Merge branch 'code-mode-ui/shiki' into code-mode-ui/trajectory-spans

.agents/notes/implemented/feature/2026-07-26-code-dispatch-log-spill.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
-2026-07-26-code-dispatch-log-spill.md: 2668c195a43ae1f6011c09413338a23caf75401e
-2026-07-26-code-dispatch-log-spill.zh.md: e084ae80d7fed864c7f296b1fd6db713acf7a2b0
+2026-07-26-code-dispatch-log-spill.md: 65af7808c493867cb13042a4f169ffdf05eb4538
+2026-07-26-code-dispatch-log-spill.zh.md: e1293e62f9de9860300428c5c0d25c5404dc76f9

.agents/notes/implemented/feature/2026-07-26-code-dispatch-log-spill.md

@@ -14,7 +14,7 @@ Since the full-content dispatch logging landed, a `run_code` program that reads
 
 **A log-shaping waterfall on the registry, and the spill policy as its first listener.**
 
-- **Seam**: `tools/code-dispatch-log` — a scope-filtered waterfall the bridge runs (via `registry.shapeDispatchLog`, contained: a throwing listener falls back to the unshaped content) over each settled sub-dispatch before appending `tool/code-dispatch`. The payload (`CodeDispatchLog`) carries the outer execution, the hoisted `agent` routing key, the sub-call identity, and the default content. Only the durable copy is shapeable — the program already received the complete value across the worker boundary, and the model sees neither.
+- **Seam**: `tools/code-dispatch-log` — a scope-filtered waterfall the bridge runs (via `registry.shapeDispatchLog`, contained: a throwing listener falls back to the unshaped content, with total error formatting so a hostile thrown value cannot escape the containment) over each settled sub-dispatch before appending `tool/code-dispatch`. The payload (`CodeDispatchLog`) carries the outer execution, the hoisted `agent` routing key, the sub-call identity, and the default content — the RENDERED result projection a native `tool/result` would carry (the program itself received the structured `value`). Only the durable copy is shapeable; the model sees neither. Shaping runs OFF the program path as tracked side work, but bounded: past `maxParallelSubCalls` pending log tasks the ordered commit lane holds, so a slow spill backend backpressures the run instead of accumulating unbounded pending I/O; run settlement still drains every task inside the open turn.
 - **Policy**: `dsh-spill-policy` registers a second arm on the new seam sharing the exact replacement pipeline of its model-facing arm (same `maxInlineBytes` cap, same preview + locator + within-cap invariant, same best-effort fallbacks), with the artifact labeled `dispatch` under the sub-call id. UIs and replay read the full text through the spill artifact exactly as they do for spilled native results, so the native-parity rendering story survives bounding.
 - **One deliberate asymmetry**: the model-facing arm skips `read` (the `read → spill → read again` loop); the dispatch-log arm bounds `read` sub-calls too — a log copy is not model context, so the loop cannot happen, and `read` is precisely the tool that produces huge logs.
 

.agents/notes/implemented/feature/2026-07-26-code-dispatch-log-spill.zh.md

@@ -14,7 +14,7 @@ Status: implemented
 
 **在注册表上增设一个日志整形 waterfall（瀑布式事件），spill 策略作为其第一个监听器。**
 
-- **Seam**：`tools/code-dispatch-log`，一个按作用域过滤的 waterfall，由桥接层在追加 `tool/code-dispatch` 之前对每个已结算的子分发运行（经由 `registry.shapeDispatchLog`，且故障被兜住：监听器抛出异常时回退到未整形的内容）。载荷（`CodeDispatchLog`）携带外层执行、提升出来的 `agent` 路由键、子调用标识与默认内容。可整形的只有持久副本：程序已经跨 worker 边界收到了完整的值，而模型两者都看不到。
+- **Seam**：`tools/code-dispatch-log`，一个按作用域过滤的 waterfall，由桥接层在追加 `tool/code-dispatch` 之前对每个已结算的子分发运行（经由 `registry.shapeDispatchLog`，且故障被兜住：监听器抛出异常时回退到未整形的内容，并用全防御的错误格式化确保恶意抛出值无法逃出兜底）。载荷（`CodeDispatchLog`）携带外层执行、提升出来的 `agent` 路由键、子调用标识与默认内容——即原生 `tool/result` 所载的渲染后结果投影（程序本身收到的是结构化 `value`）。可整形的只有持久副本；模型两者都看不到。整形作为被跟踪的旁路工作在程序路径之外运行，但有界：待处理日志任务超过 `maxParallelSubCalls` 时有序提交车道会暂停，因此慢速 spill 后端会对整个 run 施加背压，而不是无限累积待完成 I/O；run 结算仍会在开放轮次内排空全部任务。
 - **策略**：`dsh-spill-policy` 在新 seam 上注册第二个分支，与其面向模型的分支共用一模一样的替换流水线（同样的 `maxInlineBytes` 上限、同样的预览 + 定位符 + 不超上限不变式、同样的尽力而为回退），产物以 `dispatch` 为标签，记在子调用 id 名下。UI 与回放通过 spill 产物读取全文，方式与读取被 spill 的原生结果完全相同，因此与原生同等保真的渲染在施加边界之后依然成立。
 - **一处有意的不对称**：面向模型的分支跳过 `read`（避免 `read → spill → read again` 循环）；分发日志分支则连 `read` 子调用也施加边界：日志副本不是模型上下文，该循环因此不可能发生，而 `read` 恰恰是会产生巨大日志的那个工具。
 

.agents/notes/implemented/feature/2026-07-26-code-mode-live-parallel-dispatch.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
-2026-07-26-code-mode-live-parallel-dispatch.md: f0d13456d63779fb89b9af4cb09bc90c37356a21
-2026-07-26-code-mode-live-parallel-dispatch.zh.md: 5554ab0456f13a2bbc6d5b18e515930f954c6682
+2026-07-26-code-mode-live-parallel-dispatch.md: b4afc21be902d8ed3e5bee2ad1a540413a864f25
+2026-07-26-code-mode-live-parallel-dispatch.zh.md: 409e4cbf9ea3d1b4d1bbe0cd86b494429ebb8a3d

.agents/notes/implemented/feature/2026-07-26-code-mode-live-parallel-dispatch.md

@@ -15,7 +15,7 @@ Two gaps remained after the first two PRs. Sub-call rows appeared only when each
 **One lifecycle pair, one scheduling contract, shared with native.**
 
 - **Event pair**: `tool/code-dispatch-start` (parent/sub ids, name, normalized args) is appended when the scheduler actually starts a call — not at submission, so a queued call abandoned by run settlement logs nothing. The existing `tool/code-dispatch` settles the pair (same `subCallId`); every started call settles exactly once (aborts settle as `isError` outcomes through the pipeline). Timing = the two events' `time` fields. Both stay log-only; model context is untouched; format stays v0.
-- **Bridge scheduler**: submitted calls are classified at submission via `registry.executionMode` (the SAME fail-closed `isConcurrencySafe` contract the loop uses) and start strictly in submission order. Consecutive parallel-classified calls overlap up to `maxParallelSubCalls` (a validated registry `Config` field, default 10 — the loop scheduler's own default; `1` restores serial dispatch); an exclusive call drains the pool, runs alone, and bars later calls. This is the loop's group semantics adapted to calls that arrive over time instead of in one parsed batch. Run settlement aborts in-flight dispatches and abandons queued-unstarted ones (binding rejection, no events), then drains to quiescence before the outer result closes the turn.
+- **Bridge scheduler**: submitted calls are classified at start time via `registry.executionMode` (the SAME fail-closed `isConcurrencySafe` contract the loop uses) and start strictly in submission order. One single-lane driver owns every ORDERED stage — the start append, `prepare` (pre-execute/guards), the head-of-line `finalize`/`finish` commit (post-execute + context deferral + settle append) — so ordered policy stages never overlap each other and only the around-dispatch/body stage runs concurrently, exactly the native loop's sequencing (`fillPool` awaits `startCall` then `commitReady`). Consecutive parallel-classified calls overlap up to `maxParallelSubCalls` (a `Config` field validated by the Loader schema AND re-validated at direct construction, default 10 — the loop scheduler's own default; `1` restores serial dispatch); an exclusive call drains the pool, runs alone, and holds its barrier until its COMMIT completes (post-execute included), like a native exclusive group. Run settlement aborts in-flight dispatches and abandons queued-unstarted ones (binding rejection, no events), then drains to quiescence — including a commit already mid-flight when the program returned — before the outer result closes the turn.
 - **Client**: `CodeSubCall` widens to `RunningToolCall | ToolResultNode` — a start event lands the running shape in the dispatch index (rows derive the running ring from the shape, exactly as for native in-flight calls), and its settle replaces the entry in place, preserving start order under parallel completion and carrying the start's `time` as `callTime` (duration source). A settle with no observed start (window cut mid-pair, or a pre-start-event log) appends directly, so old logs keep rendering.
 - **SDK prompt**: the model-facing "calls execute sequentially" sentence is replaced with the true contract (independent safe calls may overlap under `Promise.all`; dependent work sequences with `await`) — a model-visible change, re-recorded across every code-mode snapshot.
 

.agents/notes/implemented/feature/2026-07-26-code-mode-live-parallel-dispatch.zh.md

@@ -15,7 +15,7 @@ Status: implemented
 **一对生命周期事件，一份调度契约，与原生共用。**
 
 - **事件对**：`tool/code-dispatch-start`（父/子 id、名称、规范化参数）在调度器真正启动某个调用时才追加，而非在提交时，因此因 run 结算而被放弃的排队调用不会留下任何日志。既有的 `tool/code-dispatch` 结算该事件对（`subCallId` 相同）；每个已启动的调用恰好结算一次（中止也会作为 `isError` 结果经由流水线结算）。计时即这两个事件的 `time` 字段。两个事件都保持仅日志；模型上下文不受影响；格式保持 v0。
-- **桥接层调度器**：已提交的调用在提交那一刻就经 `registry.executionMode` 分类（与 loop 所用完全相同的 fail-closed `isConcurrencySafe` 契约），并严格按提交顺序启动。连续被分类为可并行的调用可以重叠执行，上限为 `maxParallelSubCalls`（经校验的注册表 `Config` 字段，默认值 10，即 loop 调度器自身的默认值；设为 `1` 即恢复串行分发）；独占调用则先排空池、独自运行，并阻挡其后的调用。这是把 loop 的分组语义适配到另一种场景：调用随时间陆续到达，而非作为单个已解析的批次一次性到达。run 结算时会中止仍在运行的分发，并放弃已排队未启动的分发（绑定调用被拒绝，不产生事件），随后排空到完全停稳，之后外层结果才结束该轮次。
+- **桥接层调度器**：已提交的调用在启动那一刻经 `registry.executionMode` 分类（与 loop 所用完全相同的 fail-closed `isConcurrencySafe` 契约），并严格按提交顺序启动。所有有序阶段——start 事件追加、`prepare`（pre-execute/守卫）、队首 `finalize`/`finish` 提交（post-execute + 上下文延迟提交 + settle 事件追加）——由单一驱动车道独占执行，因此有序策略阶段彼此绝不重叠，只有 around-dispatch/工具体阶段并发运行，与原生 loop 的时序完全一致（`fillPool` 先 await `startCall` 再 `commitReady`）。连续被分类为可并行的调用可以重叠执行，上限为 `maxParallelSubCalls`（`Config` 字段，Loader schema 校验之外直接构造时也重新校验，默认值 10，即 loop 调度器自身的默认值；设为 `1` 即恢复串行分发）；独占调用则先排空池、独自运行，且其屏障保持到自身提交（含 post-execute）完成为止，与原生独占分组一致。run 结算时会中止仍在运行的分发，并放弃已排队未启动的分发（绑定调用被拒绝，不产生事件），随后排空到完全停稳——包括程序返回时已在途的提交——之后外层结果才结束该轮次。
 - **client 侧**：`CodeSubCall` 拓宽为 `RunningToolCall | ToolResultNode`：start 事件把运行中形状写入分发索引（行组件从该形状推导出运行指示环，与原生运行中的调用处理完全一致），其结算事件则原位替换该条目，即使并行完成也保持启动顺序不变，并把 start 事件的 `time` 作为 `callTime`（时长来源）带入。未观察到对应 start 的结算事件（窗口切在事件对中间，或日志录制于 start 事件引入之前）会直接追加，因此旧日志仍能照常渲染。
 - **SDK 提示词**：面向模型的「调用按顺序执行」一句替换为真实契约（相互独立的安全调用可以在 `Promise.all` 下重叠执行；相互依赖的工作以 `await` 顺序衔接）；这是模型可见的变更，每一份 code-mode 快照都已重新录制。
 

.agents/notes/implemented/process/2026-07-26-web-syntax-highlighting-shiki.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
-2026-07-26-web-syntax-highlighting-shiki.md: 79ad2153b8883fda92205dada300fd194834129b
-2026-07-26-web-syntax-highlighting-shiki.zh.md: 4cb3f0ceadebc4837108463c149262bf8e36f93d
+2026-07-26-web-syntax-highlighting-shiki.md: b329e35f1d0ce7b3de454758403a09f67056b5af
+2026-07-26-web-syntax-highlighting-shiki.zh.md: 8e9d1f0d0c38ce64bcb5da1262538da762f70b12

.agents/notes/implemented/process/2026-07-26-web-syntax-highlighting-shiki.md

@@ -15,7 +15,7 @@ The client rendered every code surface — markdown fences in assistant prose, t
 **Shiki in its synchronous fine-grained form, as one `ui-primitives` singleton, themed exclusively through CSS custom properties.**
 
 - **Dependency**: `shiki/core` + `@shikijs/langs`, composed via `createHighlighterCoreSync` with `createJavaScriptRegexEngine({ forgiving: true })` — no oniguruma WASM, no async init, bundle-friendly. Grammar allowlist: `typescript` (embeds JS), `shellscript`, `json` — the languages the harness actually renders; everything else falls back to a geometry-identical plain block, never an error. Prior art: the VitePress site already renders all documentation code through shiki, and TextMate grammars materially beat regex highlighters on TypeScript — the payload that matters here.
-- **Singleton**: `ui-primitives/src/markdown/highlight.ts` lazily creates one `HighlighterCore` per document and exposes `highlightToHtml(code, lang)` (undefined = render plain). The shared `CodeBlock` component owns both arms; its shiki arm injects the generated span tree via `dangerouslySetInnerHTML` — sanctioned because shiki emits a static span tree computed from the code text (no user HTML passes through, no scripts/handlers), shiki's own documented consumption path.
+- **Singleton**: `ui-primitives/src/markdown/highlight.ts` creates one `HighlighterCore` per document and exposes `highlightToHtml(code, lang)` (undefined = render plain). Engine + grammar construction is a ~120-175ms long task, so the module pre-warms the singleton in a deferred task at plugin boot (the lazy path stays as the correctness fallback), keeping the cost off the render path where a stream's finalize swap would jank. The alias table is a `Map`, not an object: fence info strings are assistant-authored, so a label like `constructor` must miss instead of resolving an inherited property and crashing shiki. The shared `CodeBlock` component owns both arms; its shiki arm injects the generated span tree via `dangerouslySetInnerHTML` — sanctioned because shiki emits a static span tree computed from the code text (no user HTML passes through, no scripts/handlers), shiki's own documented consumption path.
 - **Theming**: shiki's `createCssVariablesTheme` routes every token color through `--shiki-*` custom properties; the VALUES live in a new `ui-theme/styles/shiki.css` token sheet (light on `:root`, dark on `body[data-ds-dark-theme]` — the same cascade as every other sheet), imported by the shell's `base.css` chain. Component CSS stays tokens-only; no literal color ever enters JS or component sheets. Background/foreground alias the existing markdown code-block tokens so highlighted and plain blocks agree.
 - **Surfaces**: markdown fences (`MarkdownText`'s `pre` component routes single-string fences through `CodeBlock`), the `run_code` expanded program body (ToolRow's code variant, `lang="typescript"`), and the details panel's Input args (`lang="json"`). Output stays plain deliberately — tool output is arbitrary text, and guessing a grammar would mis-highlight more than it helps.
 

.agents/notes/implemented/process/2026-07-26-web-syntax-highlighting-shiki.zh.md

@@ -15,7 +15,7 @@ client 过去把每一处代码表面——assistant 正文里的 markdown 围
 **采用同步细粒度形态的 shiki，作为 `ui-primitives` 里的一个单例，主题化完全经由 CSS 自定义属性完成。**
 
 - **依赖**：`shiki/core` + `@shikijs/langs`，经 `createHighlighterCoreSync` 搭配 `createJavaScriptRegexEngine({ forgiving: true })` 组装——不带 oniguruma WASM、没有异步初始化、对 bundle 友好。语法（grammar）白名单：`typescript`（内嵌 JS）、`shellscript`、`json`——即 harness 实际会渲染的那几种语言；其余一律回退到几何完全一致的纯文本块，绝不报错。先例：VitePress 站点已经通过 shiki 渲染全部文档代码；而在 TypeScript（正是此处要紧的载荷）上，TextMate 语法实质性优于正则高亮器。
-- **单例**：`ui-primitives/src/markdown/highlight.ts` 为每个 document 惰性创建一个 `HighlighterCore`，并公开 `highlightToHtml(code, lang)`（undefined 即渲染为纯文本）。共享的 `CodeBlock` 组件同时拥有两条分支；其 shiki 分支经 `dangerouslySetInnerHTML` 注入生成的 span 树——此用法获准，因为 shiki 输出的是从代码文本计算出的静态 span 树（不流经任何用户 HTML，没有脚本或事件处理器），这正是 shiki 自身文档载明的消费路径。
+- **单例**：`ui-primitives/src/markdown/highlight.ts` 为每个 document 创建一个 `HighlighterCore`，并公开 `highlightToHtml(code, lang)`（undefined 即渲染为纯文本）。引擎加语法的构建是一次约 120-175ms 的长任务，因此模块在插件启动时用延迟任务预热单例（惰性路径保留为正确性兜底），把这笔开销挪出渲染路径——否则流式 finalize 交换的那一刻会卡顿。别名表用 `Map` 而非对象：fence 信息串由 assistant 撰写，诸如 `constructor` 这样的标签必须落空，而不是解析到继承属性并让 shiki 崩溃。共享的 `CodeBlock` 组件同时拥有两条分支；其 shiki 分支经 `dangerouslySetInnerHTML` 注入生成的 span 树——此用法获准，因为 shiki 输出的是从代码文本计算出的静态 span 树（不流经任何用户 HTML，没有脚本或事件处理器），这正是 shiki 自身文档载明的消费路径。
 - **主题化**：shiki 的 `createCssVariablesTheme` 让每一种 token 颜色都经由 `--shiki-*` 自定义属性路由；取值本身住在新增的 `ui-theme/styles/shiki.css` token 表里（亮色在 `:root`、暗色在 `body[data-ds-dark-theme]`——层叠方式与其余每张样式表相同），由壳的 `base.css` 导入链引入。组件 CSS 保持只用 token；任何字面颜色都不进入 JS 或组件样式表。背景/前景以别名指向既有的 markdown 代码块 token，使高亮块与纯文本块彼此一致。
 - **表面**：markdown 围栏代码块（`MarkdownText` 的 `pre` 组件把单字符串围栏路由到 `CodeBlock`）、`run_code` 展开后的程序正文（ToolRow 的 code 变体，`lang="typescript"`），以及 details 面板的 Input 参数（`lang="json"`）。输出有意保持纯文本——工具输出是任意文本，硬猜一种语法，带来的误高亮会多于帮助。