Merge remote-tracking branch 'origin/master' into worktree/sidebar-subagent-activity

# Conflicts:
#	packages/client/runtime/README.i18n.yaml

.agents/notes/implemented/feature/2026-06-18-compaction-capability-seam.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/2026-06-18-compaction-capability-seam.md
-2026-06-18-compaction-capability-seam.md: 8dcbe74429a620027a570124383442b969c12196
-2026-06-18-compaction-capability-seam.zh.md: 27b63f29c2e6f35637185b47c882ae42e5d41088
+2026-06-18-compaction-capability-seam.md: 390d091fd6e6f7fa1694ba78b90522b70948fe92
+2026-06-18-compaction-capability-seam.zh.md: d72793e6adde911a8623ac1c6b65df899381923e

.agents/notes/implemented/feature/2026-06-18-compaction-capability-seam.md

@@ -33,7 +33,7 @@ This is not a coupling smell — it is the contract's domain. The "only cordis"
 
 An earlier draft put the full algorithm (the retention walk, token-summing, text extraction) as concrete methods on the interface. That recouples the contract to one strategy: a backend that wants a different retention policy or event sequence would have to fight inherited concrete code. Making all three operations abstract puts every *how* decision in the backend and keeps the interface a statement of *what*. Token measurement is not a compaction hook at all; the singleton service lets multiple consumers share one per-session replay fold.
 
-`compactIfNeeded(agent, trigger, signal)` takes an explicit `'pressure' | 'context-overflow'` trigger and cancellation. It reads only the latest durable routed request; no header means no work, while any routed provider/model target uses the singleton estimator. `compactNow(agent, signal)` requires an idle agent and performs one useful balanced reduction even below pressure, returning `null` without writes when none exists. `compactRegion(start, end, agent, signal?)` uses `agent.session` as its single session identity and keeps an optional signal for explicit callers. The default summarizer resolves its target from explicit config, the latest logged routed target, then agent options, and records the provider/model pair after any `llm/stream` routing. It replays the routed request's prefix and appends the compaction directive as a trailing user message so the provider's warm KV cache is reused — see the [summary prefix-cache Agent Note](../bug-fix/2026-07-21-compaction-summary-prefix-cache-reuse.md). The call sets the provider-neutral `GenerateOptions.purpose` to `compaction`; adapters may map that purpose to model-hidden transport metadata, and the DeepSeek adapter sends `x-deepseek-harness-compact: 1`.
+`compactIfNeeded(agent, trigger, signal)` takes an explicit `'pressure' | 'context-overflow'` trigger and cancellation. It reads only the latest durable routed request; no header means no work, while any routed provider/model target uses the singleton estimator. `compactNow(agent, signal)` requires an idle agent and performs one useful balanced reduction even below pressure, returning `null` without writes when none exists. `compactRegion(start, end, agent, signal?)` uses `agent.session` as its single session identity and keeps an optional signal for explicit callers. The default summarizer resolves its target from explicit config, the latest logged routed target, then agent options, and records the provider/model pair after any `llm/stream` routing. It replays the routed request's prefix and appends the compaction directive as a trailing user message so the provider's warm KV cache is reused — see the [summary prefix-cache Agent Note](../bug-fix/2026-07-21-compaction-summary-prefix-cache-reuse.md). The result carries `llmStreamCall: true` because it consumed exactly one call through this context's LLM service; a subclass sets that marker only under the same condition, since retained `rawOutput` alone does not identify the call path. The call sets the provider-neutral `GenerateOptions.purpose` to `compaction`; adapters may map that purpose to model-hidden transport metadata, and the DeepSeek adapter sends `x-deepseek-harness-compact: 1`.
 
 ### Automatic pressure runs after successful durable step work
 
@@ -76,7 +76,7 @@ Because `SurfaceEventType` is closed, the summary cannot ride on a `compact/*` e
 ```
 compact/start    → log-only. Acquires the lock.
 [summarize older range via the backend]
-compact/summary  → log-only. Provenance: raw summary, range, shadowed seqs, token count.
+compact/summary  → log-only. Provenance: raw summary, local-call marker, range, shadowed seqs, token count.
 user/message     → canonical checkpoint source + surfaceOp { op:'replace', start, end }.
                    THE surface mutation (framed summary).
                    deriveMessages() renders it as a user-role message.
@@ -131,4 +131,4 @@ The lifecycle boundary makes crash state unambiguous:
 - **Loop:** Tests pin pre-step after the preceding `step/end` and before the next `step/start`, actual `agent/request` routing, closed failed steps, fresh retry numbering, and complete thrown/in-band overflow → compaction → reconstructed retry composition.
 - **Manual:** Maintenance serialization, marker ordering, injection retention, live/stale orphan classification, cancellation, close/flush failures, command mapping, and the queued TUI journey are pinned without a model key.
 - **With-key e2e:** A real model and bash session with lowered limits triggers compaction, records a complete `compact/start…end` pair, shrinks the surface, and finishes the task.
-- **Snapshot gap:** The summarization call is session-associated and logs `compact/summary`, but ordinary transcript replay does not derive its auxiliary response; keyless assembled coverage therefore needs an explicit replay override.
+- **Snapshot:** The assembled context-overflow scenario derives the auxiliary call from `compact/summary` only when `llmStreamCall: true` proves that the local LLM service consumed it; canonical reconstructed blocks pin the complete recovery without provider delta partitioning.

.agents/notes/implemented/feature/2026-06-18-compaction-capability-seam.zh.md

@@ -33,7 +33,7 @@ Status: implemented
 
 早期草案将完整算法（保留遍历、token 求和、文本提取）作为接口上的具体方法。这会将契约重新耦合到一种策略：想要不同保留策略或事件排序的后端必须与继承来的具体代码对抗。将三个操作都设为抽象，把所有*怎么做*的决策放在后端，并让接口保持为*做什么*的声明。token 测量根本不是压缩钩子；单例服务使多个消费方能够共享逐会话的回放折叠。
 
-`compactIfNeeded(agent, trigger, signal)` 接受显式的 `'pressure' | 'context-overflow'` 触发原因与取消信号。它只读取最新的持久化已路由请求；没有 header 就不执行工作，任何已路由的提供方/模型目标都使用单例估算器。`compactNow(agent, signal)` 要求 agent 处于 idle，即使未达到压力也进行一次有效的平衡缩减；不存在这种范围时返回 `null`，且不写入任何内容。`compactRegion(start, end, agent, signal?)` 将 `agent.session` 作为唯一会话身份，并为显式调用方保留可选 signal。默认摘要器依次从显式配置、最新记录的已路由目标和 agent 选项解析目标，并在任何 `llm/stream` 路由后记录提供方/模型对。它回放已路由请求的前缀，并将压缩指令追加为尾部 user 消息，从而复用提供方的热 KV cache；见[摘要前缀缓存 Agent Note](../bug-fix/2026-07-21-compaction-summary-prefix-cache-reuse.md)。该调用将提供方无关的 `GenerateOptions.purpose` 设为 `compaction`；适配器可以将此用途映射为对模型隐藏的传输元数据，DeepSeek 适配器会发送 `x-deepseek-harness-compact: 1`。
+`compactIfNeeded(agent, trigger, signal)` 接受显式的 `'pressure' | 'context-overflow'` 触发原因与取消信号。它只读取最新的持久化已路由请求；没有 header 就不执行工作，任何已路由的提供方/模型目标都使用单例估算器。`compactNow(agent, signal)` 要求 agent 处于 idle，即使未达到压力也进行一次有效的平衡缩减；不存在这种范围时返回 `null`，且不写入任何内容。`compactRegion(start, end, agent, signal?)` 将 `agent.session` 作为唯一会话身份，并为显式调用方保留可选 signal。默认摘要器依次从显式配置、最新记录的已路由目标和 agent 选项解析目标，并在任何 `llm/stream` 路由后记录提供方/模型对。它回放已路由请求的前缀，并将压缩指令追加为尾部 user 消息，从而复用提供方的热 KV cache；见[摘要前缀缓存 Agent Note](../bug-fix/2026-07-21-compaction-summary-prefix-cache-reuse.md)。该结果携带 `llmStreamCall: true`，因为生成它时恰好通过此上下文的 LLM 服务发起了一次调用；只有满足相同条件时，子类才设置该标记，因为单有保留的 `rawOutput` 并不能判定调用路径。该调用将提供方无关的 `GenerateOptions.purpose` 设为 `compaction`；适配器可以将此用途映射为对模型隐藏的传输元数据，DeepSeek 适配器会发送 `x-deepseek-harness-compact: 1`。
 
 ### 成功的持久步骤工作完成后运行自动压力检查
 
@@ -76,7 +76,7 @@ retry → next numbered step/start      ⟵ derives from the replacement surface
 ```
 compact/start    → log-only. Acquires the lock.
 [summarize older range via the backend]
-compact/summary  → log-only. Provenance: raw summary, range, shadowed seqs, token count.
+compact/summary  → log-only. Provenance: raw summary, local-call marker, range, shadowed seqs, token count.
 user/message     → canonical checkpoint source + surfaceOp { op:'replace', start, end }.
                    THE surface mutation (framed summary).
                    deriveMessages() renders it as a user-role message.
@@ -131,4 +131,4 @@ compact/end      → log-only. Releases the lock (carries `error` on a recoverab
 - **循环测试：** 测试固定 pre-step 发生在前一个 `step/end` 之后、下一个 `step/start` 之前，使用实际 `agent/request` 路由，关闭失败步骤，分配新的重试编号，并覆盖完整的抛出/带内溢出 → 压缩 → 重建重试组合。
 - **手动测试：** 无需模型密钥即可固定 maintenance 串行化、标记顺序、注入保留、活动／陈旧未匹配标记分类、取消、闭合／flush 失败、命令映射以及排队 TUI 流程。
 - **带密钥 e2e：** 真实模型和 bash 会话在降低的限制下触发压缩，记录完整的 `compact/start…end` 对，缩小 surface，并完成任务。
-- **快照缺口：** 摘要调用与会话关联并记录 `compact/summary`，但普通 transcript（文本记录）回放不会派生其辅助响应；因此，要实现无密钥的组装态覆盖，就必须显式提供回放 override。
+- **快照：** 组装后的上下文溢出场景仅在 `llmStreamCall: true` 证明本地 LLM 服务执行了辅助调用时，才从 `compact/summary` 派生该调用；规范重建的块在不固定提供方增量切分的情况下固定完整恢复过程。

.agents/notes/implemented/feature/2026-06-24-workspace-context.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/2026-06-24-workspace-context.md
-2026-06-24-workspace-context.md: 3c21b7599e6bb2dd759e4a040c1b999f1b0516cd
-2026-06-24-workspace-context.zh.md: e913ecb22b4692e1ee7b96d4f5effe2606775cf8
+2026-06-24-workspace-context.md: 53a580aee50752b7c6daeff5caa42ba6409c8885
+2026-06-24-workspace-context.zh.md: dc1d6b5dc2c9f576e7298c891a3a1145f123a8db

.agents/notes/implemented/feature/2026-06-24-workspace-context.md

@@ -14,7 +14,7 @@ The lifecycle has two distinct classes of content. The initial applicable chain
 
 ## Decision
 
-The implementation lives in `packages/context/workspace-context` as `@deepseek-ai/dsh-workspace-context`. It is a request-context extension, not a core service or a filesystem backend. The shared demo spine and Host Runtime mount it from an explicit `{ maxBytes } | false` deployment choice; `dsh web` enables a 65,536-byte budget while the Host Runtime's headless consumer disables it. The plugin consumes `agent/pre-step`, `tools/post-execute`, and the optional `ctx.fs` capability.
+The implementation lives in `packages/context/workspace-context` as `@deepseek-ai/dsh-workspace-context`. It is a request-context extension, not a core service or a filesystem backend. The shared demo spine and Host Runtime mount it from an explicit `{ maxBytes } | false` deployment choice; `dsh web` enables a 65,536-byte budget while the Host Runtime's headless consumer disables it. The plugin consumes `agent/pre-step`, immutable `tools/result` outcomes, `session/event` boundaries, and the optional `ctx.fs` capability.
 
 The plugin does not statically inject `fs`. Providerless product trees therefore boot normally and the plugin no-ops until a filesystem provider exists. All production reads go through that provider. Candidate probes resolve each path and stat the result, so a final-component symlink is followed to its target: a link to a regular file loads, while a missing path or a non-file target is a confirmed absence. Following repository-owned links across the trust boundary is a deliberate reversal of the original no-follow probe; the [instruction-symlink follow note](2026-07-21-follow-instruction-symlinks.md) owns that decision and its residual risk. The step signal and dynamic tool execution signal propagate through resolution, metadata probes, and streaming reads, so cancellation does not wait for an unrelated filesystem scan. A resolve or stat exception is classified as unavailable: it skips only that candidate and is never interpreted as the deletion of an already-loaded scope.
 
@@ -38,7 +38,7 @@ The baseline is a user-role `<system-reminder>` with `Instructions from: <path>`
 
 ### Dynamic Discovery And Refresh
 
-After a successful first-party `read`, `write`, or `edit` call, the `tools/post-execute` listener reconciles the touched descendant chain and every scope already known to the session. A newly reached scope is returned through `additionalContexts` for the next request using an `Additional instructions from: <path>` system-reminder. Under Code Mode, `run_code` defers sub-dispatch contexts onto its outer result, so the same update is appended only after the parent result rather than being injected mid-call.
+After a successful first-party `read`, `write`, or `edit` call, the immutable `tools/result` observer reconciles the touched descendant chain and every scope already known to the session, then queues an `Additional instructions from: <path>` system-reminder in the agent inbox for the next request. Under Code Mode, successful sub-dispatch touches bubble through opaque parent execution tokens until the top-level result settles. A touch produced inside an agent-loop step does not begin its asynchronous projection until the durable `step/end`; a direct tool execution outside an open step projects immediately. The two boundaries keep result and step adjacency deterministic without making the tool pipeline await filesystem discovery.
 
 A content edit appends `Updated instructions from: <path>`, states that the new content replaces the previous content, and includes the complete current file. If precedence changes from one candidate to another, the message also names the previous path and says it no longer applies. If no candidate remains, the plugin appends `Instructions removed: <path>` and states that the previously loaded instructions no longer apply.
 
@@ -50,7 +50,7 @@ Shell commands are not discovery triggers. Local bash calls start fresh shells,
 
 Every workspace context event stores versioned metadata with `{ action, scope, path, digest? }`, where `digest` is SHA-1 over the loaded content. Complete baselines additionally carry `baseline: true` and `baselineIdentity`. The model-facing prompt has no HTML comments, hidden markers, or headings that are parsed back into state.
 
-At reconciliation time the plugin scans workspace-sourced `user/message` events and derives the latest state for each visible scope. A short per-session pending map begins only after the immutable top-level `tools/result` proves an `additionalContexts` entry survived every post-execute listener, then covers the interval before the loop appends that context to the log. Each entry records the open `{ turn, step }`: an equal durable `user/message` at or after its sequence boundary confirms and removes it, while a matching `step/end` arriving first means the loop discarded its context buffer, so the plugin removes both the pending entry and its version-cache fast path. A nested Code Mode result stages its changes under the parent's opaque execution token so repeated sub-dispatches in one run do not duplicate them; the parent result rolls that provisional state back and commits only contexts retained by outer policy.
+At reconciliation time the plugin scans workspace-sourced `user/message` events and derives the latest state for each visible scope. Successful nested touches aggregate under their parent execution token, including when a later composite result is blocked; the top-level result transfers them either to the open session step or directly to a per-agent projection queue. A `step/end` releases its staged touches only after that boundary is durable, and the next `agent/pre-step` waits for the serialized projections. Each projection composes against visible history plus the current inbox and replaces the single pending workspace context instead of accumulating intermediate renderings.
 
 An unchanged path and digest is suppressed. A logged removal is a tombstone, so a reappearing candidate becomes a new `set`. Resume works from persisted metadata: a compatible visible baseline supplies comparison state rather than causing another complete baseline to be appended. If compaction removes an instruction event from the visible surface, that state no longer suppresses a later load, matching the fact that the model can no longer see it. Only changes actually included under the byte budget enter metadata or pending state, so an omitted file remains eligible on a later touch.
 
@@ -62,7 +62,7 @@ There is intentionally no watcher. Detection occurs at the next successful struc
 
 `maxBytes` is required and applies separately to a rendered baseline or one dynamic reconciliation batch; there is no implicit or unbounded render budget. Non-positive and non-finite values disable loading. When content exceeds the budget, broader files are omitted before the most-specific file is truncated. A visible `Workspace instruction budget ...` notice names omitted and truncated paths and byte counts, and output never exceeds the configured bytes.
 
-`maxSourceBytes` is a positive per-file cap with a 1 MiB default. The loader checks reported size before reading and still consumes content through `streamText()` with a running UTF-8 byte count, so missing/stale metadata cannot force an unbounded allocation. An oversized winning candidate is unavailable rather than a reason to fall through to another same-directory name. The plugin deliberately keeps no process-wide cache and never retains instruction prose. It keeps only `{ path, version, digest }` per effective scope in a `WeakMap<Session, Map<scope, state>>`: a matching provider `FsVersion` plus matching effective prompt state skips the read, while a changed version triggers a bounded read and SHA-1 confirmation. SHA-1 remains the cross-provider content identity persisted in visible structured metadata; provider versions are only an in-memory invalidation fast path. Cache transitions for model-visible changes commit only when the corresponding context survives the complete tool-result policy chain, and are invalidated if that accepted context is later dropped with its aborted step before reaching the log.
+`maxSourceBytes` is a positive per-file cap with a 1 MiB default. The loader checks reported size before reading and still consumes content through `streamText()` with a running UTF-8 byte count, so missing/stale metadata cannot force an unbounded allocation. An oversized winning candidate is unavailable rather than a reason to fall through to another same-directory name. The plugin deliberately keeps no process-wide cache and never retains instruction prose. It keeps only `{ path, version, digest }` per effective scope in a `WeakMap<Session, Map<scope, state>>`: a matching provider `FsVersion` plus matching effective prompt state skips the read, while a changed version triggers a bounded read and SHA-1 confirmation. SHA-1 remains the cross-provider content identity persisted in visible structured metadata; provider versions are only an in-memory invalidation fast path. Dynamic cache transitions occur while a serialized projection composes the one desired inbox context, and the next pre-step waits for that projection before deciding what enters the request.
 
 ## Alternatives considered
 
@@ -78,7 +78,7 @@ There is intentionally no watcher. Detection occurs at the next successful struc
 
 ## Consequences
 
-Workspace guidance is isolated per session and shared by the demo front doors, Web Host, and every tool presentation mode. Initial, nested, and changed instructions are durable and replayable. The generic session/agent context contract carries typed source data through injected messages and post-tool `additionalContexts` arrays without flattening entries.
+Workspace guidance is isolated per session and shared by the demo front doors, Web Host, and every tool presentation mode. Initial, nested, and changed instructions are durable and replayable. The generic session/agent context contract carries typed source data through inbox-staged and durably entered user messages without flattening entries.
 
 Repository text remains untrusted input. Lower-authority user-role framing, explicit precedence language, and delimiter escaping reduce risk but do not eliminate prompt injection. Following a candidate symlink to its target widens that surface to off-tree content, so the permission and sandbox layers that confine `ctx.fs` to trusted roots are the boundary that treats workspace files as data rather than authority (the [instruction-symlink follow note](2026-07-21-follow-instruction-symlinks.md) owns the residual risk).
 

.agents/notes/implemented/feature/2026-06-24-workspace-context.zh.md

@@ -14,7 +14,7 @@ Status: implemented
 
 ## 决策
 
-该实现在 `packages/context/workspace-context` 中，包（package）名为 `@deepseek-ai/dsh-workspace-context`。它是请求上下文扩展，不是核心服务或文件系统后端。共享 demo 主干与 Host Runtime 根据显式的 `{ maxBytes } | false` 部署选择挂载它；`dsh web` 启用 65,536 字节预算，Host Runtime 的 headless 消费方则禁用它。该插件使用 `agent/pre-step`、`tools/post-execute` 和可选的 `ctx.fs` 功能。
+该实现在 `packages/context/workspace-context` 中，包（package）名为 `@deepseek-ai/dsh-workspace-context`。它是请求上下文扩展，不是核心服务或文件系统后端。共享 demo 主干与 Host Runtime 根据显式的 `{ maxBytes } | false` 部署选择挂载它；`dsh web` 启用 65,536 字节预算，Host Runtime 的 headless 消费方则禁用它。该插件使用 `agent/pre-step`、不可变的 `tools/result` 结果、`session/event` 边界和可选的 `ctx.fs` 功能。
 
 插件不会静态注入 `fs`。因此，不带提供方的产品树仍能正常启动；在文件系统提供方出现之前，插件保持无操作。所有生产读取都通过该提供方完成。候选项探测会解析每个路径并对结果执行 stat，因此会跟随最终路径组件的符号链接至其目标：指向普通文件的链接会被加载，缺失路径或非文件目标则确认为不存在。允许仓库拥有的链接跨越信任边界，是对最初不跟随探测方式的刻意反转；[跟随指令符号链接记录](2026-07-21-follow-instruction-symlinks.md)负责说明该决策及其残余风险。步骤信号与动态工具执行信号会贯穿解析、元数据探测和流式读取，因此取消不会等待无关的文件系统扫描。解析或 stat 异常归类为不可用：它只跳过该候选项，绝不被解释为已经加载的作用域被删除。
 
@@ -38,7 +38,7 @@ Status: implemented
 
 ### 动态发现与刷新
 
-第一方 `read`、`write` 或 `edit` 调用成功后，`tools/post-execute` 监听器会协调被触及的后代路径链，以及该会话已经知道的每个作用域。新到达的作用域通过 `additionalContexts` 返回，并在下一次请求中使用 `Additional instructions from: <path>` system-reminder。在 Code Mode 下，`run_code` 会把子分发上下文延后至其外层结果，因此同一更新只会在父结果之后追加，而不会在调用中途注入。
+第一方 `read`、`write` 或 `edit` 调用成功后，不可变的 `tools/result` 观察器会协调被触及的后代路径链，以及该会话已经知道的每个作用域，然后在 agent inbox 中排入一条 `Additional instructions from: <path>` system-reminder，供下一次请求使用。在 Code Mode 下，成功的子分派 touch 会沿不透明的父级执行 token 逐层上浮，直到顶层结果落定。在 agent loop 步骤内产生的 touch，须等持久 `step/end` 后才开始异步投影；打开的步骤之外直接执行工具时，则立即投影。这两个边界在不让工具流水线等待文件系统发现的前提下，保证结果／步骤的相邻关系具有确定性。
 
 内容编辑会追加 `Updated instructions from: <path>`，说明新内容取代先前内容，并包含当前的完整文件。如果优先级从一个候选项变为另一个，消息还会指出先前路径并说明它不再适用。如果没有候选项保留，插件会追加 `Instructions removed: <path>`，并说明先前加载的指令不再适用。
 
@@ -50,7 +50,7 @@ shell 命令不会触发发现。本地 bash 调用会启动全新的 shell，
 
 每个工作区上下文事件都会存储带版本的元数据，其形态为 `{ action, scope, path, digest? }`；`digest` 是对已加载内容计算的 SHA-1。完整基线还会额外携带 `baseline: true` 和 `baselineIdentity`。模型可见提示词中没有 HTML 注释、隐藏标记，也没有会被解析回状态的标题。
 
-协调时，插件扫描带工作区来源的 `user/message` 事件，并派生每个可见作用域的最新状态。一个简短的逐会话待处理映射只会在不可变的顶层 `tools/result` 证明某个 `additionalContexts` 条目经过所有 post-execute 监听器后仍然保留时开始记录；随后，它覆盖循环将该上下文追加到日志之前的间隔。每个条目记录开启状态的 `{ turn, step }`：如果相同的持久 `user/message` 出现在其序列边界或之后，该条目得到确认并被移除；如果匹配的 `step/end` 先到达，则说明循环丢弃了上下文缓冲区，插件会同时移除待处理条目及其版本缓存快速路径。嵌套的 Code Mode 结果会把变更暂存在父级的不透明执行 token 下，确保一次运行中的重复子分发不会产生重复项；父级结果会回滚这份临时状态，并且只提交外层策略保留的上下文。
+协调时，插件扫描带工作区来源的 `user/message` 事件，并派生每个可见作用域的最新状态。即使后续复合结果被拦截，成功的嵌套 touch 也会聚合到父级执行 token 下；顶层结果会将它们交给打开的会话步骤，或直接交给逐 agent 投影队列。`step/end` 只会在自身边界持久化后释放其暂存的 touch；下一次 `agent/pre-step` 会等待串行投影完成。每次投影都会根据可见历史和当前 inbox 进行组合，并替换唯一一条待处理工作区上下文，而不会累积中间渲染结果。
 
 路径和 digest 均未变化时会被抑制。日志中的移除操作是一条墓碑记录，因此重新出现的候选项会成为新的 `set`。恢复操作从持久化元数据继续工作：兼容的可见基线会提供比较状态，而不会导致再次追加完整基线。如果压缩从可见表面移除某条指令事件，该状态不再抑制后续加载，这与模型已经无法看见它的事实一致。只有真正纳入字节预算的变更才会进入元数据或待处理状态，因此被省略的文件在之后的触碰中仍有资格加载。
 
@@ -62,7 +62,7 @@ shell 命令不会触发发现。本地 bash 调用会启动全新的 shell，
 
 `maxBytes` 是必填项，分别作用于渲染后的基线或单个动态协调批次；系统不存在隐式或无界的渲染预算。非正数或非有限值会禁用加载。内容超过预算时，系统会先省略较宽泛的文件，再截断最具体的文件。可见的 `Workspace instruction budget ...` 提示会指出被省略和截断的路径与字节数，并且输出绝不超过配置字节数。
 
-`maxSourceBytes` 是正数的逐文件上限，默认为 1 MiB。loader 会在读取前检查报告的大小，同时仍通过 `streamText()` 消费内容并持续统计 UTF-8 字节数，因此缺失／陈旧的元数据无法迫使其进行无界分配。过大的胜出候选项会被视为不可用，而不是改为同目录中的下一个名称。插件刻意不保留进程级缓存，也绝不保留指令正文。它只为每个有效作用域保存 `{ path, version, digest }`，并将这些状态放在 `WeakMap<Session, Map<scope, state>>` 中：提供方 `FsVersion` 与有效提示词状态同时匹配时跳过读取；版本变化则触发有界读取和 SHA-1 确认。SHA-1 仍是持久化在可见结构化元数据中的跨提供方内容标识；提供方版本只作为内存中的失效快速路径。模型可见变更的缓存转换只有在相应上下文通过完整的工具结果策略链后才会提交；如果该已接受上下文随后与中止步骤一起被丢弃、未能进入日志，缓存转换就会失效。
+`maxSourceBytes` 是正数的逐文件上限，默认为 1 MiB。loader 会在读取前检查报告的大小，同时仍通过 `streamText()` 消费内容并持续统计 UTF-8 字节数，因此缺失／陈旧的元数据无法迫使其进行无界分配。过大的胜出候选项会被视为不可用，而不是改为同目录中的下一个名称。插件刻意不保留进程级缓存，也绝不保留指令正文。它只为每个有效作用域保存 `{ path, version, digest }`，并将这些状态放在 `WeakMap<Session, Map<scope, state>>` 中：提供方 `FsVersion` 与有效提示词状态同时匹配时跳过读取；版本变化则触发有界读取和 SHA-1 确认。SHA-1 仍是持久化在可见结构化元数据中的跨提供方内容标识；提供方版本只作为内存中的失效快速路径。动态缓存转换发生在串行投影组合唯一一条目标 inbox 上下文时；下一次 pre-step 会等待该投影完成，再决定哪些内容进入请求。
 
 ## 考虑过的替代方案
 
@@ -78,7 +78,7 @@ shell 命令不会触发发现。本地 bash 调用会启动全新的 shell，
 
 ## 后果
 
-工作区指引按会话隔离，并由 demo 入口、Web Host 与每一种工具展示模式共享。初始、嵌套与变更指令都保持持久且可回放。通用的会话／agent 上下文契约通过注入消息与工具执行后的 `additionalContexts` 数组携带带类型的来源数据，而不会把条目展平。
+工作区指引按会话隔离，并由 demo 入口、Web Host 与每一种工具展示模式共享。初始、嵌套与变更指令都保持持久且可回放。通用的会话／agent 上下文契约通过先在 inbox 中暂存、再持久进入的 user 消息携带带类型的来源数据，而不会把条目展平。
 
 仓库文本仍是不受信任的输入。低权威 user 角色框架、显式优先级说明和分隔符转义可以降低风险，但无法消除提示词注入。跟随候选符号链接到目标，会把该攻击面扩大至树外内容；因此，把 `ctx.fs` 限制在可信根目录内的权限与沙箱层才是真正的边界，它们让系统把工作区文件当作数据而不是权威（[跟随指令符号链接记录](2026-07-21-follow-instruction-symlinks.md)负责说明残余风险）。
 

.agents/notes/implemented/feature/2026-07-05-skill-system.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/2026-07-05-skill-system.md
-2026-07-05-skill-system.md: dd2fb1d22949f55ea7cb2c9f280e7cfbfcbbb226
-2026-07-05-skill-system.zh.md: 96656a8e1dfc2ae1ce7301ba29e7739349b6aab6
+2026-07-05-skill-system.md: a998d70ec934aed4bf7ce32aa711abd47b508a1d
+2026-07-05-skill-system.zh.md: 4fa7c4fd657c2f41f16b30679ec95e61a75f8a0c

.agents/notes/implemented/feature/2026-07-05-skill-system.md

@@ -14,9 +14,11 @@ DeepSeek Harness uses the same primitive so project-specific review, plugin-auth
 
 `@deepseek-ai/dsh-skill` is the pure provider registry (`ctx.skills`), `@deepseek-ai/dsh-skill-local` is the shipped local filesystem provider, and `@deepseek-ai/dsh-tool-skill` owns the durable session catalog and model-facing loader tool. `dsh-agent-spine-demo` loads the registry, local provider, and consumer by default so TUI, headless, and ACP apps get the same behavior while embedded or remote providers contribute skills without changing the registry or consumer. Its `skills` config forwards `registry`, `local`, and `tool` branches to those owners.
 
+Dedicated packaged providers can contribute immutable skills without filesystem discovery. The shipped CLI declares `@deepseek-ai/dsh-skill-badge` disabled by default; enabling its composition row contributes the official badge instructions through the same registry and consumer ([decision](2026-08-06-bundled-dsh-badge-skill.md)).
+
 Provider plugins register synchronously during `apply()`. Provider membership is direct effect-owned state: registration and disposal invalidate completed catalogs synchronously, and discovery reads the current provider map on demand rather than observing registry-change events. Provider catalogs return ranked candidates from awaited `list()` calls, where remote providers perform initialization, authentication, and discovery while honoring the lookup abort signal. The registry validates each candidate, resolves same-name skills first-wins by rank, provider registration order, and provider-local order, then sorts summaries by skill name for deterministic consumers. It caches only completed catalog snapshots and retries when a provider/runtime revision changes during discovery, so an unload cannot freeze a stale, unresolvable skill into a session catalog. Runtime `ctx.skills.register(...)` remains a convenience for embedded in-process skills and uses project-over-user priority; `runtime` is reserved as the registry-owned provider name.
 
-The local provider scans cwd-sensitive project roots, custom roots, and user roots in first-wins rank order: project `.dsh`, project `.agents`, `customSkillDirs`, user `.dsh`, then user `.agents`. The user `.dsh/skills` scan skips `.system` so a system-owned directory is not treated as normal user content. DeepSeek Harness does not ship built-in system skills; embedded or remote providers supply additional skills when configured.
+The local provider scans cwd-sensitive project roots, custom roots, and user roots in first-wins rank order: project `.dsh`, project `.agents`, `customSkillDirs`, user `.dsh`, then user `.agents`. The user `.dsh/skills` scan skips `.system` so a system-owned directory is not treated as normal user content. The local provider does not synthesize built-in system skills; configured bundled roots and dedicated providers supply additional skills.
 
 Each skill is either `<name>/SKILL.md` or `<name>.md` with YAML frontmatter. `name` and `description` are required; `whenToUse`, `metadata`, `disable-model-invocation`, and `user-invocable` are optional. Names are kebab-case. The invocation fields project into a typed nested policy as defined by the [independent model and user invocation decision](2026-07-28-skill-invocation-policy.md); the parser rejects the old camel-case spellings. YAML frontmatter is parsed with the `yaml` package instead of `js-yaml` or a hand-written parser: `yaml` is the already-declared modern parser for this package's limited frontmatter needs, and a narrow parser would either reject valid YAML users expect to work or grow into an unreviewed YAML subset.
 
@@ -48,7 +50,7 @@ The data structures and catalog/tool contract are documented in [skills.md](../.
 
 The agent-core spine includes one catalog contributor, one local provider, and one model-facing tool. Skill discovery is cwd-sensitive, so callers that create agents with different session cwd values can observe different project skill overrides by design.
 
-The catalog is deterministic for a fixed root set and runtime registration revision, but disk changes are not watched; discovery is memoized until runtime registration invalidates the cache or the process restarts.
+The catalog is deterministic for a fixed root set and runtime registration revision. The local provider watches configured roots and invalidates completed catalogs after relevant disk changes; runtime registration and provider disposal also invalidate them.
 
 ## Deferred
 

.agents/notes/implemented/feature/2026-07-05-skill-system.zh.md

@@ -14,9 +14,11 @@ DeepSeek Harness 使用同一原语，使项目特定的评审、插件编写和
 
 `@deepseek-ai/dsh-skill` 是纯提供方注册表（`ctx.skills`），`@deepseek-ai/dsh-skill-local` 是随附的本地文件系统提供方，`@deepseek-ai/dsh-tool-skill` 负责持久化会话目录与面向模型的 loader 工具。`dsh-agent-spine-demo` 默认加载注册表、本地提供方和消费方，使 TUI、headless 与 ACP（Agent Client Protocol）应用获得相同行为，同时嵌入式或远程提供方可在不修改注册表或消费方的前提下贡献 skill。其 `skills` 配置将 `registry`、`local` 和 `tool` 分支分别转发给对应的所有者。
 
+专用的随包提供方可以贡献不可变的 skill，无需文件系统发现。交付的 CLI（命令行界面）默认将 `@deepseek-ai/dsh-skill-badge` 声明为禁用；启用其组合配置行，就会通过同一个注册表和消费方贡献官方徽章指令（见[决策](2026-08-06-bundled-dsh-badge-skill.md)）。
+
 提供方插件在 `apply()` 期间同步注册。提供方成员资格是由直接 effect 持有的状态：注册与 dispose（资源释放）同步地使已完成的目录失效，发现操作按需读取当前提供方映射而非监听注册表变更事件。提供方目录从等待的 `list()` 调用返回排序后的候选项，远程提供方在此过程中执行初始化、认证和发现，同时遵守查找的 abort 信号。注册表校验每个候选项，按排名、提供方注册顺序和提供方内部顺序以先到先得方式解决同名 skill 冲突，然后按 skill 名称排序摘要以保证消费方获得确定性结果。它仅缓存已完成的目录快照，并在发现过程中提供方/运行时修订版本发生变化时重试，因此卸载操作不会将一个陈旧且不可解析的 skill 冻结到会话目录中。运行时 `ctx.skills.register(...)` 仍作为嵌入式进程内 skill 的便捷方式保留，使用 project 优先于 user 的优先级；`runtime` 保留为注册表拥有的提供方名称。
 
-本地提供方按先到先得的排名顺序扫描 cwd 敏感的项目根目录、自定义根目录和用户根目录：项目 `.dsh`、项目 `.agents`、`customSkillDirs`、用户 `.dsh`，然后是用户 `.agents`。用户 `.dsh/skills` 扫描跳过 `.system`，以免系统拥有的目录被当作普通用户内容处理。DeepSeek Harness 不随附内置系统 skill；嵌入式或远程提供方在配置后提供额外 skill。
+本地提供方按先到先得的排名顺序扫描 cwd 敏感的项目根目录、自定义根目录和用户根目录：项目 `.dsh`、项目 `.agents`、`customSkillDirs`、用户 `.dsh`，然后是用户 `.agents`。用户 `.dsh/skills` 扫描跳过 `.system`，以免系统拥有的目录被当作普通用户内容处理。本地提供方不会合成内置系统 skill；已配置的 bundled 根目录和专用提供方会提供额外 skill。
 
 每个 skill 是 `<name>/SKILL.md` 或带 YAML frontmatter 的 `<name>.md`。`name` 和 `description` 为必填；`whenToUse`、`metadata`、`disable-model-invocation` 和 `user-invocable` 为可选。名称采用 kebab-case。调用字段会投影到类型化的嵌套策略中，具体由[模型与用户独立调用决策](2026-07-28-skill-invocation-policy.md)定义；解析器会拒绝旧的驼峰拼写。YAML frontmatter 使用 `yaml` 包（package）解析，而非 `js-yaml` 或手写解析器：`yaml` 是本包有限 frontmatter 需求已声明的现代解析器，窄解析器要么拒绝用户预期可用的合法 YAML，要么膨胀为一个未经评审的 YAML 子集。
 
@@ -48,7 +50,7 @@ DeepSeek Harness 使用同一原语，使项目特定的评审、插件编写和
 
 agent-core 主干包含一个目录贡献者、一个本地提供方和一个面向模型的工具。Skill 发现是 cwd 敏感的，因此以不同会话 cwd 值创建 agent 的调用方可以按设计观察到不同的项目 skill 覆盖。
 
-目录对于固定的根目录集合和运行时注册修订版本是确定性的，但不监视磁盘变化；发现结果被缓存，直到运行时注册使缓存失效或进程重启。
+目录对于固定的根目录集合和运行时注册修订版本是确定性的。本地提供方会监视已配置的根目录，并在发生相关磁盘变化后使已完成的目录失效；运行时注册和提供方释放也会使其失效。
 
 ## 延后
 

.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: aed5ac1ceb02130ce97a8c83c0f77869fdc32146
-2026-07-06-sandbox.zh.md: db95b1a5b7a7cae1e0fcdd8deba9dcb6ad020a67
+2026-07-06-sandbox.md: 583b388815cd9b2b9cf94ce393839169ce3ffac3
+2026-07-06-sandbox.zh.md: e435b671a42ca5c3ea4f6800bf91d6e006da35d3

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

@@ -64,7 +64,7 @@ Left open, for the phase that needs them: whether network restriction arrives as
 
 The launcher is a ~300-line C program (plain C11 over the raw Landlock UAPI — no libraries beyond a statically linked musl, so the audit surface is that one file plus the kernel's stable syscall contract): `--ro <path>` / `--rw <path>` grants, `--`, the wrapped argv; it installs the ruleset on itself and `exec`s (rulesets are inherited across `execve`, and it sets `no_new_privs` before restricting); `--probe` enforces a maximal ruleset in a short-lived child and exits 0 only when the kernel actually enforces; every launcher failure exits 125 without running the child and prints a fatal `landlock-run:` line. A successfully exec'd child may also return 125, so status alone is not launcher evidence. An older ABI prints the exact `landlock-run: partial enforcement (older Landlock ABI)` notice before it executes the child, so that line is not fatal evidence.
 
-The Landlock launcher source and package workspace live at `native/landlock-run`, next to the harness consumers. The standalone [`node-addon-landlock-run`](https://github.com/deepseek-harness/node-addon-landlock-run) repository is the release mirror used to pack and publish the npm package family; `native/README.md` owns the export procedure. Platform binaries are selected by npm, and the entry package owns path resolution, probing, CLI flags, the fatal prefix, and the partial-enforcement notice while the harness maps sandbox modes to grants. Versioning the entry point with its binaries keeps probe parsing and launch syntax aligned.
+The Landlock launcher source and package family live at `native/landlock-run`, next to the harness consumers and inside the root pnpm workspace. The [in-repository Landlock release decision](../process/2026-08-06-in-repository-landlock-release.md) owns the shared lockfile, native build, pack rehearsal, and npm publication boundary. Platform binaries are selected by npm, and the entry package owns path resolution, probing, CLI flags, the fatal prefix, and the partial-enforcement notice while the harness maps sandbox modes to grants. Versioning the entry point with its binaries keeps probe parsing and launch syntax aligned.
 
 Backend profiles share the mode contract but differ in necessary host grants. Landlock and Seatbelt allow only `/dev/null` in read-only mode; workspace-write also permits their required host temp roots. Each wrap carries backend-specific denial signatures. Landlock reports partial enforcement on older ABIs that cannot govern every operation, while successful bwrap and Seatbelt profiles report full enforcement.
 
@@ -118,7 +118,7 @@ fs/web/todo execute in-process, so their sandbox semantics are policy at their s
 ### Testing
 
 - **Unit:** pin platform selection and profiles, direct provider-argv handoff, spawn-level failures with invalid-workdir controls, missing/non-executable/missing-interpreter evidence, malformed-runner negative controls, confined `BASH_ENV` ordering, structured runner classification (including partial-Landlock notice-only child outcomes, gated fatal evidence, child exits 126/127, and foreground/background parity), per-call mode/root resolution, per-process facts, escalation validation and outcomes, permission preset folding and write-through, and runtime-context ordering and materialization.
-- **Keyless real-runner:** exercise bwrap, Landlock, and Seatbelt against real filesystem effects at provider and bash-consumer layers; one real Cordis context concurrently drives two project sessions through shipped bash and fs tools, proving own-root success and sibling-root denial. Packed-install coverage proves the registry launcher remains executable. CI rejects a silent all-skip.
+- **Keyless real-runner:** exercise bwrap, Landlock, and Seatbelt against real filesystem effects at provider and bash-consumer layers; one real Cordis context concurrently drives two project sessions through shipped bash and fs tools, proving own-root success and sibling-root denial. Packed-install coverage installs the current checkout's native tarballs and proves the launcher remains executable and byte-identical. CI rejects a silent all-skip.
 - **With-key:** start the real ACP composition in read-only mode, let a model-driven bash write hit the runner's denial marker, then drive the bridge answerer and disk effect through granted and rejected workspace-write retries; unavailable credentials or runners self-skip.
 - **Snapshot:** pin the atomic current-policy context and both scripted approval branches. A real ACP example scenario places its session under the user home while the deployment fallback points at `/tmp`, then pins both the workspace-write runtime-context message and a successful deployment-selected mutation; this distinguishes session-root resolution from the process fallback without depending on runner-specific denial text. A POSIX fake partial-Landlock provider pins direct bash `false` as an ordinary child result and a missing provider executable as foreground/background infrastructure failure through the assembled app. Other snapshots start unconfined so unrelated fixtures remain platform-independent.
 
@@ -128,15 +128,15 @@ Each phase gets its full design when picked up, validated against the code at th
 
 - **Second consumer** — `subagent-acp` optionally confines child agents (per-call policy; unconfined default — a child agent must write its own persistence).
 - **More environments** — an environment-coherent capability group example (e.g. bash+fs against one container).
-- **Windows chain** — `PLATFORM_CHAINS.win32` is reserved and empty (fail-closed); filling it means a confinement runner from the AppContainer/restricted-token family, shipped from its own repository on the `node-addon-landlock-run` template, plus its profile dialect, denial signatures, and runner-failure rules. Wrapping the third-party landstrip runner instead was [considered and rejected](../../rejected/feature/2026-07-26-evaluate-landstrip-for-windows-sandbox-rung.md) — not battle-tested enough for a security invariant.
+- **Windows chain** — `PLATFORM_CHAINS.win32` is reserved and empty (fail-closed); filling it means a confinement runner from the AppContainer/restricted-token family, shipped from the main repository under `native/` following the `@deepseek-ai/node-addon-landlock-run` template, plus its profile dialect, denial signatures, and runner-failure rules. Wrapping the third-party landstrip runner instead was [considered and rejected](../../rejected/feature/2026-07-26-evaluate-landstrip-for-windows-sandbox-rung.md) — not battle-tested enough for a security invariant.
 
 ## Alternatives considered
 
 - **Command-string heuristic preflight** — rejected: cannot understand expansion/subprocesses/symlinks; the strict attempt (run it, let the kernel decide) is the only trustworthy denial signal.
 - **Functionally probe even a platform's sole backend** — rejected: probing arbitrates between candidates; with one there is nothing to decide, and probe cost taxes the first confined command of every session (prohibitive for heavy future backends). The runner's own exec-time fail-closed refusal plus structured `runnerFailureRules` classification carries the safety property instead.
-- **Commit the built launcher binaries** — rejected: a binary in a diff is unreviewable and churns history; reviewed source + native CI builds + the launcher repo's byte-pinned publish rehearsal keep bytes out of every tree.
+- **Commit the built launcher binaries** — rejected: a binary in a diff is unreviewable and churns history; reviewed source + native CI builds + the main repository's byte-pinned publish rehearsal keep bytes out of every tree.
 - **Compile the launcher on install** — rejected: pushes a C toolchain onto every consumer; a fallback that exists only where a compiler happens to be is not a fallback.
-- **Cross-compile both architectures from one builder** — rejected: requires carrying a pinned cross toolchain (rustup targets, zig, or a container image) solely to rebuild two ~70 KB binaries; per-architecture native runners already exist and each builds its own platform package (the `node-addon-require-builtin` model, the launcher repo's own pipeline).
+- **Cross-compile both architectures from one builder** — rejected: requires carrying a pinned cross toolchain (rustup targets, zig, or a container image) solely to rebuild two ~70 KB binaries; per-architecture native runners already exist and each builds its own platform package (the `node-addon-require-builtin` model, retained by the main repository's native pipeline).
 - **No fallback (bwrap or fail closed)** — rejected: concentrates failure on the hosts a sandbox matters most, degrading to `danger-full-access` by resignation.
 - **Keep the mechanism inside `dsh-bash-sandbox`** — rejected: blocks the existing second consumer, makes future phases read mode out of a bash plugin's config, and cannot express escalation.
 - **Config-fixed mode on the provider** — rejected: one mode per process; cannot serve concurrent consumers with different policies nor the one-shot widened retry.
@@ -175,7 +175,7 @@ Costs and accepted limits:
 - **The Seatbelt rung leans on Apple's deprecated-but-shipped `sandbox-exec` CLI.** As darwin's sole candidate it is selected without probing, so a future removal under a usable workdir surfaces as a runner-attributable spawn failure and an executable refusal through its fatal signature — both become `SANDBOX_UNAVAILABLE`, and the command never runs; fail closed, never open.
 - **Landlock confinement is only as complete as the running kernel's ABI.** Reported as `enforcement: 'partial'` rather than refused — the deliberate trade that keeps the fallback available on older-kernel hosts.
 - **Runner attribution uses an in-band protocol.** Exit status plus stderr cannot cryptographically identify the writer, so a confined child can mimic a fatal runner line and status to cause an availability/diagnostic false attribution. The conjunction and exact notice exclusion reduce accidental matches; this is not a sandbox bypass because the child is already confined.
-- **The launcher arrives as a registry dependency.** Trusted through its own repository's release pipeline (reviewed C source, native CI builders, byte-pinned publish rehearsal) plus this repo's version pin — the real-kernel e2e legs are what vouch for behavior through the installed bytes.
+- **The launcher is a workspace dependency in source and an npm dependency after publication.** The main repository tests reviewed C source, native CI builds, and byte-pinned local tarballs together before publishing the same package family; the real-kernel e2e legs vouch for behavior through those installed bytes.
 - **The model may over-ask.** Escalating without denial grounding, or picking `danger-full-access` where `workspace-write` suffices: the description steers and the enum forces the ladder, but the human prompt is the actual gate; the `approval/asked` reasons make over-asking auditable, and a `prepend` policy answerer can auto-reject patterns a deployment never wants.
 - **The advertised target set is static while the effective mode is per-session** (schemas are registry-global) — a session already at the widest mode is still offered the fields. Harmless by construction: the strict-wider check at execution, not the enum, is the safety boundary — a non-widening request fails with its own text and never prompts anyone.
 - **A granted escalation is not a working sandbox.** An unavailable backend still fails closed even for a granted escalation to a confining mode — at `confine()` when the platform has no chain or every probe fails, through the spawn channel when the selected executable cannot start, or through a structured rule when a started runner refuses — while a granted `danger-full-access` run never touches the provider at all: there the grant, not the probe, is the authority.
@@ -187,7 +187,7 @@ Costs and accepted limits:
 - **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 task stamps `sandbox.runnerFailed` and renders its own marker. A `SubprocessService` that synchronously throws the same provenanced `ENOENT`/`EACCES` shape 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 — Windows today?** `confine()` throws the fail-closed `SANDBOX_UNAVAILABLE` and the command never spawns; `win32` is a reserved EMPTY chain, pinned by test to fail closed identically until a Windows runner fills it (§ Deferred phases).
-- **`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 registry-installed Landlock launcher, and the verdict is cached for the provider's lifetime.
+- **`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?** No — `SandboxMode` claims FILE effects only; the bwrap profile deliberately does not unshare pid, and no backend claims network. Whether network restriction becomes its own knob is left open in § The seam.
 - **Which tools actually run confined?** OS subprocesses through `ctx.bash` — the bash tools, and hook commands transitively — plus the filesystem tools (`read`/`write`/`edit`) through the sandboxed `ctx.fs` provider (the [cross-family fs sandbox RFC](2026-07-14-cross-family-fs-sandbox.md)): bash confines via the OS runner, fs via an in-process path fence, both keying off the same `ctx.sandboxPolicy` mode. web/todo stay in-process and unfenced (web's only effect is network, outside the file-effect mode vocabulary).
 - **Does a granted escalation persist?** No. The grant is consumed by the exact foreground or background call that asked; every neighboring call keeps its own effective mode. A later background denial surfaces through `task_output` and may ground a new exact-command retry.

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

@@ -64,7 +64,7 @@ OS 子进程约束适用于 bash 执行器（包括钩子命令），后续还
 
 launcher 是一个约 300 行的 C 程序（纯 C11，直接使用 Landlock UAPI——除静态链接的 musl 外无其他库，因此审计面仅为该文件加内核的稳定 syscall 契约）：`--ro <path>` / `--rw <path>` 授权，`--`，被包装的 argv；它为自身安装规则集并执行 `exec`（规则集跨 `execve` 继承，且它在限制前设置 `no_new_privs`）；`--probe` 在一个短生命周期子进程中强制最大规则集，仅当内核确实强制时才以 0 退出；所有 launcher 失败都会以 125 退出且不运行子进程，并打印一行致命的 `landlock-run:` 诊断。成功完成 exec 的子进程也可能返回 125，因此仅凭退出状态不能作为 launcher 失败的证据。较旧的 ABI 会在执行子进程之前打印精确的 `landlock-run: partial enforcement (older Landlock ABI)` 通知，因此该行不是致命证据。
 
-Landlock launcher 源码和包工作区位于 `native/landlock-run`，与 harness 消费方同仓。独立的 [`node-addon-landlock-run`](https://github.com/deepseek-harness/node-addon-landlock-run) 仓库是用于打包并发布 npm 包族的发布镜像；导出流程归 `native/README.md` 所有。平台二进制由 npm 选择，入口包拥有路径解析、探测、CLI 参数、致命前缀和部分强制执行通知，而 harness 将沙箱模式映射为授权。将入口点与其二进制一起版本化，使探测解析和启动语法保持对齐。
+Landlock launcher 源码和包家族位于 `native/landlock-run`，与 harness 消费方同仓，并属于根 pnpm workspace。[仓库内 Landlock 发布决策](../process/2026-08-06-in-repository-landlock-release.md)负责共享锁文件、原生构建、打包演练和 npm 发布边界。平台二进制由 npm 选择，入口包拥有路径解析、探测、CLI 参数、致命前缀和部分强制执行通知，而 harness 将沙箱模式映射为授权。将入口点与其二进制一起版本化，使探测解析和启动语法保持对齐。
 
 后端 profile 共享模式契约但在必要的主机授权上有所不同。Landlock 和 Seatbelt 在 read-only 模式下仅允许 `/dev/null`；workspace-write 还允许各自所需的主机临时目录根。每次包装携带后端特定的拒绝签名。Landlock 在较旧的 ABI 无法管控所有操作时报告 partial enforcement，而成功的 bwrap 和 Seatbelt profile 报告 full enforcement。
 
@@ -118,7 +118,7 @@ fs/web/todo 在进程内执行，因此它们的沙箱语义是各自 seam 层
 ### 测试
 
 - **单元测试：** 固定平台选择和 profile、直接交接提供方返回的 argv、带有无效 workdir 对照的 spawn 层失败、runner 缺失／不可执行／解释器缺失证据、格式错误 runner 阴性对照、受约束的 `BASH_ENV` 求值顺序、结构化 runner 分类（包括只有部分强制执行通知的子进程结果、带门控的致命证据、子进程退出码 126／127，以及前台／后台一致性）、按调用的模式/根目录解析、按进程事实、升级验证和结果、权限 preset fold 和写入透传，以及运行时上下文排序与具体化。
-- **Keyless 真实 runner：** 在提供方和 bash 消费方层面对 bwrap、Landlock 和 Seatbelt 执行真实文件系统效果测试；一个真实 Cordis 上下文通过已交付的 bash 和 fs 工具并发驱动两个项目会话，证明在自身根目录写入成功、在兄弟根目录写入被拒绝。打包安装测试证明注册表 launcher 保持可执行。CI 拒绝静默全跳过。
+- **Keyless 真实 runner：** 在提供方和 bash 消费方层面对 bwrap、Landlock 和 Seatbelt 执行真实文件系统效果测试；一个真实 Cordis 上下文通过已交付的 bash 和 fs 工具并发驱动两个项目会话，证明在自身根目录写入成功、在兄弟根目录写入被拒绝。打包安装测试会安装当前 checkout 的原生 tarball，并证明 launcher 保持可执行且字节完全一致。CI 拒绝静默全跳过。
 - **With-key：** 以只读模式启动真实 ACP 组合，让模型驱动的 bash 写入命中 runner 的拒绝标记，再通过已授权与被拒绝的 workspace-write 重试驱动 bridge 应答器和磁盘效果；不可用的凭证或 runner 自动跳过。
 - **快照：** 固定原子化的当前策略上下文和两个脚本化的 approval 分支。一个真实 ACP 示例场景把会话放在用户主目录下，同时让部署后备根目录指向 `/tmp`，然后固定 workspace-write 运行时上下文消息与一次成功的、由部署选定的变更；这能区分会话根目录解析与进程后备值，而不依赖 runner 特定的拒绝文本。一个模拟 Landlock 部分强制执行行为的 POSIX 提供方会在组装后的应用中固定直接执行 bash `false` 时仍得到普通子进程结果，并固定提供方可执行文件缺失时在前台／后台均为基础设施失败。其他快照以无约束启动，使无关 fixture（测试前置数据）保持平台无关。
 
@@ -128,15 +128,15 @@ fs/web/todo 在进程内执行，因此它们的沙箱语义是各自 seam 层
 
 - **第二个消费方**——`subagent-acp` 可选地约束子 agent（按调用策略；默认无约束——子 agent 必须写入自己的持久化）。
 - **更多环境**——环境一致的能力组示例（如 bash+fs 对一个容器）。
-- **Windows 链**——`PLATFORM_CHAINS.win32` 保留为空（失败关闭）；填充它意味着来自 AppContainer/restricted-token 家族的约束 runner，从其自己的仓库按 `node-addon-landlock-run` 模板交付，加上其 profile 方言、拒绝签名和 runner 失败规则。改为包装第三方 landstrip runner 的方案[经考虑后已驳回](../../rejected/feature/2026-07-26-evaluate-landstrip-for-windows-sandbox-rung.md)——它所经受的实战检验还不足以承载安全不变式。
+- **Windows 链**——`PLATFORM_CHAINS.win32` 保留为空（失败关闭）；填充它意味着来自 AppContainer/restricted-token 家族的约束 runner，由主仓库在 `native/` 下按 `@deepseek-ai/node-addon-landlock-run` 模板交付，加上其 profile 方言、拒绝签名和 runner 失败规则。改为包装第三方 landstrip runner 的方案[经考虑后已驳回](../../rejected/feature/2026-07-26-evaluate-landstrip-for-windows-sandbox-rung.md)——它所经受的实战检验还不足以承载安全不变式。
 
 ## 曾考虑的替代方案
 
 - **命令字符串启发式预检**：否决。无法理解展开/子进程/符号链接；严格尝试（运行它，让内核决定）是唯一可信的拒绝信号。
 - **即使平台仅有一个后端也功能性探测**：否决。探测用于在候选者之间仲裁；只有一个时无需决策，且探测开销对每个会话的首次约束命令征税（对未来重量级后端而言代价过高）。runner 自身执行时的失败关闭拒绝加结构化 `runnerFailureRules` 分类承载了安全属性。
-- **提交构建好的 launcher 二进制**：否决。diff 中的二进制不可审查且膨胀历史；经审查的源码 + 原生 CI 构建 + launcher 仓库的字节固定发布演练使二进制远离所有代码树。
+- **提交构建好的 launcher 二进制**：否决。diff 中的二进制不可审查且膨胀历史；经审查的源码 + 原生 CI 构建 + 主仓库的字节固定发布演练使二进制远离所有代码树。
 - **安装时编译 launcher**：否决。将 C 工具链强加给每个消费方；仅在碰巧有编译器时才存在的备选不是备选。
-- **从一个构建器交叉编译两种架构**：否决。仅为重建两个约 70 KB 的二进制就需要携带一个固定的交叉工具链（rustup targets、zig 或容器镜像）；每架构的原生 runner 已存在，各自构建自己的平台包（`node-addon-require-builtin` 模式，launcher 仓库自己的流水线）。
+- **从一个构建器交叉编译两种架构**：否决。仅为重建两个约 70 KB 的二进制就需要携带一个固定的交叉工具链（rustup targets、zig 或容器镜像）；每架构的原生 runner 已存在，各自构建自己的平台包（`node-addon-require-builtin` 模式，由主仓库的原生流水线保留）。
 - **无备选（bwrap 或失败关闭）**：否决。将失败集中在沙箱最重要的主机上，最终因放弃而降级到 `danger-full-access`。
 - **将机制保留在 `dsh-bash-sandbox` 内部**：否决。阻塞既有的第二个消费方，使未来阶段从一个 bash 插件的配置中读取模式，且无法表达升级。
 - **提供方上的配置固定模式**：否决。每进程一个模式；无法服务具有不同策略的并发消费方，也无法表达一次性放宽重试。
@@ -175,7 +175,7 @@ fs/web/todo 在进程内执行，因此它们的沙箱语义是各自 seam 层
 - **Seatbelt 层级依赖 Apple 已弃用但仍交付的 `sandbox-exec` CLI。** 作为 darwin 的唯一候选，它无需探测即被选中，因此在 workdir 可用时，未来移除会表现为可归因于 runner 的 spawn 失败，可执行文件拒绝则通过其致命签名体现——两者都会变为 `SANDBOX_UNAVAILABLE`，且命令绝不会运行；失败关闭，绝不开放。
 - **Landlock 约束的完整度取决于运行内核的 ABI。** 报告为 `enforcement: 'partial'` 而非拒绝——这是有意的权衡，使备选在旧内核主机上仍可用。
 - **Runner 归因使用带内协议。** 退出状态与 stderr 无法以密码学方式识别写入者，因此受限子进程可以模仿 runner 的致命诊断行和状态，造成可用性或诊断误归因。多项证据的合取与精确通知排除减少了意外匹配；这不是沙箱绕过，因为子进程已经受到限制。
-- **launcher 作为注册表依赖到达。** 通过其自身仓库的发布流水线（经审查的 C 源码、原生 CI 构建器、字节固定的发布演练）加上本仓库的版本固定获得信任——真实内核 e2e 测试环节会验证安装产物的实际行为。
+- **launcher 在源码中是 workspace 依赖，发布后是 npm 依赖。** 主仓库会在发布同一个包家族之前，一起测试经审查的 C 源码、原生 CI 构建和字节固定的本地 tarball；真实内核 e2e 测试环节会验证这些安装字节的实际行为。
 - **模型可能过度请求。** 在没有拒绝依据的情况下升级，或在 `workspace-write` 足够时选择 `danger-full-access`：描述引导且枚举强制阶梯，但人的提示词是实际门控；`approval/asked` 原因使过度请求可审计，且 `prepend` 策略应答器可以自动拒绝部署永远不想要的模式。
 - **公布的目标集是静态的，而有效模式是按会话的**（schema 是注册表全局的）——已处于最宽模式的会话仍被提供这些字段。构造上无害：执行时的严格放宽检查（而非枚举）是安全边界——非放宽请求以自身文本失败且不提示任何人。
 - **授权的升级不等于可工作的沙箱。** 不可用的后端即使对授权升级到约束模式也仍然失败关闭——平台没有链或所有探测失败时在 `confine()` 阶段失败，所选可执行文件无法启动时通过 spawn 通道失败，已启动的 runner 拒绝时则通过结构化规则失败——而授权的 `danger-full-access` 运行根本不触及提供方：此时授权（而非探测）是权威。
@@ -187,7 +187,7 @@ fs/web/todo 在进程内执行，因此它们的沙箱语义是各自 seam 层
 - **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 `SubprocessService` 同步抛出同样带有来源信息的 `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——今天的 Windows？** `confine()` 抛出失败关闭的 `SANDBOX_UNAVAILABLE`，命令永不 spawn；`win32` 是保留的空链，由测试固定为同样失败关闭，直到 Windows runner 填充它（§ 延迟阶段）。
-- **`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，选择落到注册表安装的 Landlock launcher，结论在提供方生命周期内缓存。
+- **`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？** 不——`SandboxMode` 仅声称文件操作；bwrap profile 刻意不 unshare pid，没有后端声称网络。网络限制是否成为自己的旋钮留在 § seam 中开放。
 - **哪些工具实际在约束下运行？** 通过 `ctx.bash` 的 OS 子进程——bash 工具及传递性的钩子命令——再加上通过沙箱化 `ctx.fs` 提供方运行的文件系统工具（`read`/`write`/`edit`，见[跨工具族 fs 沙箱 Agent Note](2026-07-14-cross-family-fs-sandbox.md)）：bash 通过 OS runner 约束，fs 通过进程内路径围栏约束，二者都以同一个 `ctx.sandboxPolicy` 模式为键。web/todo 仍在进程内且不受限制（web 的唯一效果是网络，不在文件效果模式词汇内）。
 - **授权的升级会持久化吗？** 不会。授权由发起请求的确切前台或后台调用消费；每个相邻调用保留自己的有效模式。后续的后台拒绝通过 `task_output` 呈现，并且可以作为一次新的精确命令重试的依据。

.agents/notes/implemented/feature/2026-08-02-win32-in-process-folder-dialog.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/2026-08-02-win32-in-process-folder-dialog.md
-2026-08-02-win32-in-process-folder-dialog.md: 91a1ed0d7b1c1938a5e038ce36f1ca90bf3c9e82
-2026-08-02-win32-in-process-folder-dialog.zh.md: 6b90dc1c5fa0042b3e2bcbea8ed554f1f0ea2acf
+2026-08-02-win32-in-process-folder-dialog.md: 5389293605169ce5ca269a127de5f609b8b7dd11
+2026-08-02-win32-in-process-folder-dialog.zh.md: ef81bc1f65b2859de1eb60f74f79e4869da4746b

.agents/notes/implemented/feature/2026-08-02-win32-in-process-folder-dialog.md

@@ -14,7 +14,7 @@ The Windows directory picker's primary tier was a spawned PowerShell script arou
 
 ## Alternatives considered
 
-- **A prebuilt native helper (`native/` family like `node-addon-landlock-run`).** Rejected: a mirror repository, an npm package family, MSVC provisioning, and a release handoff — all to ship ~150 lines of C the repository cannot exercise on CI (no real-Windows lane); koffi delivers the same COM surface with zero new supply chain.
+- **A prebuilt native helper (`native/` family like `@deepseek-ai/node-addon-landlock-run`).** Rejected: another npm package family, MSVC provisioning, and a Windows build/release lane — all to ship ~150 lines of C the repository cannot currently exercise on CI (no real-Windows lane); koffi delivers the same COM surface with zero new supply chain.
 - **An N-API in-process addon.** Rejected for the same CI/toolchain reasons plus owned C++ for STA threading and message pumping that a child process + koffi express in TypeScript.
 - **Keep PowerShell primary and probe versions.** Rejected: the picker stays hostage to shell packaging (6 vs 7, Store aliases, profiles), and 5.1's legacy dialog remains the floor wherever pwsh is absent; the fallback-trigger widening alone was accepted into the fallback tier instead.
 - **Blocking the main thread for the modal call.** Rejected outright: the web host must keep serving RPC while the dialog is open.

.agents/notes/implemented/feature/2026-08-02-win32-in-process-folder-dialog.zh.md

@@ -14,7 +14,7 @@ Windows 目录选择器的主层此前是围绕 WinForms `FolderBrowserDialog`
 
 ## 考虑过的替代方案
 
-- **预编译原生助手（`native/` 家族，如 `node-addon-landlock-run`）。** 否决：镜像仓库、npm 包家族、MSVC 供给和发布交接——只为交付约 150 行 CI 无法执行的 C（没有真 Windows 通道）；koffi 以零新增供应链提供同一 COM 面。
+- **预编译原生助手（`native/` 家族，如 `@deepseek-ai/node-addon-landlock-run`）。** 否决：再增加一个 npm 包家族、MSVC 供给和 Windows 构建／发布通道——只为交付约 150 行目前无法在 CI 中执行的 C（现有 CI 没有真 Windows 通道）；koffi 以零新增供应链提供同一 COM 面。
 - **N-API 进程内插件。** 否决：同样的 CI／工具链原因，另加需要自有 C++ 处理 STA 线程与消息泵，而子进程 + koffi 用 TypeScript 就能表达。
 - **保留 PowerShell 为主层并探测版本。** 否决：选择器仍被 shell 打包形态挟持（6 与 7、Store 别名、profile），且没有 pwsh 的机器地板仍是 5.1 的旧版对话框；仅把回退触发条件的拓宽吸收进回退层。
 - **在主线程上阻塞模态调用。** 直接否决：对话框打开期间 web 宿主必须继续服务 RPC。

.agents/notes/implemented/feature/2026-08-06-bundled-dsh-badge-skill.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/2026-08-06-bundled-dsh-badge-skill.md
+2026-08-06-bundled-dsh-badge-skill.md: 512f67ca347ca311a1f80fef932f6af8c91b0fe9
+2026-08-06-bundled-dsh-badge-skill.zh.md: 88fcadf66944cb91419441be3916ae04968be663

.agents/notes/implemented/feature/2026-08-06-bundled-dsh-badge-skill.md

@@ -0,0 +1,25 @@
+# Agent Note: Bundled dsh badge skill
+
+Status: implemented
+
+English | [中文](2026-08-06-bundled-dsh-badge-skill.zh.md)
+
+## Problem
+
+The [Cordis tutorial](../../../../docs/cordis-tutorial/index.md) uses an official “powered by dsh” badge across its pages, but the shipped CLI has no reusable instructions or explicit opt-in provider for applying the same attribution elsewhere.
+
+## Decision
+
+`@deepseek-ai/dsh-skill-badge` is a native Cordis plugin that registers one immutable bundled provider on `ctx.skills`. The provider owns the `dsh-badge` summary, instruction body, and PNG resource base; `dsh-tool-skill` remains the sole owner of model-facing catalog and loader rendering.
+
+The shipped CLI composition declares `skill-badge` as disabled. Enabling that existing row is the explicit opt-in; disabled installations advertise no badge skill and gain no model-visible content.
+
+The provider uses the bundled rank after project, custom, and user filesystem sources, so a user-owned `dsh-badge` definition can override it through the ordinary registry precedence contract. Provider disposal removes the contribution through the registry-owned effect.
+
+## Alternatives considered
+
+**Mount packaged files through `dsh-skill-local`.** Rejected because filesystem discovery, parsing, and watching add lifecycle machinery that an immutable single-skill provider does not need.
+
+## Consequences
+
+The badge instructions and source PNG are versioned with DSH and resolve through a packaged directory resource base. The provider has no configuration surface. Package tests pin provider lifecycle and the official PNG bytes, while a keyless assembled-application snapshot pins the enabled catalog and loaded skill body.

.agents/notes/implemented/feature/2026-08-06-bundled-dsh-badge-skill.zh.md

@@ -0,0 +1,25 @@
+# Agent Note: 内置 dsh 徽章 skill
+
+Status: implemented
+
+[English](2026-08-06-bundled-dsh-badge-skill.md) | 中文
+
+## 问题
+
+[Cordis 教程](../../../../docs/cordis-tutorial/index.md)的各个页面都使用官方「powered by dsh」徽章，但交付的 CLI（命令行界面）既没有用于在其他位置应用同样署名的可复用指令，也没有可显式选择加入的提供方。
+
+## 决策
+
+`@deepseek-ai/dsh-skill-badge` 是一个原生 Cordis 插件，会在 `ctx.skills` 上注册一个不可变的内置提供方。该提供方负责 `dsh-badge` 的摘要、指令正文和 PNG 资源基底；`dsh-tool-skill` 仍是面向模型的目录与 loader 渲染的唯一归属方。
+
+交付的 CLI 组合将 `skill-badge` 声明为禁用。启用这个现有配置行就是显式选择加入；禁用它的安装实例不会公开任何徽章 skill（技能），也不会获得任何模型可见内容。
+
+该提供方使用排在项目、自定义及用户文件系统来源之后的内置 rank，因此用户自有的 `dsh-badge` 定义可通过注册表的常规优先级契约覆盖它。提供方释放时，注册表拥有的 effect 会移除该贡献。
+
+## 曾考虑的替代方案
+
+**通过 `dsh-skill-local` 挂载随包文件。** 否决，因为文件系统发现、解析和监视会引入生命周期机制，而不可变的单一 skill 提供方并不需要这些机制。
+
+## 后果
+
+徽章指令和源 PNG 随 DSH 一同纳入版本管理，并通过以随包目录为基础的资源基底解析。该提供方没有配置面。包测试固定提供方生命周期和官方 PNG 的字节内容；无密钥的组装应用快照则固定启用后的目录和已加载的 skill 正文。

.agents/notes/implemented/feature/2026-08-08-dsh-run-headless-command.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/2026-08-08-dsh-run-headless-command.md
+2026-08-08-dsh-run-headless-command.md: d1cc0573d1d94bb7b30e98aa685d2b7d89878a99
+2026-08-08-dsh-run-headless-command.zh.md: 7b48dfd708efb197babcfcbeaf4f8a4ca76cde3f

.agents/notes/implemented/feature/2026-08-08-dsh-run-headless-command.md

@@ -0,0 +1,41 @@
+# Agent Note: `dsh run` owns one-shot headless execution
+
+Status: implemented
+
+English | [中文](2026-08-08-dsh-run-headless-command.zh.md)
+
+## Problem
+
+The product launcher attached optional task text to its generic profile boot: `dsh --profile headless "task"`. That made one argv shape mean either a long-lived profile or a one-shot run according to a row discovered only after composition. The parser's `ProfileInvocation` carried optional task state, help presented a profile implementation detail as the user command, and a custom profile could accept a task only through the same overloaded root.
+
+The former `dsh -p` spelling was already absent from the parser, so restoring it or detecting it specially would add compatibility machinery to a pre-release interface. A separate application-file proposal also used the `run` verb, leaving two incompatible owners for one top-level command.
+
+## Decision
+
+One-shot execution owns an explicit grammar:
+
+```text
+dsh run [--profile <name>] [--patch <path>...] <task...>
+```
+
+`--profile` defaults to `headless` and remains available for custom one-shot compositions. `--patch` is repeatable and occupies the existing overlay layer. Commander joins the variadic task arguments with spaces and rejects a missing or blank task before boot.
+
+The [profile plugin bundle decision](../architecture/2026-08-05-profile-plugin-bundles.md) owns the composition selected by this grammar.
+
+`RunInvocation` is a separate `DshInvocation` member. The generic profile invocation no longer carries task text, and its root command accepts no positional arguments. Both dispatch paths call the existing deep `runProfile` module: `profile` omits `task`, while `run` supplies it. There is no shallow `run.ts` forwarding module and no alias, warning, or custom detector for former spellings; they fail through the ordinary Commander grammar. A one-shot profile without `headless-runner` still fails through the existing composed-row check, while booting a profile that contains that row without a task points to `dsh run --profile <name> "<task>"`.
+
+The `run` verb belongs to one-shot task execution. Launching an application file must choose another command name; two top-level meanings selected by positional shape would recreate the ambiguity this command removes.
+
+The runner's user-visible contract stays the same: a fresh persisted session, browser observation URL on stderr, final assistant text on stdout, completed/non-completed exit mapping, and bounded signal shutdown. The product-level keyless acceptance exposed that the in-process mux consumer could lag the same-process `agent/status: idle` notification and derive output before reading the final frames. The idle notification now captures the authoritative final session sequence, and the runner waits until the ordered mux reaches that boundary (or the stream ends) before deriving text and exit reason. This enforces the existing idle-to-idle contract without adding a wire field or a timing delay.
+
+## Alternatives considered
+
+- **Keep task text on `dsh --profile`.** Rejected because profile boot and one-shot execution remain one grammar whose meaning depends on a late composition check.
+- **Preserve `dsh -p` or the positional profile form as aliases.** Rejected under the pre-release stance: compatibility branches would outlive the interface they were meant to retire.
+- **Make `--profile headless` mandatory under `run`.** Rejected because the shipped one-shot surface should have the shortest canonical spelling, while optional `--profile` preserves plugin-defined one-shot compositions.
+- **Give `dsh run` to application-file launch and choose another headless verb.** Rejected because `run` describes executing a task through the harness; application-file ownership would make the product's primary one-shot command less direct and collide with custom one-shot profiles.
+- **Add `apps/cli/src/run.ts`.** Rejected because it would only forward to `runProfile`, splitting command ownership without hiding any complexity.
+
+## Consequences
+
+This is an intentional breaking CLI change. Documentation, help, parser tests, built-bin acceptance, PTY shutdown coverage, and the assembled keyless snapshot use `dsh run`. Existing custom one-shot profiles keep working through `--profile`; long-lived profiles and config dumps retain their existing root grammar. The competing application-file command must be renamed and rebased separately rather than sharing or overloading `run`.

.agents/notes/implemented/feature/2026-08-08-dsh-run-headless-command.zh.md

@@ -0,0 +1,41 @@
+# Agent Note: `dsh run` 负责一次性 headless 执行
+
+Status: implemented
+
+[English](2026-08-08-dsh-run-headless-command.md) | 中文
+
+## 问题
+
+产品启动器过去把可选任务文本挂在通用 profile 启动命令上：`dsh --profile headless "task"`。于是，同一种 argv 形态会表示常驻 profile 或一次性运行，具体含义取决于组合完成后才发现的配置行。解析器的 `ProfileInvocation` 携带可选任务状态，帮助信息把 profile 的实现细节呈现为用户命令，自定义 profile 也只能通过同一个过载的根命令接收任务。
+
+解析器中已经没有原来的 `dsh -p` 写法，因此恢复该写法或加入特殊检测，会给预发布接口增加兼容机制。另一个应用文件提案也使用 `run` 动词，使同一个顶层命令同时归属两个互不兼容的功能。
+
+## 决策
+
+一次性执行采用明确语法：
+
+```text
+dsh run [--profile <name>] [--patch <path>...] <task...>
+```
+
+`--profile` 默认为 `headless`，同时保留对自定义一次性组合的支持。`--patch` 可重复使用，并沿用既有 overlay 层的位置。Commander 用空格拼接可变数量的任务参数，并在启动前拒绝缺失或空白任务。
+
+[profile 插件组合包决策](../architecture/2026-08-05-profile-plugin-bundles.md)负责该语法所选择的组合。
+
+`RunInvocation` 是单独的 `DshInvocation` 成员。通用 profile 调用不再携带任务文本，其根命令也不接受位置参数。两条分派路径都调用已有的深层 `runProfile` 模块：`profile` 省略 `task`，`run` 则提供该字段。实现中没有只负责转发的浅层 `run.ts` 模块，也没有面向旧写法的别名、警告或自定义检测器；旧写法会按普通 Commander 语法失败。缺少 `headless-runner` 的一次性 profile 仍会触发既有的组合行检查；如果启动的 profile 包含该行却未提供任务，错误会指向 `dsh run --profile <name> "<task>"`。
+
+`run` 动词只负责一次性任务执行。应用文件启动必须选择其他命令名；如果让两个顶层含义由位置参数形态决定，就会重新引入本命令消除的歧义。
+
+运行器面向用户的契约保持不变：创建新的持久化会话，在 stderr 打印浏览器观察 URL，在 stdout 打印最终 assistant 文本，将完成／未完成映射为退出状态，并执行有界的信号关闭。产品级无密钥验收用例发现，进程内 mux 消费方可能落后于同进程的 `agent/status: idle` 通知，在读到最终帧之前就生成输出。idle 通知现在会捕获权威的会话最终事件序号，运行器则等待有序 mux 到达该边界（或流结束），再生成文本和退出原因。这一机制在不增加 wire 字段或定时延迟的前提下，落实了既有的 idle-to-idle 契约。
+
+## 考虑过的替代方案
+
+- **把任务文本保留在 `dsh --profile` 上。** 不予采纳：profile 启动和一次性执行仍共用同一套语法，其含义取决于较晚发生的组合检查。
+- **保留 `dsh -p` 或位置参数 profile 形式作为别名。** 不予采纳：根据预发布立场，这些兼容分支会比本应退役的接口存续更久。
+- **要求在 `run` 下必须指定 `--profile headless`。** 不予采纳：已交付的一次性接口应采用最短的规范写法，同时用可选的 `--profile` 保留插件定义的一次性组合。
+- **把 `dsh run` 交给应用文件启动，并为 headless 选择另一个动词。** 不予采纳：`run` 描述的是通过 harness 执行任务；若归应用文件所有，产品的主要一次性命令会更不直接，并与自定义一次性 profile 冲突。
+- **新增 `apps/cli/src/run.ts`。** 不予采纳：它只会转发到 `runProfile`，拆分命令归属，却没有隐藏任何复杂度。
+
+## 后果
+
+这是一次有意为之的 CLI（命令行界面）破坏性变更。文档、帮助信息、解析器测试、构建后二进制验收、PTY 关闭覆盖和组装应用的无密钥快照都使用 `dsh run`。现有自定义一次性 profile 可继续通过 `--profile` 工作；常驻 profile 和配置 dump 保留既有的根命令语法。与之竞争的应用文件命令必须单独改名并 rebase，不得共享或重载 `run`。