that fence the open route validates the body at the wire: an `application/json` media type (exact essence, not a substring match), a 64 KiB bound with a drained 413, string field types, only probed-available catalog ids, and an absolute path naming an existing directory.
+- **The DSH-version gate is deleted.** It existed because the plugin rode release-to-release against an interface it did not own; a first-party package is versioned with the repository, so the gate, its `sessionStorage`/`localStorage` trust ledger, and the argv-walking version probe have no referent.
+- **The last choice persists through `createSnapshotStore(..., { persist })`** (`dsh.open-in-app.choice`), replacing hand-rolled `localStorage` access. A fresh store has no platform-specific choice; the component uses the first available host entry until the user chooses one.
+
+The pair lives in `packages/host/` and `packages/client/` because that is what the halves are: the probe/launch side is host infrastructure beside the webserver it consumes, and the button is a client surface beside the other `ui-*` packages. Review moved it there from a single dual-half package in `packages/workspace/` (see Alternatives).
+
+## Alternatives considered
+
+**Vendor the plugin's `lib/` as-is under `packages/`.** Fastest, but the hand-authored JavaScript fails typecheck, coverage, i18n, JSDoc, and invariant gates wholesale; keeping it exempt would create a package class the repository deliberately does not have.
+
+**A Typert Remote instead of raw webServer routes.** The apps/open calls fit the Remote RPC shape, but the icon route serves binary PNGs, which the JSON RPC vocabulary does not carry; splitting icons onto a raw route while apps/open ride Remote gives two transports for one feature. Raw routes also match the original's client, and `webhook-github` establishes the validated-raw-route pattern.
+
+**Extend `host/apiproxy`'s `openPath` instead of a new open endpoint.** `openPath` opens one path with the OS-default application; this feature's subject is *which* application, with availability probing and per-application launchers — a different contract. Both share `dsh-native-command`.
+
+**One dual-half package in `packages/workspace/` (the shape that shipped first, following `dsh-session-log-export`).** Split during review into the host/client pair: the workspace group's contract is host-side only, the feature consumes `webServer` rather than `workspaceRegistry`, and registering the single package on the Client compiler aggregate forced the host route and catalog tests to pose as `.client.spec.ts`. The split puts each half's tests on its own compiler face and its dependencies in the right sections; the wire contract stayed in one home via the host package's `./shared` subpath.
+
+**A configurable catalog (cordis.yml-defined applications).** Deferred: each entry couples discovery, launch arguments, process policy, and icon behavior, so a user-facing settings owner and validation rules are required before accepting arbitrary commands. User-supplied labels are user data and do not conflict with locale-owned product copy. Codex and Orca demonstrate the likely extension: maintained built-in presets plus configurable custom handlers.
+
+**Enumerating every installed application through operating-system APIs.** Rejected as the menu's authority: Launch Services, Windows registration data, and XDG desktop entries can locate applications, but they do not establish which applications accept a workspace directory or which launch protocol opens it correctly. OS-native identifiers remain useful locator inputs for maintained presets; custom handlers cover the long tail without guessing launch semantics.
+
+**Shipping static icons for Windows/Linux entries (Codex bundles PNGs per target).** Rejected: the icon route serves real host icons on all three platforms, and bundling third-party product artwork adds an asset pipeline and trademark surface for cosmetic gain.
+
+**Subprocess-heavy detection (the shape that shipped first): `open -Ra` per macOS entry, `which`/`where.exe` per CLI, and a re-resolution on every launch.** Replaced during review: a list resolution spawned ~26 children on macOS, display-name Launch Services queries are weaker evidence than an on-disk bundle, the repository already carries in-process PATH resolution (`ctx.subprocess.resolveExecutable()`), and re-detecting on click put the probe deadline on the interactive path. A batched-`mdfind` fallback for relocated macOS bundles was also considered and left out with the review's known-paths instruction; the miss is recorded as a Known Limitation. A native LaunchServices/NSWorkspace lookup would need an addon the repository does not carry — deferred, with known `.app` paths as the stand-in.
+
+**A public application-discovery Service Definition.** Deferred on the single-consumer rule: the resolver stays a package-internal module until a second GUI discovery consumer exists.
+
+**128px Windows icons through an `SHDefExtractIcon` P/Invoke (`Add-Type`) script.** Deferred: `ExtractAssociatedIcon` is the stock .NET surface with no compiled snippet, and 32px only softens slightly at the button's 15-18 CSS px on high-DPI displays; the P/Invoke variant is a script-local upgrade if that softness matters in practice. Following the user's active Linux icon theme was likewise left out — hicolor is the freedesktop fallback every theme inherits from — so themed desktops may see the stock icon.
+
+## Consequences
+
+- The Web profile gains the header button wherever the host probes at least one installed catalog application on macOS, Windows, or Linux, with zero rendering elsewhere (empty probed catalog → the component returns null).
+- The community plugin's install path remains valid but redundant; its original routes and browser choice key are separate from `open-in-app`, so installations using the first-party feature should remove the community plugin to avoid duplicate header controls.
+- Resolution and icons run lazily, once per host process, so an application installed while dsh runs appears only after restart — accepted; the uninstall direction self-heals through the `ENOENT` single-entry refresh.
+- The catalog is compile-time fixed; extending it means editing `OPEN_IN_APP_CATALOG` and both locale dictionaries together (README Known Limitations). Platform coverage is uneven — several Git GUIs and terminals are macOS-only entries, Windows icons are limited to the 32px stock .NET extraction, Linux follows hicolor rather than the active theme, and CLI-only entries without a desktop record keep the generic icon.
+- Coverage: resolver logic (every locator kind over temp filesystems, registry-dump and desktop-entry fixtures, an injected env/home/PATH table), per-platform icon extraction, the three routes (real Loader + real WebServer composition, including the one-pass cache, the `ENOENT` refresh, and HMR-safety disposal), controller wire behavior, and component presentation are unit-tested to the per-file 100% gate; no snapshot is added because the shipped keyless snapshot fixtures assert session-driven output, which this browser-side control never touches. The web ARIA goldens disable the `open-in-app` and `ui-open-in-app` rows, and the Host-only preset e2e composition disables the host row: the button reflects whatever applications the running machine has installed, so its presence and label are host facts no cross-platform golden can pin.

.agents/notes/implemented/feature/2026-08-25-promote-open-anywhere-plugin.zh.md

@@ -0,0 +1,57 @@
+# Agent Note: 将 open-anywhere 从社区插件转正为第一方包
+
+Status: implemented
+
+[English](2026-08-25-promote-open-anywhere-plugin.md) | 中文
+
+## 问题
+
+社区插件 `@dsh-plugins/open-anywhere`（gitlab.deepseek.com/Ciyou/dsh-open-anywhere）在会话头部增加一个 "Open In..." 分体按钮，可在 Finder、Cursor、VS Code、Xcode、Git GUI 或终端中打开会话的 workspace 目录。它以手写 `lib/` JavaScript 形式经 `dsh plugin add` 安装：无类型、无测试、直接调用 `node:child_process`、手搓下拉菜单和 style 标签、自带针对 rc6–rc8 的浏览器端 DSH 版本门禁，并靠探测 `process.argv` 猜测运行中的 dsh 版本。用户希望该功能成为 Web profile 的内置部分，而 bundle 安装路径给不了这一点——且外部形态几乎违反了仓库的所有约定（locale 拥有文案、逐文件覆盖率、wire 边界校验、主机命令的能力接缝）。
+
+## 决定
+
+第一方功能命名为 `open-in-app`：它选择在 Harness 主机上打开 workspace 目录的应用，不表示另一台机器或目的位置。
+
+该功能的第一方归属是一对包：`@deepseek-ai/dsh-host-open-in-app` 位于 `packages/host/open-in-app/`（探测、目录与启动路由），`@deepseek-ai/dsh-client-ui-open-in-app` 位于 `packages/client/ui-open-in-app/`（分体按钮），由 `dsh-web-app` bundle 的 `open-in-app` 与 `ui-open-in-app` 两行挂载进 Web profile。转正是重写，不是 vendoring：
+
+- **一对 host/client 包，沿用 `directory-picker-browse`/`ui-directory-picker-browse` 的配对结构**：host 包的 `src/index.ts` 在 `ctx.webServer` 上注册三条 HTTP 路由（`GET /open-in-app/apps`、`GET /open-in-app/icon/<id>`、`POST /open-in-app/open`）；ui 包的 `src/client/index.ts` 经标准 slot/inject 通货把分体按钮注册进 `conversation.session.header.utilities`，文案在类型化的 `open-in-app` locale 命名空间中，样式为 `--dsw-*` token 上的 CSS Modules（原插件手工注入的 style 标签与内联下拉被 `Menu` 原语替代），节点半边是让插件出现在主机名册上的空 apply。路由路径与 wire 载荷类型只有一个家：host 包浏览器安全的 `./shared` 子路径（只有常量与类型）；client bundle 经 client tsdown preset 的 `INLINE_SAFE` 条目将其内联，与 `dsh-session` 各 wire 切片同一通道。host 根入口只导出 Loader 所需的插件实体与类型；目录、resolver、launcher 与图标 helper 保持源码内部可见。
+- **一趟解析产出已验证的启动器；点击绝不重新检测。** 主机把整个目录每进程惰性解析一次，产出目录 id 到 `OpenInAppResolvedLaunch` 的映射——本机实际持有的启动器，绝不是裸的安装记录。`GET /apps` 提供映射的 keys，`POST /open` 直接启动其值；启动时发现可执行文件已消失（spawn `ENOENT`）会只作废该条目、重解析一次，无法再证明时把它从列表移除（卸载自愈，新安装等重启）。
+- **解析来源按平台务实且廉价。** macOS 在已知应用目录（`/Applications`、`~/Applications`）查条目的 bundle 拼写，启动 `open -a <解析出的 bundle>`；Xcode 跟随 `xcode-select -p`。Windows 依次读 `App Paths`、只在能证明磁盘可执行文件时采用的 Uninstall 记录、已知路径，以及采用版本化安装目录的应用中最新的目录——每趟每根一条批量 `reg.exe query`。GitHub Desktop 会同时解析版本化可执行文件与随包提供的 `cli.js`，不经命令 shell 调用受支持的 `github open <path>` 行为。CLI 名称经 `ctx.subprocess.resolveExecutable()` 进程内解析（PATH/PATHEXT stat，无 shell、无 `which`/`where.exe` 子进程）；CLI 不在 PATH 上的 Linux GUI 条目回退到其 XDG desktop 条目验证过的 `TryExec`/`Exec`，且只有主机声明了 display server 时才提供 `xdg-open`。其余主机命令（`xcode-select`、`reg.exe`、图标提取）经 `@deepseek-ai/dsh-native-command`（argv，绝不走 shell）执行。
+- **三个独立期限取代单一 `commandTimeoutMs`**：`probeTimeoutMs`（解析命令）、`iconTimeoutMs`（图标提取命令）、`launchWatchMs`（早期失败看护窗口），调整一种操作的超时不再改变其他操作的响应时间——随发行 bundle 保守地保留 10 秒命令期限（超时是失败上界而非延迟预算），看护窗口 1 秒，它才是约束 open 路由与按钮忙碌态挂起一次成功启动时长的量。启动以清理过凭据的环境（`dsh-subprocess` 的 `scrubbedParentEnv`）叠加适配器显式环境后 detached 派生；Windows GUI launcher 默认保持可见，只有负责另行打开 GUI 的 CLI 适配器会显式隐藏自己的进程。看护窗口关闭时仍在运行的启动器计为已启动，绝不会被杀死或等待（kitty 与 JetBrains IDE 在整个窗口生命周期内保持前台）。
+- **有主机来源的平台都提取图标。** macOS 把解析出的 bundle 的 `.icns` 转 128px PNG（`plutil` + `sips`）；Windows 用生成的 PowerShell 脚本以位置式 `-File` 参数（路径不经命令行解析）提取解析出的可执行文件的关联图标为 32px PNG；Linux 沿 spec 的 desktop 条目 `Icon=` 查 hicolor 主题与 pixmaps 目录（PNG 或 SVG，纯文件系统）。任何失败应答 404，浏览器保持通用占位图形。
+- **启动目录是按平台声明条目的数据表**（`OPEN_IN_APP_CATALOG`）：每个应用 id 按平台（`darwin`/`win32`/`linux`）声明一条按序尝试的启动器来源链——`fixed`（随系统内置）、`app`（已知目录的 bundle 拼写）、`xcode`（`xcode-select -p`）、`cli`（进程内 PATH 解析）、`file`（`${VAR}`/`~/` 展开后第一个存在的候选文件）、`scan`（带版本号安装目录取最新）、`app-paths` 与 `install-record`（Windows 注册表两层）、`github-desktop`（版本化可执行文件与随包 CLI）、`desktop`（Linux XDG desktop 条目）——加启动器 argv 模板（`{path}` token 原位携带目录，否则目录追加在末尾）、可选环境与 Windows 可见性策略，以及可选回退启动器（`xed` 之后的 `open -a <bundle>`）。白名单对齐 Codex 的 "Open In" 目标列表：编辑器与 IDE（VS Code、VS Code Insiders、Cursor、Windsurf、Zed、Sublime Text、Xcode、Android Studio、七个 JetBrains IDE）、转正插件原有的 Git GUI、终端（Terminal、iTerm2、Ghostty、Warp、kitty、Windows Terminal、Git Bash、GNOME Terminal、Konsole）与各平台文件管理器。文件管理器与平台终端使用独立 id（`finder`/`explorer`/`filemanager`），而非一个 id 配平台标签，因为标签是静态浏览器词典条目，且每台主机只会探测到其中一个。文件管理器条目直接经 `dsh-native-command` 的路径打开器启动（`shell-open`，OS shell 的 open verb，携带完整父环境，而非 detached 的清理环境 spawn），因为直接 spawn `explorer.exe <dir>` 不能可靠地弹出窗口。封闭 union 以 `assertNever` 收尾。
+- **所有路由都运行在组合 connection 服务的信任栅栏之后**（`requestRejection`：挫败 DNS rebinding 与跨站调用的 Host/Origin 栅栏，加上浏览器认证），与 API gateway 施加在其 WebSocket upgrade 上的守卫相同；机制的唯一出处是 `src/index.ts` 的模块注释。在该栅栏之上，open 路由在 wire 边界校验请求体：`application/json` 媒体类型（精确 essence，而非子串匹配）、以排空后 413 的方式把 body 限制在 64 KiB、校验字段类型、只接受探测为可用的目录 id，并要求指向现存目录的绝对路径。
+- **删除了 DSH 版本门禁。** 它存在是因为插件逐版本骑乘一个它不拥有的接口；第一方包与仓库同版本发布，门禁、它的 `sessionStorage`/`localStorage` 信任台账和 argv 遍历版本探测都失去了所指。
+- **上次选择经 `createSnapshotStore(..., { persist })` 持久化**（`dsh.open-in-app.choice`），替代手写 `localStorage` 访问。新存储没有平台特定的初始选择；用户首次选择前，组件使用主机提供的第一个可用条目。
+
+这对包放在 `packages/host/` 与 `packages/client/`，因为两个半边本来就是这两种东西：探测/启动侧是主机基础设施，与它消费的 webserver 同组；按钮是客户端表面，与其他 `ui-*` 包同组。评审把它从 `packages/workspace/` 的单个双半边包迁到这里（见替代方案）。
+
+## 考虑过的替代方案
+
+**将插件的 `lib/` 原样 vendor 进 `packages/`。** 最快，但手写 JavaScript 会整体不过 typecheck、覆盖率、i18n、JSDoc 和 invariant 门禁；为其保留豁免会造出仓库刻意不设的包类别。
+
+**用 Typert Remote 而非裸 webServer 路由。** apps/open 调用符合 Remote RPC 形态，但 icon 路由提供二进制 PNG，JSON RPC 词汇承载不了；把 icon 拆去裸路由而 apps/open 走 Remote 会让一个功能有两种传输。裸路由也匹配原插件的客户端，且 `webhook-github` 已确立带校验裸路由的先例。
+
+**扩展 `host/apiproxy` 的 `openPath` 而非新 open 端点。** `openPath` 用系统默认应用打开一个路径；本功能的主体是*用哪个*应用，带可用性探测和逐应用启动器——是不同的契约。两者共享 `dsh-native-command`。
+
+**放在 `packages/workspace/` 的单个双半边包（最初交付的形态，沿用 `dsh-session-log-export`）。** 评审期拆成 host/client 对：workspace 组的契约是 host-side only，该功能消费的是 `webServer` 而非 `workspaceRegistry`，且单包整体注册在 Client 编译聚合面迫使主机路由与目录测试伪装成 `.client.spec.ts`。拆分让每个半边的测试落在自己的编译面、依赖落在正确的区段；wire 契约经 host 包的 `./shared` 子路径保持唯一出处。
+
+**可配置目录（cordis.yml 定义应用）。** 延后：每个条目耦合发现、启动参数、进程策略与图标行为，因此接受任意命令前需要明确的用户设置归属与校验规则。用户提供的 label 属于用户数据，不与 locale 拥有的产品文案冲突。Codex 与 Orca 展示了可能的扩展形态：维护过的内置 preset 加可配置 custom handler。
+
+**通过操作系统 API 枚举所有已安装应用。** 不作为菜单真源：Launch Services、Windows 注册信息与 XDG desktop 条目可以定位应用，但不能证明每个应用都能接收 workspace 目录，也不能给出正确打开目录所需的启动协议。OS 原生标识仍可作为维护过的 preset 的 locator 输入；custom handler 用于覆盖长尾，而不是猜测启动语义。
+
+**为 Windows/Linux 条目内置静态图标（Codex 为每个目标打包 PNG）。** 拒绝：图标路由在三个平台都提供主机真实图标，为装饰性收益打包第三方产品图形会引入资产管线与商标风险面。
+
+**重子进程检测（最初交付的形态）：macOS 每条目一次 `open -Ra`、CLI 各一次 `which`/`where.exe`、每次启动重新解析。** 评审期替换：一次列表解析在 macOS 上派生约 26 个子进程，按显示名查 Launch Services 的证据弱于磁盘上的 bundle，仓库已有进程内 PATH 解析（`ctx.subprocess.resolveExecutable()`），且点击时重新检测把探测期限放上了交互路径。为挪位 macOS bundle 考虑过批量 `mdfind` 兜底，依评审的 known-paths 指示未采用；漏检记入已知限制。原生 LaunchServices/NSWorkspace 查询需要仓库尚无的 addon——延后，以已知 `.app` 路径为替身。
+
+**公共的应用发现 Service Definition。** 按单消费者规则延后：在出现第二个 GUI 发现消费者之前，resolver 保持为包内模块。
+
+**经 `SHDefExtractIcon` P/Invoke（`Add-Type`）脚本取 128px Windows 图标。** 延后：`ExtractAssociatedIcon` 是不需编译片段的 .NET 标准面，32px 在按钮 15-18 CSS px 的尺寸上仅在高分屏略微发软；若实际在意，P/Invoke 变体是脚本内局部升级。同理未追用户的 Linux 图标主题——hicolor 是所有主题继承的 freedesktop 兜底——自定义主题桌面会看到原版图标。
+
+## 后果
+
+- 只要主机在 macOS、Windows 或 Linux 上探测到至少一个已安装的目录应用，Web profile 就会出现头部按钮；其余情况零渲染（探测目录为空 → 组件返回 null）。
+- 社区插件的安装路径仍然有效但已冗余；其原始路由与浏览器选择键独立于 `open-in-app`，因此使用第一方功能的安装应移除社区插件，避免出现重复的头部控件。
+- 解析与图标每主机进程惰性执行一次，dsh 运行期间安装的应用要重启后才出现——接受；卸载方向经 `ENOENT` 单条目刷新自愈。
+- 目录在编译期固定；扩展它意味着同时编辑 `OPEN_IN_APP_CATALOG` 与两份 locale 词典（README 已知限制）。平台覆盖不均——若干 Git GUI 与终端仅有 macOS 条目；Windows 图标受限于 .NET 标准接口的 32px 提取，Linux 跟随 hicolor 而非当前主题，没有 desktop 记录的纯 CLI 条目则保留通用图标。
+- 覆盖：resolver 逻辑（每种 locator 在临时文件系统上、注册表转储与 desktop 条目 fixture、注入的 env/home/PATH 表）、逐平台图标提取、三条路由（真实 Loader + 真实 WebServer 组合，含单趟缓存、`ENOENT` 刷新与 HMR 安全处置）、controller wire 行为和组件呈现都以逐文件 100% 门禁做了单元测试；不新增 snapshot，因为随仓库发布的免密 snapshot fixture 断言会话驱动的输出，而这个纯浏览器侧控件不触及它。Web ARIA golden 禁用 `open-in-app` 与 `ui-open-in-app` 两行，Host-only 的 preset e2e 组合禁用 host 行：按钮反映运行机器实际安装了哪些应用，其出现与否和标签都是主机事实，跨平台 golden 无法钉住。

.agents/notes/implemented/feature/2026-08-27-continuable-subagent-human-inbox-control.i18n.yaml

@@ -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/2026-08-27-continuable-subagent-human-inbox-control.md
+2026-08-27-continuable-subagent-human-inbox-control.md: cf5dfd070dfd600fb64bc529b4a1476a258c1181
+2026-08-27-continuable-subagent-human-inbox-control.zh.md: 081fb84f75afcc94339ae3bf627480081bc56d89

.agents/notes/implemented/feature/2026-08-27-continuable-subagent-human-inbox-control.md

@@ -0,0 +1,51 @@
+# Agent Note: Human inbox controls for continuable subagents
+
+Status: implemented
+
+English | [中文](2026-08-27-continuable-subagent-human-inbox-control.zh.md)
+
+## Problem
+
+Continuable children use the same Agent loop and inbox as ordinary Agents, but the human delivery path exposed only FIFO follow-up. The Client discarded its existing Queue/Steer choice when it selected the dedicated subagent prompt Remote, and the generic Session ownership fence rejected every queue mutation for a subagent-owned identity. The browser therefore hid controls that the live child inbox already supported.
+
+Opening generic Session control indiscriminately would weaken the subagent ownership rule. Prompt delivery still needs exact live direct-parent authorization and cold-resume accounting, while queue mutation must reject one-shot, unknown, corrupt, and cold children. A valid continuable descriptor in the child's own log suffix identifies which live subagent-owned Sessions may use occurrence mutation. Settlement must also retain an idle Agent while a delivery a driver will claim is still pending, and must not tear down an Agent that claimed the idle phase for a maintenance task after `whenIdle()` resolved.
+
+## Decision
+
+A live continuable child exposes the ordinary human inbox controls without adding another queue, Remote endpoint, queue action, or Host-facing subagent operation. One-shot children remain read-only.
+
+The existing `SubagentPromptRequest` carries `delivery: 'queue' | 'steer'`. The Client forwards the mode already selected by `Session.prompt(content, mode)` through `subagent.prompt`. The Remote still requires the exact live direct parent and then uses one package-internal continuation-manager delivery operation. Queue calls `Agent.followup(message)`; steer calls `Agent.steer(message)`. Both paths share child locking, cold resume, final parent reauthorization, caller-signal cutoff, `MessageId` creation, rollback, and disposal-race handling. This human choice adds no public scheduling method or model tool; the separately owned `sendMessage()` and model-facing `send_message` operation keep their fixed adjacent-Agent Steer semantics.
+
+The browser gives a continuable child the ordinary busy Enter/Cmd+Enter Queue/Steer preference, QueueDock Edit/Remove/Steer actions, and empty-draft steer-all gesture. Send and Stop remain independent controls. Composer prompts still require the live parent because their Remote mints new admitted work. QueueDock mutations address already-live inbox work directly, so they remain available when the parent is offline; the parent-offline composer stays locked.
+
+The existing `session.updateQueue(itemId, action)` resolves the exact live Agent and admits a subagent-owned Session only when its current projected identity is continuable and the descriptor sequence belongs to the child's own non-seed suffix. A live one-shot Agent and a missing, inherited-only, or invalid identity retain the ownership failure. An absent Agent returns `queue-item-not-found` and does not cold-resume the child. The target Session id is sufficient human authority for a live inbox occurrence mutation; a parent address is not required. Edit and Remove retain their complete existing `nextTurn` and `nextStep` semantics, including plugin-injected context, while Steer requires a queued occurrence and an Agent that reports running when the command begins.
+
+The continuation manager keeps no second message-reservation state. One private `SubagentInbox` delegates Queue and Steer to the Agent inbox and owns the Activation's existing closing promise. Natural settlement waits for `Agent.whenIdle()`, an empty child Inbox, and disposal of every owned child. The manager confirms the Inbox, owned-child set, and wake generation under the child lock, then flushes final Session state while admission remains open. The final child-lock decision revalidates the Session sequence and the same residency facts, then synchronously starts an `Agent.runMaintenance()` task whose entry claims the idle phase and closes the wrapper in the same JavaScript turn. Every pending Inbox occurrence retains the Activation regardless of its delivery mode or provenance. Manager-owned deliveries, Inbox claims or discards, and owned-child release renew the wake generation. Direct Agent work accepted during the flush either changes the final Session or residency observation, remains active and prevents the final maintenance task from starting, or completes before revalidation.
+
+QueueDock Steer uses the Agent's best-effort delivery after the command admits a running queued occurrence. If the queued occurrence was claimed first, `queue-item-not-found` leaves its ordinary Queue delivery underway. If active cancellation wins during the synchronous transfer, Agent steering appends the message to `nextTurn`, latches a wake, and the Session command still succeeds. The selected message moves behind the remaining Queue in that fallback case. Newly composed Steer uses the same fallback and remains deliverable when it misses the nearest step.
+
+This decision partially supersedes the human-control exclusions in [Web subagent catalog and human continuation](2026-07-27-web-subagent-conversations.md), [Continuable subagents](2026-07-28-continuable-subagent-conversations.md), [Steer a queued Web message](../../archived/feature/2026-07-30-web-queue-steer-action.md), and [Steer the whole Web queue with an empty-draft Cmd/Ctrl+Enter](../../archived/feature/2026-08-06-web-queue-steer-all-gesture.md). The active records own catalog authorization and Activation lifecycle; the archived records preserve the original QueueDock Steer and gesture decisions.
+
+## Alternatives considered
+
+**Add `SubagentRuntime.steer()` and a new Remote.** Rejected because human prompt delivery already has a mode-bearing Client method and one authenticated Remote. A new public operation would expand both the service and model-adjacent surface without adding an execution primitive.
+
+**Add `subagents.updateQueue`.** Rejected because `session.updateQueue` already owns exact inbox occurrence mutation and its race failures. The projected continuable identity provides the narrow ownership-fence exception without adding another operation.
+
+**Route every subagent control through generic Session APIs.** Rejected because prompt and cancellation require subagent lineage authorization, cold-resume accounting, and dedicated failure mapping. Only live inbox occurrence mutation has enough target-local state to use the narrow ownership-fence exception.
+
+**Restrict continuable queue mutation to `nextTurn`.** Rejected because human inbox parity intentionally includes editing or removing pending steering and injected context. If a plugin needs a stronger transaction around its `nextStep` input, that protection belongs to the shared Agent inbox semantics rather than a subagent-only restriction.
+
+**Track waking work by `MessageId` and transfer that record across mutation.** Rejected because it duplicates the Inbox's pending set with a second activity ledger and couples residency to occurrence identity. `whenIdle()` waits for existing Agent activity, `Inbox.hasPending` conservatively retains every occurrence, the Activation generation invalidates stale observations, and the final maintenance task atomically joins idle ownership to admission closure. This choice can retain quiet injected context, but it avoids both an additional mutation protocol and silent loss of accepted steering.
+
+**Derive residency from `MessageSource.kind`, treating `plugin` as parked context.** Rejected because `kind` records who produced a message, not how it was delivered, and `MessageSourceMap` is merge-extensible. Plugins steer with a plugin source (`cordis-host-runner` failure reports, blocking Stop hooks) and hosts inject with non-plugin sources (`dsh-experimental-agent-team` quiet mail), so the correspondence fails in both directions. Treating all pending occurrences alike avoids that unsupported inference.
+
+## Consequences
+
+Continuable child conversations and ordinary Sessions share one human inbox interaction model and one Agent-loop queue. Human steering can affect a resident or cold-resumed child without changing public model controls. QueueDock remains useful for a live child after its parent goes offline, while new messages continue to respect direct-parent authorization.
+
+The generic Session command has one narrow ownership-fence exception for a live subagent-owned Agent with a valid own-suffix continuable identity. Because the operation addresses either inbox destination, a caller that knows a pending `MessageId` can edit or remove plugin-supplied next-step input, exactly as on an ordinary Session. QueueDock renders only `queued`-placement rows, so no browser gesture reaches that input; an edit there also keeps the original producer's `MessageSource`, which would attribute human text to that producer.
+
+Inbox notifications retain their occurrence semantics and do not carry continuation residency. Claim and discard notifications only wake settlement after pending work changes; `whenIdle()`, the final idle-phase maintenance task, `Inbox.hasPending`, the owned-child set, the Activation generation, and the Session sequence decide whether disposal is safe without depending on scheduler ordering, message identity, or provenance. The final flush precedes the closing cutoff, so a detached hook, job completion, or direct Agent delivery accepted during that await invalidates the observation instead of being stopped by the resulting disposal. Maintenance that remains active prevents the final task from claiming the idle phase; maintenance that starts and finishes during the flush has completed before disposal. A child left holding only injected context remains resident even though no driver is obliged to claim it; without a later waking delivery, queue removal, or manager teardown, that child and its live ancestors can remain resident for the process lifetime. A replayed Inbox follows the same conservative rule without reconstructing how each pending message was delivered.
+
+Model-side scheduling remains fixed rather than caller-selectable. The adjacent-Agent `send_message` tool always uses Steer, while only the browser human path chooses Queue or Steer.

.agents/notes/implemented/feature/2026-08-27-continuable-subagent-human-inbox-control.zh.md

@@ -0,0 +1,51 @@
+# Agent Note: 可继续 subagent 的人类 inbox 控制
+
+Status: implemented
+
+[English](2026-08-27-continuable-subagent-human-inbox-control.md) | 中文
+
+## 问题
+
+可继续子级与普通 Agent 使用相同的 agent loop（智能体循环）和 inbox，但人类投递路径只公开 FIFO 后续轮次。Client 选择专用 subagent prompt Remote 时会丢弃既有的 Queue／Steer 选择，通用 Session ownership fence 又拒绝 subagent 所有身份的全部 queue 变更。因此，浏览器隐藏了在线子级 inbox 已经支持的控制。
+
+无差别开放通用 Session 控制会削弱 subagent 所有权规则。Prompt 投递仍需要确切在线直接父级鉴权与冷恢复记账，而 queue 变更必须拒绝一次性、未知、损坏和冷子级。child 自身 log suffix 中的有效 continuable descriptor 可标识哪些在线 subagent-owned Session 能使用 occurrence mutation。Settlement 还必须在 idle Agent 仍有会被 driver 认领的待投递工作时保留该 Agent，也不得拆除在 `whenIdle()` 兑现后才占用 idle 阶段执行 maintenance 任务的 Agent。
+
+## 决策
+
+在线可继续子级公开普通的人类 inbox 控制，不增加另一套 queue、Remote endpoint、queue action 或面向 Host 的 subagent 操作。一次性子级继续只读。
+
+现有 `SubagentPromptRequest` 携带 `delivery: 'queue' | 'steer'`。Client 把 `Session.prompt(content, mode)` 已选出的 mode 经 `subagent.prompt` 原样转发。Remote 仍要求确切在线直接父级，随后使用一个包内 continuation manager 投递操作。Queue 调用 `Agent.followup(message)`；steer 调用 `Agent.steer(message)`。两条路径共享 child lock、冷恢复、最终父级重新鉴权、调用方 signal 截止、`MessageId` 创建、回滚与 dispose 竞态处理。该人类选择不新增公开调度方法或模型工具；由其他决策拥有的 `sendMessage()` 与面向模型的 `send_message` 操作保留固定的相邻 Agent Steer 语义。
+
+浏览器为可继续子级提供普通的繁忙态 Enter／Cmd+Enter Queue／Steer 偏好、QueueDock Edit／Remove／Steer 操作，以及空草稿 steer-all 手势。Send 与 Stop 继续是独立控制。Composer prompt 会创建新的已准入工作，因此仍要求在线父级。QueueDock 变更直接寻址已经在线的 inbox 工作，所以父级离线时仍可使用；父级离线的 composer 继续锁定。
+
+现有 `session.updateQueue(itemId, action)` 会解析确切在线 Agent，并且只有 subagent-owned Session 的当前 projected identity 为 continuable、descriptor 序号属于 child 自身的非 seed suffix 时才会准入。在线 one-shot Agent 以及缺失、仅继承或无效的 identity 都会继续触发所有权失败。Agent 不存在时返回 `queue-item-not-found`，且不会冷恢复子级。对在线 inbox occurrence 变更而言，目标 Session id 已是充分的人类权限；无需 parent 地址。Edit 与 Remove 保留既有完整 `nextTurn` 和 `nextStep` 语义，包括插件注入的 context；Steer 要求排队 occurrence，且 command 开始时 Agent 必须报告 running。
+
+Continuation manager 不保留第二套消息 reservation 状态。一个私有 `SubagentInbox` 会把 Queue 与 Steer 委托给 Agent inbox，并持有 Activation 既有的 closing promise。自然结算会等待 `Agent.whenIdle()`、child Inbox 为空以及所拥有的每个子级完成 dispose。管理器会在 child lock 内确认 Inbox、owned-child set 与 wake generation，再在准入保持开放时 flush 最终 Session 状态。最终 child-lock 决策会重新验证 Session 序号与相同的驻留事实，然后同步启动一个 `Agent.runMaintenance()` 任务；该任务的入口会占用 idle 阶段，并在同一个 JavaScript turn 内关闭包装层。每个待处理 Inbox occurrence 都会保留 Activation，无论其投递模式或来源如何。由 manager 所有的投递、Inbox claim 或 discard，以及所拥有子级的释放都会更新 wake generation。flush 期间直接接受的 Agent 工作要么改变最终 Session 或驻留观察，要么保持活跃并阻止最终 maintenance 任务启动，要么在重验前完成。
+
+QueueDock Steer 在 command 准入一个正在运行的排队 occurrence 后，采用 Agent 的 best-effort 投递。如果排队 occurrence 先被 claim，`queue-item-not-found` 表示其普通 Queue 投递已经开始。如果活跃取消在同步转移期间先发生，Agent steering 会把消息追加到 `nextTurn`、锁存唤醒，Session command 仍然成功。在该 fallback 情况下，选中消息会移到 Queue 剩余项之后。新组合的 Steer 使用同样的 fallback，错过最近步骤时仍保证可投递。
+
+本决策部分取代 [Web subagent 目录与人类 continuation](2026-07-27-web-subagent-conversations.zh.md)、[可继续 subagent](2026-07-28-continuable-subagent-conversations.zh.md)、[Steer Web 已排队消息](../../archived/feature/2026-07-30-web-queue-steer-action.md)和[用空草稿 Cmd/Ctrl+Enter steer 整个 Web queue](../../archived/feature/2026-08-06-web-queue-steer-all-gesture.md)中的人类控制排除项。活跃记录拥有目录鉴权与 Activation 生命周期；归档记录保留最初的 QueueDock Steer 与手势决策。
+
+## 考虑过的替代方案
+
+**新增 `SubagentRuntime.steer()` 与 Remote。** 拒绝，因为人类 prompt 投递已经拥有带 mode 的 Client 方法和一个已鉴权 Remote。新的公开操作会扩大 service 与模型相邻接口，却不增加执行原语。
+
+**新增 `subagents.updateQueue`。** 拒绝，因为 `session.updateQueue` 已经拥有准确 inbox occurrence 变更及其竞态失败。Projected continuable identity 提供狭窄的 ownership-fence 例外，无需新增操作。
+
+**把所有 subagent 控制都路由到通用 Session API。** 拒绝，因为 prompt 与取消需要 subagent 血缘鉴权、冷恢复记账与专用失败映射。只有在线 inbox occurrence 变更拥有足够的目标本地状态，可使用狭窄的 ownership-fence 例外。
+
+**把可继续 queue 变更限制在 `nextTurn`。** 拒绝，因为人类 inbox 对齐有意包括编辑或删除待处理 steering 与注入 context。如果插件需要围绕其 `nextStep` 输入建立更强事务，该保护应属于共享 Agent inbox 语义，而非 subagent 专属限制。
+
+**按 `MessageId` 跟踪唤醒工作，并在 mutation 中转移该记录。** 拒绝，因为这会用第二套活动账本重复 Inbox 的待处理集合，并让驻留依赖 occurrence 身份。`whenIdle()` 会等待既有 Agent 活动，`Inbox.hasPending` 保守地保留每个 occurrence，Activation generation 会让过期观察失效，而最终 maintenance 任务则以原子方式衔接 idle ownership 与准入关闭。这项选择可能保留静默注入的 context，但既避免额外的 mutation 协议，也避免静默丢失已接受的 steering。
+
+**用 `MessageSource.kind` 推导驻留，把 `plugin` 视为停放 context。** 拒绝，因为 `kind` 记录的是消息由谁产生，而非如何投递，且 `MessageSourceMap` 可合并扩展。插件会以 plugin 来源 steer（`cordis-host-runner` 的失败报告、阻断式 Stop hook），host 也会以非 plugin 来源 inject（`dsh-experimental-agent-team` 的静默邮件），因此该对应关系在两个方向上都不成立。统一对待所有待处理 occurrence 可以避免这种没有依据的推断。
+
+## 结果
+
+可继续子级会话与普通 Session 共享一套人类 inbox 交互模型和一套 Agent-loop queue。人类 steering 可以影响驻留或冷恢复的子级，而不改变公开模型控制。父级离线后，QueueDock 对在线子级仍有用；新消息则继续遵守直接父级鉴权。
+
+通用 Session command 为拥有有效自身 suffix continuable identity 的在线 subagent-owned Agent 提供一个狭窄的 ownership-fence 例外。因为该操作可寻址两个 inbox 目标，知道待处理 `MessageId` 的调用方可以像操作普通 Session 一样，编辑或删除插件提供的 next-step 输入。QueueDock 只渲染 `queued` placement 的行，因此没有浏览器手势能到达该输入；在那里编辑还会保留原产出方的 `MessageSource`，从而把人类文本归属给该产出方。
+
+Inbox notification 保留 occurrence 语义，不携带 continuation 驻留状态。Claim 与 discard notification 只负责在待处理工作变化后唤醒 settlement；`whenIdle()`、最终 idle 阶段 maintenance 任务、`Inbox.hasPending`、owned-child set、Activation generation 与 Session 序号无需依赖调度顺序、消息身份或来源即可决定何时安全 dispose。最终 flush 位于 closing cutoff 之前，因此 detached hook、job completion 或直接 Agent 投递只要在该 await 期间被接受，就会让观察失效，而不会被随后发生的 dispose 停止。仍然活跃的 maintenance 会阻止最终任务占用 idle 阶段；在 flush 期间开始并结束的 maintenance 已在 dispose 前完成。仅持有被注入 context 的 child 即使没有 driver 必须认领它，也会保持驻留；如果之后没有唤醒投递、queue removal 或 manager teardown，该 child 及其在线祖先可以在进程生命周期内一直驻留。重放出的 Inbox 遵循同一条保守规则，无需重建每条待处理消息的投递方式。
+
+模型侧调度保持固定，不由调用方选择。相邻 Agent 的 `send_message` 工具始终使用 Steer，只有浏览器人类路径选择 Queue 或 Steer。

.agents/notes/implemented/feature/2026-09-07-model-switch-notice.i18n.yaml

@@ -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/2026-09-07-model-switch-notice.md
+2026-09-07-model-switch-notice.md: cb2f915f13b5fa277977b7a768a3531074126f2a
+2026-09-07-model-switch-notice.zh.md: b6dd8da02732e4b6c6b2102aef6074b8f136d37d

.agents/notes/implemented/feature/2026-09-07-model-switch-notice.md

@@ -0,0 +1,29 @@
+# Agent Note: Model-visible route-change notices
+
+Status: implemented
+
+English | [中文](2026-09-07-model-switch-notice.zh.md)
+
+## Problem
+
+Session history identifies message roles but does not tell a newly selected model which route generated earlier assistant turns. In the motivating session, the user switched from `deepseek-v4-flash` to `deepseek-v4-flash-vision-exp`. The new model saw image placeholders saying that a text-only model had omitted the images and inferred that the limitation described its own image capability.
+
+## Decision
+
+`installModelSelection` compares the provider/model selection captured during prompt assembly with the latest durable request header. When they differ, it appends `[model changed: assistant turns above this point were generated by <previous>; the session continues with <next>]` as an identified user-role message to a downstream `agent/pre-step` decision that would send a model request.
+
+Routes from the same provider use model ids; cross-provider routes use `provider/model`. Initial selection, unchanged routes, reasoning-effort-only changes, rejected steps, and aborted steps add no notice. An empty first decision and a decision that removes offered messages remain no-request results. An empty continuation after a tool call receives the notice because the loop would still request the model from retained history. A selection changed during pre-step processing waits for the next prompt assembly.
+
+The accepted message is logged through the existing `user/message` event before the request header and therefore appears in model input, Chat, and Trajectory. The request header remains the durable record that the new route was used. If a step fails before that header is logged, the next request step repeats the notice because the durable previous route remains unchanged.
+
+## Alternatives considered
+
+**Put the notice in the system prompt.** A transient prompt change would not be reconstructable from the session log and would not mark the exact point where ownership of assistant turns changed.
+
+**Create a separate package or capability.** The behavior only coordinates the prompt snapshot and request route already owned by `installModelSelection`; it has no independent service, provider, or consumer roles.
+
+**Show the change only in the client.** Client-only presentation would leave the newly selected model without the fact needed to interpret earlier assistant turns.
+
+## Consequences
+
+Each emitted route-change notice becomes retained user-role history and consumes context tokens on later requests. A failure before request-header persistence can retain more than one identical notice. Existing session events represent the behavior, so the session format does not change.

.agents/notes/implemented/feature/2026-09-07-model-switch-notice.zh.md

@@ -0,0 +1,29 @@
+# Agent Note: 模型可见的路由切换提示
+
+Status: implemented
+
+[English](2026-09-07-model-switch-notice.md) | 中文
+
+## 问题
+
+会话历史会标明消息角色，但不会告诉新选中的模型，先前的 assistant 消息由哪个路由生成。在引出本改动的会话中，用户从 `deepseek-v4-flash` 切换到 `deepseek-v4-flash-vision-exp`。新模型看到图片占位文本称纯文本模型省略了图片，于是误以为该限制描述的是自己的图片能力。
+
+## 决策
+
+`installModelSelection` 会比较提示词组装时捕获的提供方和模型选择与最新的持久请求 header。两者不同时，它会把 `[model changed: assistant turns above this point were generated by <previous>; the session continues with <next>]` 作为带标识的 user 角色消息，追加到下游原本会发出模型请求的 `agent/pre-step` 决策中。
+
+同一提供方的路由只使用模型 ID；跨提供方的路由使用 `provider/model`。首次选择、未变化的路由、仅推理强度变化、被拒绝的步骤和被取消的步骤都不会增加提示。首次空决策与移除待处理消息后得到的空决策都不会产生请求。工具调用后的空续步仍会基于保留历史请求模型，因此会收到提示。pre-step 处理期间发生的选择变更会等待下一次提示词组装。
+
+被接纳的消息会在请求 header 前通过现有 `user/message` 事件落盘，因此会出现在模型输入、Chat 和 Trajectory 中。请求 header 仍是新路由已被使用的持久记录。如果步骤在该 header 落盘前失败，持久记录中的先前路由保持不变，所以下一个请求步骤会再次发出提示。
+
+## 考虑过的替代方案
+
+**把提示放进系统提示词。** 临时提示词变更无法从会话日志重建，也不能标出 assistant 消息由哪个模型生成的分界点。
+
+**新建单独的包或能力。** 该行为只协调 `installModelSelection` 已经负责的提示词快照与请求路由，没有独立的服务、提供方或消费方角色。
+
+**只在客户端展示切换。** 仅由客户端展示时，新选中的模型仍然缺少解释先前 assistant 消息所需的信息。
+
+## 影响
+
+每条实际发出的路由切换提示都会成为保留的 user 角色历史，并在后续请求中占用上下文 token。请求 header 落盘前发生的失败可能保留多条相同提示。该行为使用现有会话事件，因此会话格式不变。

.agents/notes/implemented/process/2026-07-21-serial-cross-platform-ci-reference.i18n.yaml

@@ -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/process/2026-07-21-serial-cross-platform-ci-reference.md
-2026-07-21-serial-cross-platform-ci-reference.md: e7d1caefabe3e90a84ea8eaea67381ea4f63f6f7
-2026-07-21-serial-cross-platform-ci-reference.zh.md: 57279e662aaf0c361bc2641cc5c6e757c4199ebb
+2026-07-21-serial-cross-platform-ci-reference.md: edb81b643d0cef2e5bc807005a9016324b8430ab
+2026-07-21-serial-cross-platform-ci-reference.zh.md: 41fd9c032038f2a312978acf995febfdab34aeaa

.agents/notes/implemented/process/2026-07-21-serial-cross-platform-ci-reference.md

@@ -16,7 +16,7 @@ Real-kernel sandbox proofs require specific hosted operating systems and archite
 
 ## Decision
 
-[CI](../../../../.github/workflows/ci.yml) (pull-request-only) and [CI master](../../../../.github/workflows/ci-master.yml) (master-push + workflow_dispatch) give pull-request and master-push events complementary responsibilities. Pull requests run consolidated Linux and Wine-hosted Windows jobs plus the Node compatibility and Python contracts on standard GitHub-hosted capacity; an independent native Windows job reports the complete Windows inventory without participating in the required aggregate. On a push to `master`, the active references are `serial / linux (self-hosted standby)` on the in-house `vm-backup` pool and `serial / windows (self-hosted standby)` on the in-house `dsh-win-ci` pool — the hot-standby drills that continuously re-prove the failover targets described in the [failover runbook](2026-07-26-ci-failover-runbook.md). There is no standard-hosted `serial / linux` definition; the standard-hosted `serial / macos` remains disabled under `TODO(hosted-serial-ci)` until its portable capacity can be restored. The current `serial / windows` definition is the in-house `dsh-win-ci` standby. The separate job definitions intentionally keep their short checkout, runtime setup, and immutable install sequences visible instead of hiding operating systems behind a matrix or reusable workflow. `workflow_dispatch` is reserved for runner benchmarks.
+[CI](../../../../.github/workflows/ci.yml) (pull-request-only) and [CI master](../../../../.github/workflows/ci-master.yml) (master-push + workflow_dispatch) give pull-request and master-push events complementary responsibilities. Pull requests run Linux, native Windows, [Node compatibility with self-hosted routing under the Linux variable and trust conditions](2026-09-06-node-compatibility-selfhosted.md), and Python checks; [platform scheduling](2026-09-06-master-only-platform-ci.md) assigns Wine and three Python runtime carriers to master pushes. On a push to `master`, the active references are `serial / linux (self-hosted standby)` on the in-house `vm-backup` pool and `serial / windows (self-hosted standby)` on the in-house `dsh-win-ci` pool — the hot-standby drills that continuously re-prove the failover targets described in the [failover runbook](2026-07-26-ci-failover-runbook.md). There is no standard-hosted `serial / linux` definition; the standard-hosted `serial / macos` remains disabled under `TODO(hosted-serial-ci)` until its portable capacity can be restored. The current `serial / windows` definition is the in-house `dsh-win-ci` standby. The separate job definitions intentionally keep their short checkout, runtime setup, and immutable install sequences visible instead of hiding operating systems behind a matrix or reusable workflow. `workflow_dispatch` is reserved for runner benchmarks.
 
 Each reference job runs `pnpm run check:ci` without any shard selector. `DSH_GATE_CONCURRENCY=1` makes the top-level aggregate execute one ready gate at a time; coverage, snapshot replay, built-bin smoke, and publication validation also receive worker counts of one. The reference jobs may run beside one another, but each host's repository gates are serial and complete. Linux installs bubblewrap before replaying snapshots, and Windows enables Developer Mode before installing the symlinked workspace.
 
@@ -28,13 +28,13 @@ The standalone [Sandbox](../../../../.github/workflows/sandbox.yml) workflow bel
 
 Master reference jobs are diagnostic and do not participate in the pull request's required `all checks passed` result. The ci-master and Sandbox workflows keep their cross-platform references on master pushes. Performance is evaluated from completed hosted-job timestamps and reported as a measurement; it is not encoded as a `timeout-minutes` value.
 
-The active serial references run on the self-hosted `vm-backup` (`serial / linux`) and `dsh-win-ci` (`serial / windows`) pools; the one remaining disabled hosted serial reference (`serial-macos`) uses `macos-latest`, and there is no standard-hosted `serial / linux` label. The required pull-request Windows job runs under Wine on `ubuntu-latest`, while the independent pull-request native job uses the hosted `dsh-windows-2025-16core` runner under normal operation and the self-hosted `[self-hosted, dsh-win-ci, windows]` pool under failover (see the [failover runbook](2026-07-26-ci-failover-runbook.md)), and is absent from the required aggregate under the [dual Windows decision](2026-08-08-native-windows-pull-request-ci.md). Required pull-request jobs use portable standard capacity under the [required-CI decision](../../archived/process/2026-07-23-portable-required-pull-request-ci.md). Higher-core hosted runners remain manual benchmarks because a correctness path must remain runnable without repository-external runner configuration.
+The active serial references run on the self-hosted `vm-backup` (`serial / linux`) and `dsh-win-ci` (`serial / windows`) pools; the one remaining disabled hosted serial reference (`serial-macos`) uses `macos-latest`, and there is no standard-hosted `serial / linux` label. The master-only Wine job runs on `ubuntu-latest`, while the pull-request native jobs use the hosted `dsh-windows-2025-16core` runner under normal operation and the self-hosted `[self-hosted, dsh-win-ci, windows]` pool under failover (see the [failover runbook](2026-07-26-ci-failover-runbook.md)), with build and targeted process checks required under the [native Windows decision](2026-08-08-native-windows-pull-request-ci.md). Required pull-request jobs use portable standard capacity under the [required-CI decision](../../archived/process/2026-07-23-portable-required-pull-request-ci.md). Higher-core hosted runners remain manual benchmarks because a correctness path must remain runnable without repository-external runner configuration.
 
 ## Alternatives considered
 
 - **Set each timeout equal to its latency target** - rejected because scheduling variance would cancel correct work and suppress the evidence needed to diagnose a regression.
 - **Trust only the concurrent primary inventory** - rejected because scheduling and validation share implementation assumptions; a serial aggregate is an independent completeness check.
-- **Run the serial references on every pull request** - rejected because they duplicate complete cross-platform aggregates and add macOS work to every change; the required jobs already execute the blocking Linux and Wine-hosted Windows contracts, and the independent native job supplies the complete Windows result.
+- **Run the serial references on every pull request** - rejected because they duplicate complete cross-platform aggregates and add macOS work to every change; the required jobs already execute the blocking Linux and native Windows checks, and the other native jobs supply the complete Windows result.
 - **Run the real-kernel Sandbox matrix on every pull request** - rejected because its four statuses do not participate in branch protection, while repeated installs, Landlock builds, and macOS unit parity consume runner capacity without changing the merge verdict. The master run retains the platform and installed-launcher signal.
 - **Use one operating-system matrix** - rejected because three named jobs make the reference surface visible without another selection mechanism.
 - **Run the serial reference on larger runners** - rejected because both required CI and its independent reference must remain runnable when organization-owned pools cannot allocate jobs.

.agents/notes/implemented/process/2026-07-21-serial-cross-platform-ci-reference.zh.md

@@ -16,7 +16,7 @@ Status: implemented
 
 ## 决策
 
-[CI](../../../../.github/workflows/ci.yml)（仅 pull request）与 [CI master](../../../../.github/workflows/ci-master.yml)（master 推送 + `workflow_dispatch`）为拉取请求事件与 master 推送事件赋予互补的职责。拉取请求在 GitHub 标准托管容量上运行合并后的 Linux 和由 Wine 承载的 Windows 作业，以及 Node 兼容性与 Python 约定；一个独立的原生 Windows 作业会报告完整的 Windows 清单，但不参与必需聚合流程。向 `master` 推送时，当前启用的参考作业是公司自有 `vm-backup` 池上的 `serial / linux (self-hosted standby)` 和 `dsh-win-ci` 池上的 `serial / windows (self-hosted standby)`——这些热备演练持续验证[故障切换手册](2026-07-26-ci-failover-runbook.zh.md)所描述的切换目标。不存在标准托管的 `serial / linux` 定义；标准托管的 `serial / macos` 仍处于禁用状态，并由 `TODO(hosted-serial-ci)` 标记，直到其可移植容量恢复。当前 `serial / windows` 定义是公司自有 `dsh-win-ci` 池的 standby。各自独立的作业定义有意显式保留简短的代码检出、运行时设置和依赖锁定的安装步骤，而不是用矩阵或可复用工作流隐藏操作系统差异。`workflow_dispatch` 仅用于运行器基准测试。
+[CI](../../../../.github/workflows/ci.yml)（仅 pull request）与 [CI master](../../../../.github/workflows/ci-master.yml)（master 推送 + `workflow_dispatch`）为拉取请求事件与 master 推送事件赋予互补的职责。拉取请求运行 Linux、原生 Windows、[仅在 Linux 变量与信任条件满足时使用自托管路由的 Node 兼容性](2026-09-06-node-compatibility-selfhosted.zh.md)与 Python 检查；[平台调度](2026-09-06-master-only-platform-ci.zh.md)将 Wine 与三个 Python 运行时载体分配给 master 推送。向 `master` 推送时，当前启用的参考作业是公司自有 `vm-backup` 池上的 `serial / linux (self-hosted standby)` 和 `dsh-win-ci` 池上的 `serial / windows (self-hosted standby)`——这些热备演练持续验证[故障切换手册](2026-07-26-ci-failover-runbook.zh.md)所描述的切换目标。不存在标准托管的 `serial / linux` 定义；标准托管的 `serial / macos` 仍处于禁用状态，并由 `TODO(hosted-serial-ci)` 标记，直到其可移植容量恢复。当前 `serial / windows` 定义是公司自有 `dsh-win-ci` 池的 standby。各自独立的作业定义有意显式保留简短的代码检出、运行时设置和依赖锁定的安装步骤，而不是用矩阵或可复用工作流隐藏操作系统差异。`workflow_dispatch` 仅用于运行器基准测试。
 
 每个参考作业均在不设置任何分片选择器的情况下运行 `pnpm run check:ci`。`DSH_GATE_CONCURRENCY=1` 使顶层聚合每次只执行一个已经就绪的门禁；覆盖率、快照回放、built-bin 冒烟测试和发布验证的 worker 数量也设为 1。各参考作业可以彼此并行，但每台主机上的仓库门禁都串行运行且完整执行。Linux 在回放快照前安装 bubblewrap，Windows 则在安装采用符号链接的工作区前启用开发人员模式。
 
@@ -28,13 +28,13 @@ macOS 参考流程使用 fork 进程运行常规 Vitest 项目。macOS arm64 上
 
 master 分支的参考作业仅用于诊断，不参与拉取请求所要求的 `all checks passed` 结果。ci-master 与 Sandbox 工作流把跨平台参考流程保留在 master 推送上。系统根据已完成托管作业的时间戳评估性能，并将其报告为测量结果，而不是写成 `timeout-minutes` 值。
 
-当前启用的参考流程运行在公司自有 `vm-backup`（`serial / linux`）与 `dsh-win-ci`（`serial / windows`）自托管池上；唯一剩余的禁用托管参考作业（`serial-macos`）使用 `macos-latest`，且不存在标准托管的 `serial / linux` 标签。拉取请求必需的 Windows 作业在 `ubuntu-latest` 上通过 Wine 运行，而独立的拉取请求原生作业在正常运行下使用托管的 `dsh-windows-2025-16core` 运行器，故障切换时使用自托管 `[self-hosted, dsh-win-ci, windows]` 池（参见[故障切换手册](2026-07-26-ci-failover-runbook.zh.md)），依据[双 Windows 决策](2026-08-08-native-windows-pull-request-ci.zh.md)不参与必需聚合流程。依据[必需 CI 决策](../../archived/process/2026-07-23-portable-required-pull-request-ci.md)，拉取请求必需作业使用可移植的标准容量。更高核心数的托管运行器仍仅用于手动基准测试，因为正确性路径必须无需仓库外部的运行器配置即可运行。
+当前启用的参考流程运行在公司自有 `vm-backup`（`serial / linux`）与 `dsh-win-ci`（`serial / windows`）自托管池上；唯一剩余的禁用托管参考作业（`serial-macos`）使用 `macos-latest`，且不存在标准托管的 `serial / linux` 标签。仅 master 触发的 Wine 作业在 `ubuntu-latest` 上运行，而拉取请求原生作业在正常运行下使用托管的 `dsh-windows-2025-16core` 运行器，故障切换时使用自托管 `[self-hosted, dsh-win-ci, windows]` 池（参见[故障切换手册](2026-07-26-ci-failover-runbook.zh.md)），依据[原生 Windows 决策](2026-08-08-native-windows-pull-request-ci.zh.md)，其中构建与定向进程检查参与必需聚合流程。依据[必需 CI 决策](../../archived/process/2026-07-23-portable-required-pull-request-ci.md)，拉取请求必需作业使用可移植的标准容量。更高核心数的托管运行器仍仅用于手动基准测试，因为正确性路径必须无需仓库外部的运行器配置即可运行。
 
 ## 曾考虑的替代方案
 
 - **将每个超时值设为相应延迟目标**：不予采纳，因为调度波动会中止原本正确的执行，并使诊断回归所需的证据无法产生。
 - **仅信任并发执行的主门禁清单**：不予采纳，因为调度逻辑与校验逻辑共享实现假设；串行聚合流程是一项独立的完整性检查。
-- **在每个拉取请求上运行串行参考作业**：不予采纳，因为这些作业会重复完整的跨平台聚合流程，并为每项改动增加 macOS 工作；必需作业已经执行阻塞性的 Linux 和由 Wine 承载的 Windows 约定，而独立原生作业提供完整的 Windows 结果。
+- **在每个拉取请求上运行串行参考作业**：不予采纳，因为这些作业会重复完整的跨平台聚合流程，并为每项改动增加 macOS 工作；必需作业已经执行阻塞性的 Linux 和原生 Windows 检查，而其它原生作业提供完整的 Windows 结果。
 - **在每个拉取请求上运行真实内核 Sandbox 矩阵**：不予采纳，因为它的四个状态不参与分支保护，而重复安装、Landlock 构建以及为保持平台一致而运行的 macOS 单元测试会消耗运行器容量，却不会改变合并裁决。master 上的运行保留平台与已安装 launcher 的信号。
 - **使用一个操作系统矩阵**：不予采纳，因为三个具名作业无需另一套选择机制，就能让参考流程的构成清晰可见。
 - **在大型运行器上运行串行参考流程**：不予采纳，因为当组织自有运行器池无法分配作业时，必需 CI 及其独立参考流程都必须仍可运行。

.agents/notes/implemented/process/2026-07-26-ci-failover-runbook.i18n.yaml

@@ -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/process/2026-07-26-ci-failover-runbook.md
-2026-07-26-ci-failover-runbook.md: b24996a4ba4dfaa4b26f88519a61f45c81efb5b5
-2026-07-26-ci-failover-runbook.zh.md: ee7339d70e4796c367f97490ea57468687464b3f
+2026-07-26-ci-failover-runbook.md: f24cb8b8239141cd1ccf468a566dba620dd3cfdd
+2026-07-26-ci-failover-runbook.zh.md: 57c4a92a3af720d9b11b7a1ce7a1515b83c77339