Merge origin/master into fix/landlock-runner-failure-classification

.agents/notes/archived/bug-fix/2026-07-27-tui-diff-card-redundant-path-header.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/2026-07-27-tui-diff-card-redundant-path-header.md
-2026-07-27-tui-diff-card-redundant-path-header.md: 708e543ff079828b4929d2a50ac697a9c846608a
-2026-07-27-tui-diff-card-redundant-path-header.zh.md: 863868ae707f37689bbc202267c5470d8c3163e9
+2026-07-27-tui-diff-card-redundant-path-header.md: 608a11892a20d020087180175eff847021dc0554
+2026-07-27-tui-diff-card-redundant-path-header.zh.md: bf7f1c1eeb994f9940b5f7dfb7db72d422293bd4

.agents/notes/archived/bug-fix/2026-07-27-tui-diff-card-redundant-path-header.md

@@ -1,6 +1,7 @@
 # Agent Note: TUI diff card dropped the duplicated file path
 
 Status: implemented
+Archived: 2026-07-31
 
 English | [中文](2026-07-27-tui-diff-card-redundant-path-header.zh.md)
 

.agents/notes/archived/bug-fix/2026-07-27-tui-diff-card-redundant-path-header.zh.md

@@ -1,6 +1,7 @@
 # Agent Note: TUI diff 卡片重复打印文件路径
 
 Status: implemented
+Archived: 2026-07-31
 
 [English](2026-07-27-tui-diff-card-redundant-path-header.md) | 中文
 

.agents/notes/archived/manifest.json

@@ -46,6 +46,9 @@
     "bug-fix/2026-07-26-intent-draft-same-tick-echo.i18n.yaml": "sha256:c623947c4fa00e6d4b51792c7972ba09582bbcb7605beb373725c0dd666f2c81",
     "bug-fix/2026-07-26-intent-draft-same-tick-echo.md": "sha256:fa8b1417b2cdd3deecbf8e55bdddd73dd3a8c6e3486fd399b0b8bdf317e56373",
     "bug-fix/2026-07-26-intent-draft-same-tick-echo.zh.md": "sha256:00ce72552dbaa11562fbc541343a5d33f9449edabbe6dd354eb879a7d4d530f8",
+    "bug-fix/2026-07-27-tui-diff-card-redundant-path-header.i18n.yaml": "sha256:8613a1cfcf4b9c7fafa78a8d8565e2a65ef0335b7b826af9b2bb32097836af55",
+    "bug-fix/2026-07-27-tui-diff-card-redundant-path-header.md": "sha256:1bd344aec5454d2a2d6e1e6a32eff035c4a99c3df409f2624b39fd32e23ee402",
+    "bug-fix/2026-07-27-tui-diff-card-redundant-path-header.zh.md": "sha256:0a1747006efb1a4b67feceb9b627a437a0f023158e90ae86e1fe8aef76485384",
     "feature/2026-06-14-acp-agent-client-protocol.i18n.yaml": "sha256:006795baa43ae962a8d125cc0f1e9f134bc2ee9fb758b6e7669e3fa0126e1918",
     "feature/2026-06-14-acp-agent-client-protocol.md": "sha256:6828c0af74bb3fb96206ca6b21c0e56a000b50e4744aad4bc2c05092f3a5a31b",
     "feature/2026-06-14-acp-agent-client-protocol.zh.md": "sha256:ba104e841a1fb84edbd3b6c8119d50445b7785255a7a8d13bb9ac8a2cb4d2e69",

.agents/notes/implemented/architecture/2026-07-19-package-owned-invariant-service.i18n.yaml

@@ -1,6 +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
-2026-07-19-package-owned-invariant-service.md: 2443a8f7d04b96f51bb798130078a7457f78b2a1
-2026-07-19-package-owned-invariant-service.zh.md: 3c71d3b7f99a507d4c0236b7ef6dc0794814cdc8
+#   pnpm run verify-translation-pairing --write .agents/notes/implemented/architecture/2026-07-19-package-owned-invariant-service.md
+2026-07-19-package-owned-invariant-service.md: e32efe9f6b3ce6b782c61db56d928e87c160dc9a
+2026-07-19-package-owned-invariant-service.zh.md: 60edaa3f6009acc516017683232ca0c07f64ec0d

.agents/notes/implemented/architecture/2026-07-19-package-owned-invariant-service.md

@@ -8,7 +8,7 @@ English | [中文](2026-07-19-package-owned-invariant-service.zh.md)
 
 Runtime invariant checks span session traces, agent state, scoped dispatch, and request reconstruction. Putting all checks in one diagnostics package makes that package import product vocabularies from unrelated domains, centralizes tests away from their owners, and requires the central package to change whenever a product package adds or removes a check.
 
-Deployments also need more than presence or absence of one plugin. A standard composition should carry the known invariant contributions while permitting a global off switch and package-selective diagnostics. Selection must remain stable when a package loads later or reloads under HMR, and disabled contributions must not allow two plugins to claim the same package name silently.
+Deployments that opt into diagnostics need more than presence or absence of one plugin. Such a composition carries the known invariant contributions while permitting a global off switch and package-selective diagnostics. Selection must remain stable when a package loads later or reloads under HMR, and disabled contributions must not allow two plugins to claim the same package name silently.
 
 Package ownership must also be exhaustive. Without a mechanical repository rule, a new package can omit the companion, dependency, or publication wiring and remain invisible to diagnostics until a maintainer notices the gap.
 
@@ -72,9 +72,9 @@ These four owners supplied the initial stateful checks. The follow-up runtime-co
 
 The generated scoped-event subject resolver lives in `dsh-scope`, beside the contract and invariant that consume it. `gen-scoped-events` uses the root TypeScript Program to enumerate `this: Scoped<Base>` declarations, infer routing-key types from real `scopeTarget(base, key)` calls, and require one unambiguous payload subject or an explicit unsupported marker. The committed runtime map imports no event-owner package, so semantic completeness does not expand either the service or scope package's runtime closure.
 
-### Standard composition and SDK output
+### Example composition and SDK output
 
-The standard agent spine mounts the service and all four stateful companion subpaths, forwarding `enabled`, `package_allowlist`, and `package_blocklist` to the service. Generated SDK Cordis composition emits the same entries. A subpath entry adds its installable root npm package rather than treating the subpath as a package name.
+The example agent spine mounts the service and all four stateful companion subpaths, forwarding `enabled`, `package_allowlist`, and `package_blocklist` to the service. Generated SDK Cordis composition emits the same entries. A subpath entry adds its installable root npm package rather than treating the subpath as a package name. The shipped `dsh` TUI and Web config trees omit the service and companions under the [shipped-config decision](../simplification/2026-08-03-omit-invariants-from-shipped-config.md).
 
 Workspace constraints recognize the separate invariant bundle, and package exports, project references, build configuration, dependency declarations, and the lockfile describe the same publication surface. Generated config catalogs, module graphs, and API documentation derive from those sources.
 
@@ -97,7 +97,7 @@ Every Vitest configuration loads a test host that mounts an explicitly enabled s
 
 - Product packages own and test their relational assertions while the service stays product-independent.
 - Every package pays the publication and dependency cost of a companion; only owners with a meaningful runtime relationship add listener or trace-state cost.
-- Standard compositions can disable all checks or select package names without changing their plugin tree.
+- Compositions that mount the diagnostics can disable all checks or select package names without changing their plugin tree.
 - Explicit companion entries make diagnostic cost and ownership visible in Cordis config and package exports.
 - One selected executable contribution adds one child fiber and its listener/state cost; a selected empty contribution has no listener or trace-state cost, while filtered registrations retain only name ownership.
 - Regex sources are deployment configuration and remain fixed until the service reloads.

.agents/notes/implemented/architecture/2026-07-19-package-owned-invariant-service.zh.md

@@ -8,7 +8,7 @@ Status: implemented
 
 运行时不变式检查跨越会话轨迹、agent 状态、作用域 dispatch 和请求重建。如果所有检查都放在一个诊断包中，该包就必须导入彼此无关的产品领域词汇，测试也会离开真正的所有者；任何产品包新增或移除检查时，都要修改中央包。
 
-部署还需要比“是否加载一个插件”更细的控制。标准组合应携带已知的不变式贡献，同时允许全局关闭或按包选择诊断。包稍后加载或在 HMR 下重载时，选择结果必须保持稳定；被过滤的贡献也不能让两个插件静默占用同一个包名。
+选择启用诊断的部署还需要比“是否加载一个插件”更细的控制。这类组合会携带已知的不变式贡献，同时允许全局关闭或按包选择诊断。包稍后加载或在 HMR 下重载时，选择结果必须保持稳定；被过滤的贡献也不能让两个插件静默占用同一个包名。
 
 包所有权还必须覆盖完整。若没有机械化的仓库规则，新包可能遗漏伴随插件、依赖或发布配置，并一直不会进入诊断范围，直到维护者发现这一缺口。
 
@@ -72,9 +72,9 @@ blocklist 匹配优先于 allowlist 匹配。每个条目都是区分大小写
 
 生成的 scoped event 主体解析表位于 `dsh-scope`，与消费它的契约和不变式相邻。`gen-scoped-events` 使用根 TypeScript Program 枚举 `this: Scoped<Base>` 声明，从真实 `scopeTarget(base, key)` 调用推断路由键类型，并要求唯一、无歧义的 payload 主体或显式 unsupported 标记。提交的运行时映射不导入事件所有者包，因此语义完整性不会扩大服务包或 scope 包的运行时依赖闭包。
 
-### 标准组合与 SDK 输出
+### 示例组合与 SDK 输出
 
-标准 agent spine 会挂载服务和四个有状态伴随子路径，并把 `enabled`、`package_allowlist` 与 `package_blocklist` 转发给服务。生成的 SDK Cordis 组合输出相同条目。子路径条目添加可安装的根 npm 包，而不会把子路径误当成包名。
+示例 agent spine 会挂载服务和四个有状态伴随子路径，并把 `enabled`、`package_allowlist` 与 `package_blocklist` 转发给服务。生成的 SDK Cordis 组合输出相同条目。子路径条目添加可安装的根 npm 包，而不会把子路径误当成包名。根据[交付配置决策](../simplification/2026-08-03-omit-invariants-from-shipped-config.md)，交付的 `dsh` TUI 与 Web 配置树会省略该服务及其伴随插件。
 
 Workspace 约束识别独立的不变式 bundle；包 exports、项目引用、构建配置、依赖声明和 lockfile 描述同一发布表面。生成的配置目录、模块图和 API 文档都从这些源派生。
 
@@ -97,7 +97,7 @@ Workspace 约束识别独立的不变式 bundle；包 exports、项目引用、
 
 - 产品包拥有并测试自己的关系断言，服务保持与产品无关。
 - 每个包都承担 companion 的发布与依赖成本；只有具备有意义运行时关系的所有者才增加 listener 或 trace 状态成本。
-- 标准组合无需改变插件树即可关闭全部检查或按包名选择。
+- 挂载诊断的组合无需改变插件树即可关闭全部检查或按包名选择。
 - 显式伴随条目让诊断成本和所有权在 Cordis 配置与包 export 中可见。
 - 每个选中的可执行贡献增加一个子 fiber 及其 listener/状态成本；选中的空贡献不增加 listener 或 trace 状态成本，被过滤注册则只保留包名占用。
 - 正则表达式源属于部署配置，在服务重载前保持固定。

.agents/notes/implemented/architecture/2026-08-01-packaged-ripgrep-search.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/2026-08-01-packaged-ripgrep-search.md
+2026-08-01-packaged-ripgrep-search.md: 7c515618a18b61bd90177a6fdf19bbd52e564209
+2026-08-01-packaged-ripgrep-search.zh.md: f2b1a12c737f772bff6a6c91c17f7453dbc89748

.agents/notes/implemented/architecture/2026-08-01-packaged-ripgrep-search.md

@@ -0,0 +1,38 @@
+# Agent Note: Packaged ripgrep spawn for glob/grep
+
+Status: implemented
+
+English | [中文](2026-08-01-packaged-ripgrep-search.zh.md)
+
+> Supersedes [bash-backed grep/glob discovery](../../archived/feature/2026-07-09-bash-backed-grep-glob-discovery.md): the v1 decision's explicitly deferred alternative — directly spawning ripgrep — is now what ships.
+
+## Problem
+
+The `glob`/`grep` tools ran through the bash executor seam, which made a system `rg` install a host dependency. On Windows and container images there is no `rg` on `PATH` by default, so the tools silently vanished there; a deployment could only discover that from the load-time probe warning. The bash seam also forced the whole model-visible argument surface through one shell-quoting helper, because a shell sat between the tool and ripgrep — the [bash-backed note](../../archived/feature/2026-07-09-bash-backed-grep-glob-discovery.md) recorded that coupling as the v1 trade-off and named direct spawn as the reasonable follow-up if the shell-string domain ever proved too sensitive. It did: every model value had to survive POSIX single-quoting, the probe had to be scripted in tests, and the executor's own timeout classification duplicated what the cooperative tool-timeout policy already owns.
+
+## Decision
+
+`@deepseek-ai/dsh-tool-fs-search` now runs the PACKAGED ripgrep binary (`@vscode/ripgrep`, an npm dependency whose optional platform packages ship the binary) through the `ctx.subprocess` seam: `runRipgrep()` spawns `rgPath` with a plain argv vector prefixed by `--no-config`, collect-mode stdout/stderr, `graceMs`, and `exec.signal` forwarded. `rgPath` resolves lazily at the first call (memoized per process): `@vscode/ripgrep` resolves its platform package at module evaluation, so a static import would turn a missing or corrupt platform package (`--omit=optional`, partial install) into a Loader-composition failure — the load-time failure mode this change exists to remove. There is no shell layer, so the shell-quoting boundary is gone from execution; the `singleQuote` helper and its shell-spawning tests are deleted with it. The raw streams request the seam's diagnostic-tail collect shape (no spill files — the tool never reads a raw spill path; a lossy stdout read fails as `SEARCH_RAW_OUTPUT_OVERFLOW`). The terminate grace and the stderr tail budget are validated `Config` fields (`graceMs` default 3000, `stderrMaxBytes` default 64 KiB), no longer inherited from bash-local's config. Registration is unconditional — the load-time `command -v rg` probe and the conditional registration decision are deleted, and with them the "rg not found" warning. The package injects `tools`, `systemPrompt`, and `subprocess`.
+
+Exit semantics stay tool-owned: exit 0 is success with results, exit 1 is a successful empty search, anything else classifies into the existing `SEARCH_*` vocabulary (invalid pattern, launch failure, signal kill, raw-output overflow). Timeout is the cooperative tool-call budget attached to the tool definitions: `@deepseek-ai/dsh-timeout-policy` aborts `exec.signal`, the subprocess seam's terminate escalation provides the hard kill, and the tool reports `SEARCH_ABORTED`. The working directory is the session header cwd when present, else `process.cwd()` — there is no executor config to default through anymore, so the tool owns the fallback.
+
+The `fs-glob-sampling` ACP snapshot scenario now executes the real packaged binary against a prepared workspace whose fixed mtimes pin the `--sort=modified` order, replacing the PATH-injected `rg` stand-in (POSIX-only, because the displayed paths carry `/` separators the session-log comparison cannot normalize).
+
+## Alternatives considered
+
+**Keep the bash seam and probe, but document `rg` as a required host dependency.** Rejected: the host dependency is exactly the failure this change removes, and Windows support for the discovery tools was the point of the exercise; a documented requirement is still a requirement.
+
+**Make `rgPath` injectable (a config field or env override) so tests and snapshots keep substituting a stand-in binary.** Rejected: it adds a public deployment surface whose only consumer would be test seams, and the real binary is deterministic enough to pin directly through fixture mtimes — the packaged binary is the deployment, so tests should exercise it.
+
+**Switch to a pure-JS glob/search engine (e.g. `picomatch`/`tinyglobby`).** Rejected: the [dependency-swaps audit](../../rejected/simplification/2026-07-26-dependency-swaps-rejected-by-nih-audit.md) already rejected that on the "no glob engine exists" evidence; ripgrep semantics (`--sort=modified`, VCS pruning, JSON transport, regex dialect) are the tool contract.
+
+## Consequences
+
+- The discovery tools work on every platform the packaged binary covers (darwin/linux/win32, x64/arm64) with no host install; the shipped TUI/Web rosters gain `glob`/`grep` as fixed members ([even-out-shipped-tool-rosters](../feature/2026-07-31-even-out-shipped-tool-rosters.md)).
+- The shell-string attack surface is gone: hostile patterns are inert argv elements, pinned by the integration suite, which now runs on Windows too (it previously self-skipped without a system `rg`).
+- The spawn is unconfined (a plain `ctx.subprocess` call), so `--no-config` is prepended: a host `RIPGREP_CONFIG_PATH` (or an `rg.conf` beside the binary) can otherwise inject a `--pre` preprocessor that executes an arbitrary command for every matched file. With `--no-config`, no config file — and therefore no preprocessor — can reach the search.
+- The raw-output overflow path changed shape: the old bash-backed route inherited bash-local's always-on spill and could leave an unread multi-megabyte temp file; the subprocess seam now collects without spill, and overflow is a pure error (`SEARCH_RAW_OUTPUT_OVERFLOW`, "narrow pattern, path, or include and retry") with zero content returned.
+- Load-time failure modes changed: a broken subprocess seam now fails the first search call (`SEARCH_FAILED`) instead of failing plugin load through the probe; a missing binary is a launch failure with the packaged path, not a PATH problem.
+- The integration suite's fixture dropped a filename Windows cannot represent (`"` in a name), keeping the suite replayable on every platform.
+- Regenerating `THIRD_PARTY_NOTICES.md` surfaced a latent generator bug the new dependency made visible: Node's `fs.globSync` returns OS-native separators, so on Windows the `/`-suffixed dev-area prefixes in the notices tiering never matched and dev-only packages (test tooling, support leaves) were mis-tiered as runtime. The generator now normalizes manifest paths at ingestion, and the notices are platform-independent.
+- The `@vscode/ripgrep` dependency adds its MIT row to the runtime tier, and pnpm 11's truncated virtual-store directory names needed a content-scan fallback in the notices generator's metadata lookup.

.agents/notes/implemented/architecture/2026-08-01-packaged-ripgrep-search.zh.md

@@ -0,0 +1,38 @@
+# Agent Note: glob/grep 改用打包的 ripgrep 二进制直接 spawn
+
+Status: implemented
+
+[English](2026-08-01-packaged-ripgrep-search.md) | 中文
+
+> 取代 [bash 承载的 grep/glob 发现工具](../../archived/feature/2026-07-09-bash-backed-grep-glob-discovery.md)：v1 决策中明确延期的方案——直接 spawn ripgrep——现在成为实际交付的实现。
+
+## 问题
+
+`glob`/`grep` 工具经由 bash 执行器 seam 运行，这使系统 `rg` 安装成为宿主依赖。Windows 和容器镜像的 `PATH` 默认没有 `rg`，工具在那里会静默消失；部署方只能从加载期探针警告里发现这一点。bash seam 还迫使整个模型可见参数面经过一个 shell 引号工具，因为工具与 ripgrep 之间隔着一层 shell——[bash 承载决策](../../archived/feature/2026-07-09-bash-backed-grep-glob-discovery.md) 把这种耦合记为 v1 的取舍，并把直接 spawn 列为 shell 字符串域一旦被证明过于敏感时的合理后续。它确实被证明了：每个模型值都要经受 POSIX 单引号转义，探针要在测试里脚本化，执行器自身的超时分类还与协作式工具超时策略已有的职责重复。
+
+## 决策
+
+`@deepseek-ai/dsh-tool-fs-search` 现在运行 PACKAGED（打包的）ripgrep 二进制（`@vscode/ripgrep`，一个 npm 依赖，其可选平台包随附二进制），经由 `ctx.subprocess` seam：`runRipgrep()` 以纯 argv 向量 spawn `rgPath`，向量前缀 `--no-config`，配以 collect 模式 stdout/stderr、`graceMs` 与转发的 `exec.signal`。`rgPath` 在首次调用时懒解析（进程内 memoize）：`@vscode/ripgrep` 在模块求值阶段解析其平台包，静态导入会把平台包缺失/损坏（`--omit=optional`、安装不全）变成 Loader 组合加载失败——这正是本次改动要消除的加载期失败模式。不再有 shell 层，执行路径上的 shell 引号边界随之消失；`singleQuote` 工具与其 shell spawn 测试一并删除。原始流使用 seam 的诊断尾部 collect 形态（无 spill 文件——工具从不读取原始 spill 路径；lossy stdout 读取以 `SEARCH_RAW_OUTPUT_OVERFLOW` 失败）。终止宽限与 stderr 尾部预算成为经校验的 `Config` 字段（`graceMs` 默认 3000，`stderrMaxBytes` 默认 64 KiB），不再继承自 bash-local 的配置。注册变为无条件——加载期 `command -v rg` 探针与条件注册决策被删除，连同那条 "rg not found" 警告。本包注入 `tools`、`systemPrompt` 与 `subprocess`。
+
+退出语义仍由工具拥有：退出码 0 为有结果的成功，1 为成功的空搜索，其余归入既有 `SEARCH_*` 词汇（无效模式、启动失败、信号杀死、原始输出溢出）。超时是挂在工具定义上的协作式工具调用预算：`@deepseek-ai/dsh-timeout-policy` 中止 `exec.signal`，subprocess seam 的终止升级提供硬终止，工具报告 `SEARCH_ABORTED`。工作目录为会话 header cwd（存在时），否则为 `process.cwd()`——不再有执行器配置可供默认化，因此回退由工具自己拥有。
+
+`fs-glob-sampling` ACP 快照场景改为执行真实的打包二进制，作用于一个用固定 mtime 钉住 `--sort=modified` 顺序的预制工作区，取代 PATH 注入的 `rg` 替身（仅 POSIX：展示路径携带 `/` 分隔符，会话日志比较无法归一化）。
+
+## 备选方案
+
+**保留 bash seam 与探针，仅把 `rg` 记为必需宿主依赖。** 否决：宿主依赖正是本次改动要消除的失败模式，而让发现工具支持 Windows 正是此举的目的；写进文档的依赖仍是依赖。
+
+**让 `rgPath` 可注入（配置字段或环境变量覆盖），让测试与快照继续替换替身二进制。** 否决：这会新增一个只有测试 seam 会消费的公开部署面，而真实二进制本身足够确定——通过 fixture mtime 即可直接钉住；打包二进制就是部署形态，测试应当拿它来测。
+
+**改用纯 JS 的 glob/搜索引擎（如 `picomatch`/`tinyglobby`）。** 否决：[依赖替换审计](../../rejected/simplification/2026-07-26-dependency-swaps-rejected-by-nih-audit.md) 已基于"不存在 glob 引擎"的证据否决过该方向；ripgrep 语义（`--sort=modified`、VCS 剪枝、JSON 传输、正则方言）就是工具契约。
+
+## 后果
+
+- 发现工具在打包二进制覆盖的每个平台（darwin/linux/win32，x64/arm64）上开箱即用，无需宿主安装；交付的 TUI/Web 工具清单把 `glob`/`grep` 变为固定成员（见 [拉平交付的工具清单](../feature/2026-07-31-even-out-shipped-tool-rosters.md)）。
+- shell 字符串攻击面消失：恶意模式只是惰性 argv 元素，由集成套件钉住；该套件现在也在 Windows 上运行（此前没有系统 `rg` 时它自行跳过）。
+- spawn 不受沙箱约束（普通的 `ctx.subprocess` 调用），因此前缀 `--no-config`：宿主的 `RIPGREP_CONFIG_PATH`（或二进制旁的 `rg.conf`）否则可注入 `--pre` 预处理器，对每个匹配文件执行任意命令。加上 `--no-config` 后，任何配置文件——因而任何预处理器——都无法触及搜索。
+- 原始输出溢出路径的形态改变：旧的 bash 承载路径继承了 bash-local 常开的 spill，可能留下没人读的多 MB 临时文件；subprocess seam 现在无 spill 收集，溢出是纯粹的错误（`SEARCH_RAW_OUTPUT_OVERFLOW`，"narrow pattern, path, or include and retry"），不返回任何内容。
+- 加载期失败模式改变：subprocess seam 损坏现在让首次搜索调用失败（`SEARCH_FAILED`），而非通过探针使插件加载失败；二进制缺失是带打包路径的启动失败，而不是 PATH 问题。
+- 集成套件的 fixture 去掉了 Windows 无法表示的文件名（名称含 `"`），保证套件在每个平台都能重放。
+- 重新生成 `THIRD_PARTY_NOTICES.md` 暴露了一个由新依赖带出的潜在生成器 bug：Node 的 `fs.globSync` 返回操作系统原生分隔符，因此在 Windows 上 notices 分层中带 `/` 后缀的 dev 区前缀永远匹配不上，dev-only 包（测试工具、support 叶子）被错分为 runtime。生成器现在在入口处归一化清单路径，notices 与平台无关。
+- `@vscode/ripgrep` 依赖为 runtime 层增加其 MIT 行；pnpm 11 截断的虚拟存储目录名需要在 notices 生成器的元数据查找中增加内容扫描回退。

.agents/notes/implemented/bug-fix/2026-07-30-multi-select-custom-answer-composition.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/2026-07-30-multi-select-custom-answer-composition.md
+2026-07-30-multi-select-custom-answer-composition.md: 581beec89a0f0018ec2df687f5dfe1b1b5b86d22
+2026-07-30-multi-select-custom-answer-composition.zh.md: 5c9cb59822aca3fbf49fbbdf522c76f963df3480

.agents/notes/implemented/bug-fix/2026-07-30-multi-select-custom-answer-composition.md

@@ -0,0 +1,25 @@
+# Agent Note: Multi-select custom answer composition
+
+Status: implemented
+
+English | [中文](2026-07-30-multi-select-custom-answer-composition.zh.md)
+
+## Problem
+
+The user-interaction result vocabulary carries selected option labels and optional custom text in separate fields, but its original semantics made them mutually exclusive for every question. On a multi-select question, opening or typing the custom answer discarded labels the user had already selected. The TUI returned only the custom text, and the Web host rejected a client response that preserved both fields.
+
+## Decision
+
+For a question with `multiSelect: true`, one answer item may contain both a non-empty `selected` array and non-empty `custom` text. Web drafts preserve both values regardless of whether the user selects an option or types custom text first; the TUI retains pending custom text across option/custom mode switches and projects it with checked labels from either submit mode; and the Web host accepts the combined response after applying its existing id, label, uniqueness, batch, and non-empty-text validation.
+
+Single-select and optionless questions keep exclusive semantics: custom text overrides any selected option. The result shape remains `{ id, selected, custom? }`, so no wire or tool-output schema changes.
+
+## Alternatives considered
+
+**Encode custom text as another `selected` label.** Rejected because it would erase the distinction between caller-provided option labels and human-authored text, weakening validation and forcing consumers to infer which value was custom.
+
+**Allow `selected` and `custom` together for every question.** Rejected because a single-select question represents one answer; permitting a selected option plus custom text would make its cardinality ambiguous. The combined form is limited to questions that explicitly opt into multiple answers.
+
+## Consequences
+
+Multi-select UIs can represent the user's complete answer without discarding either source. Providers and consumers retain the existing DTO, while request-aware validators interpret the allowed combination from `multiSelect`. Web component and assembled-browser coverage, TUI coverage, host-response coverage, and tool-projection coverage pin the combined result. Web, TUI, and tool-projection coverage also retain labels-only answers; assembled keyless TUI coverage pins the combined terminal flow, and single-select host coverage pins the remaining exclusivity rule.

.agents/notes/implemented/bug-fix/2026-07-30-multi-select-custom-answer-composition.zh.md

@@ -0,0 +1,25 @@
+# Agent Note: 多选题自定义答案组合
+
+Status: implemented
+
+[English](2026-07-30-multi-select-custom-answer-composition.md) | 中文
+
+## 问题
+
+用户交互结果的词汇分别通过不同字段携带选中的选项标签和可选的自定义文本，但最初的语义要求每个问题的这两个字段互斥。对于多选题，打开自定义答案或输入文本会丢弃用户已选中的标签。TUI 只返回自定义文本，而 Web 宿主会拒绝同时保留两个字段的客户端响应。
+
+## 决策
+
+对于 `multiSelect: true` 的问题，一个回答项可以同时包含非空 `selected` 数组与非空 `custom` 文本。无论用户先选择选项还是先输入自定义文本，Web 草稿都会保留两个值；TUI 在选项与自定义模式之间切换时会保留待提交的自定义文本，并在任一模式提交时将其与已勾选的标签一同投影；Web 宿主则在应用现有的 id、标签、唯一性、批次和非空文本校验后接受组合响应。
+
+单选题和无选项问题仍保持互斥语义：自定义文本会覆盖任何已选中的选项。结果形状仍为 `{ id, selected, custom? }`，因此协议或工具输出 schema 均无需变更。
+
+## 考虑过的替代方案
+
+**把自定义文本编码为另一个 `selected` 标签。** 不予采纳，因为这样会抹去调用方提供的选项标签与用户填写文本之间的区别，削弱校验，并迫使消费方推断哪个值属于自定义内容。
+
+**允许所有问题同时使用 `selected` 与 `custom`。** 不予采纳，因为单选题只表示一个回答；允许选中选项与自定义文本并存会使其基数含义模糊。组合形式仅适用于显式选择多项回答的问题。
+
+## 后果
+
+多选 UI 可以完整表达用户的回答，不会丢弃任一来源。提供方和消费方继续使用现有 DTO，而请求感知的校验器会根据 `multiSelect` 判断是否允许组合。Web 组件与组装浏览器的覆盖率、TUI 覆盖率、宿主响应覆盖率和工具投影覆盖率共同固定组合结果。Web、TUI 与工具投影覆盖率还固定了仅含标签的回答形态；组装后的无密钥 TUI 覆盖率固定终端中的组合回答流程，单选题的宿主覆盖率则固定其余的互斥规则。

.agents/notes/implemented/bug-fix/2026-07-31-resume-selector-batch-projection.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/2026-07-31-resume-selector-batch-projection.md
+2026-07-31-resume-selector-batch-projection.md: 39146527f13b20813bb6f5d5f1349ecab5724662
+2026-07-31-resume-selector-batch-projection.zh.md: 10333ea7cc5e7f2051c37f3374c5dc061bd0586e

.agents/notes/implemented/bug-fix/2026-07-31-resume-selector-batch-projection.md

@@ -0,0 +1,35 @@
+# Agent Note: Resume selector folds titles only
+
+Status: implemented
+
+English | [中文](2026-07-31-resume-selector-batch-projection.zh.md)
+
+## Problem
+
+Opening the TUI `/resume` selector called `sessionQuery.readSession()` once per listed session under an unbounded `Promise.all`. Each call re-listed the whole persistence store inside `SessionCorpus.load()` (O(N²) listings), read and decompressed the complete log, replay-validated every event through the `Session` constructor, and deep-cloned the header and events up to three times — all to derive one selector row's title, last-activity time, last `turn/end` label, provider/model route, and goal phase. On a real store (185 sessions, 87 MB compressed, ~353k events) the selector took tens of seconds to open, and the cost grew with total log size rather than session count.
+
+## Decision
+
+Selector rows fold nothing but titles, and everything else a row shows comes from metadata:
+
+- Titles come from the projection system: `session-title` already registers a `title` unit, so a live row reads the registry snapshot, a persisted row reads the durable checkpoint row (`sessionProjectionCache.cachedSnapshot`, zero I/O), and only a row without a usable checkpoint pays a `coldSnapshot` — checkpoint plus a `readFrom` tail, written back so the next scan is zero-I/O. Cold reads are bounded by the TUI `resumeScanConcurrency` config. A composition without the cache falls back to one bounded `readTitleSnapshots` batch over the logs; either path isolates a per-row failure into the disabled "Unreadable session" fallback.
+- The activity timestamp never reads a log: a live session uses its last in-memory event time; a persisted session stats the artifact named by the optional `sessionPersistence.locate()` (mtime), falling back to the header's creation time when the backend locates no per-session artifact (SQLite) or the stat fails. Any append moves the mtime, so a mere pickup boundary now floats a browsed session up — accepted as the price of a metadata-only timestamp.
+- The last-turn label, provider/model route, and goal phase columns are gone from rows. Route availability is now enforced by the Enter-time preflight, which fully reads and replay-validates the one chosen log through `readSession` before handoff.
+
+The selector overlay opens synchronously when `/resume` dispatches, before the scan settles: an `undefined` candidate set renders a "Loading sessions…" placeholder, the picker owns terminal input from its first frame, Enter reports that sessions are still loading, and Escape cancels. Closing the overlay aborts the scan through the `AbortSignal` the query methods accept; a signal-ignoring backend's late settlement is dropped by a staleness check. The finished scan swaps rows in through `setCandidates` (clearing a stale still-loading error) without replacing the overlay; a queued activation behind a closing predecessor receives an already-scanned set at construction; one catch spans listing, titles, and mtimes, so any scan failure closes the overlay and reports a notice rather than stranding the loading placeholder.
+
+No session-query or session-persistence surface changed. The shipped TUI composition gains the projection registry, storage, and projection-cache rows (mirroring the web overlay over the same `storages` root, so checkpoints written by either surface serve both); the first scan over a pre-existing store still reads each log once to seed checkpoints, and every later scan is metadata-only.
+
+## Alternatives considered
+
+**Keep per-row route/turn/goal columns via a generic batch projection (`projectSessions`).** Implemented first, then rejected: it still decompressed and parsed every log on every `/resume`, so browsing cost stayed O(total log bytes), and it grew the session-query public API for one consumer. The public seam was reverted; `readTitleSnapshots` keeps using the internal `projectMany` unchanged.
+
+**Fix only the O(N²) listing inside `SessionCorpus.load()`.** Rejected as the primary fix: the per-candidate full decompress, replay validation, and triple clone dominated on large logs. The redundant pre-listing in `load()` remains a candidate cleanup with error-semantics implications.
+
+**Surface a last-modified time through `listSnapshots`/`SessionRecord`.** Cleanest seam-wise, but touches the persistence contract, both backends, and the query record shape for what the TUI can already derive from `locate()` plus one stat. Reintroduce if a second consumer needs metadata activity times.
+
+**A bespoke persisted title index or TUI-local title cache.** Rejected: the session-projection cache already is the owned durable checkpoint system with an invalidation contract (`stateVersion`, identity binding, shrunk-log anchoring); mounting it beats adding a parallel cache.
+
+## Consequences
+
+Opening `/resume` performs one listing, one stat per persisted row, and per-row title reads that touch only checkpoint rows and log tails once checkpoints exist — O(session count) metadata instead of O(total log bytes); the fallback path without the cache remains one bounded title pass. Rows show title, timestamp, status, and id only; route problems surface as an Enter-time preflight error instead of a disabled row, and a session that fails replay is caught by preflight rather than the listing. Browsed-then-abandoned sessions float up on their pickup mtime. Fake `sessionQuery` services in TUI tests provide `readTitleSnapshots` alongside `listSessions`/`readSession`, and the test harness forwards an optional `locate`. Because the picker takes focus immediately, starting a second scan requires dismissing the current overlay first — a second `/resume` typed during a scan lands in the search field, which is the intended input capture.

.agents/notes/implemented/bug-fix/2026-07-31-resume-selector-batch-projection.zh.md

@@ -0,0 +1,35 @@
+# Agent Note: 恢复选择器只折叠标题
+
+Status: implemented
+
+[English](2026-07-31-resume-selector-batch-projection.md) | 中文
+
+## Problem
+
+打开 TUI `/resume` 选择器时，会在一个无界 `Promise.all` 中对每个列出的会话调用一次 `sessionQuery.readSession()`。每次调用都会在 `SessionCorpus.load()` 内部重新列出整个持久化存储（O(N²) 次列表查询）、读取并解压完整日志、通过 `Session` 构造函数对每个事件做回放验证，并将 header 和事件深克隆多达三次——而这一切只为推导一行选择器条目的标题、最近活动时间、最后一个 `turn/end` 标签、提供方/模型路由和目标阶段。在真实存储上（185 个会话、压缩后 87 MB、约 35.3 万个事件），选择器需要数十秒才能打开，且开销随日志总大小而非会话数量增长。
+
+## Decision
+
+选择器行除标题外不折叠任何内容，行内其余信息全部来自元数据：
+
+- 标题来自投影系统：`session-title` 已注册 `title` 投影单元，因此实时行读取注册表快照，持久化行读取持久 checkpoint 行（`sessionProjectionCache.cachedSnapshot`，零 I/O），只有没有可用 checkpoint 的行才付出一次 `coldSnapshot`——checkpoint 加 `readFrom` 尾部折叠，并写回使下次扫描零 I/O。冷读取受 TUI `resumeScanConcurrency` 配置约束。未挂载缓存的组合回退到一次对日志的有界 `readTitleSnapshots` 批量读取；两条路径都把单行失败隔离为禁用的"Unreadable session"回退。
+- 活动时间戳从不读取日志：实时会话取内存中最后一个事件的时间；持久化会话对可选 `sessionPersistence.locate()` 命名的产物做 stat（mtime），当后端定位不到按会话的产物（SQLite）或 stat 失败时回退到 header 的创建时间。任何追加都会移动 mtime，因此仅仅一次 pickup 边界也会让浏览过的会话上浮——这是元数据时间戳的代价，予以接受。
+- 行内不再有最后轮次标签、提供方/模型路由和目标阶段列。路由可用性改由 Enter 时的预检强制：预检通过 `readSession` 完整读取并回放验证选中的那一份日志后才移交。
+
+选择器 overlay 在 `/resume` 分发时同步打开，早于扫描结算：`undefined` 候选集渲染"Loading sessions…"加载占位符，选择器从第一帧起就拥有终端输入，Enter 提示会话仍在加载，Escape 取消。关闭 overlay 会通过查询方法接受的 `AbortSignal` 中止扫描；忽略信号的后端的迟到结算由过期检查丢弃。扫描完成后通过 `setCandidates`（同时清除过期的仍在加载错误）换入行数据，不替换 overlay；排在正在关闭的前任之后的排队激活会在构造时直接收到已扫描的集合；列表查询、标题与 mtime 共用同一个 catch，因此任何扫描失败都会关闭 overlay 并报告通知，而不会让加载占位符悬置。
+
+session-query 与 session-persistence 的任何表面都未改变。随附的 TUI 组合新增投影注册表、storage 与投影缓存行（镜像 web overlay，共用同一 `storages` 根，因此任一表面写下的 checkpoint 都服务两者）；对既有存储的首次扫描仍会各读取一次日志以播种 checkpoint，之后的每次扫描都只读元数据。
+
+## Alternatives considered
+
+**通过通用批量投影（`projectSessions`）保留每行的路由/轮次/目标列。** 先实现后否决：它仍在每次 `/resume` 时解压并解析全部日志，浏览开销依旧是 O(日志总字节数)，且为单一消费者扩大了 session-query 公开 API。该公开接缝已回退；`readTitleSnapshots` 继续使用内部 `projectMany`，保持不变。
+
+**只修复 `SessionCorpus.load()` 内部的 O(N²) 列表查询。** 作为主要修复被否决：在大日志上，按候选行执行的完整解压、回放验证和三重克隆才是主要开销。`load()` 中的冗余预列表查询仍是一个候选清理项，但涉及错误语义。
+
+**通过 `listSnapshots`/`SessionRecord` 暴露最后修改时间。** 从接缝角度最干净，但要触碰持久化契约、两个后端和查询记录形状，而 TUI 已能用 `locate()` 加一次 stat 得到同样的信息。若出现第二个需要元数据活动时间的消费者再引入。
+
+**专门的持久化标题索引或 TUI 本地标题缓存。** 否决：session-projection 缓存本身就是自有的持久 checkpoint 系统，并已带失效契约（`stateVersion`、身份绑定、日志收缩锚定）；挂载它优于再造一套并行缓存。
+
+## Consequences
+
+打开 `/resume` 只执行一次列表查询、每个持久化行一次 stat，标题读取在 checkpoint 就绪后只触碰 checkpoint 行和日志尾部——O(会话数) 的元数据开销，而非 O(日志总字节数)；无缓存的回退路径仍是一次有界标题扫描。行内只显示标题、时间戳、状态和 id；路由问题以 Enter 时预检错误的形式出现，而不再是禁用行；回放会失败的会话由预检而非列表阶段拦截。浏览后放弃的会话会因 pickup 的 mtime 上浮。TUI 测试中的伪造 `sessionQuery` 服务在 `listSessions`/`readSession` 之外提供 `readTitleSnapshots`，测试 harness 会转发可选的 `locate`。由于选择器立即接管焦点，启动第二次扫描需要先关闭当前 overlay——扫描期间输入的第二个 `/resume` 会落入搜索字段，这正是预期的输入捕获行为。

.agents/notes/implemented/bug-fix/2026-07-31-tui-diff-context-line-accounting.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/2026-07-31-tui-diff-context-line-accounting.md
+2026-07-31-tui-diff-context-line-accounting.md: d465568d5f6cad15ef4647be7ef936c2f5824bba
+2026-07-31-tui-diff-context-line-accounting.zh.md: dd1a3eb1acf03477059d48461219700b224b2f96

.agents/notes/implemented/bug-fix/2026-07-31-tui-diff-context-line-accounting.md

@@ -0,0 +1,33 @@
+# Agent Note: TUI diff context lines stay neutral
+
+Status: implemented
+
+English | [中文](2026-07-31-tui-diff-context-line-accounting.zh.md)
+
+## Problem
+
+Result-time filesystem diffs carry the applied change with three surrounding context lines in each `FileDiff.oldText` and `FileDiff.newText`. The TUI rendered every old-side row as removed and every new-side row as added, including the identical context present on both sides. A one-line edit therefore appeared as seven removals plus seven additions, and the footer repeated those inflated totals.
+
+## Decision
+
+The TUI compares each `FileDiff` whose old and new text are both available. Added and removed rows retain their green `+` and red `-` markers; equal context rows use the recessed body tone with a neutral two-space prefix. The footer sums only the rows classified as added or removed. `maxDiffEditLength` bounds the exact comparison by its combined added and removed line count; the default is 1000. Exceeding the bound renders the complete old side as removed and the complete new side as added, marks the footer approximate, and caches that result so redraws do not repeat the comparison. A tool result clears the pending-view cache before deriving the settled view, including when a presenter mutates and reuses the same view object.
+
+When `oldText` is `null`, the renderer cannot distinguish a create from a pending overwrite or an argument fallback whose prior text is unavailable. It therefore shows every non-empty new-side row as added, without claiming those rows were absent from an existing file. Empty new content renders no synthetic added row.
+
+This remains a consumer-side interpretation of the existing `FileDiff` contract. Filesystem tools continue to persist contextual before/after snippets, so other consumers keep their placement context and existing session logs replay with corrected TUI presentation. The TUI uses the same maintained `diff` package as `dsh-tool-fs` instead of introducing a second line-diff implementation.
+
+## Alternatives considered
+
+**Remove context from filesystem result metadata.** Rejected: contextual applied hunks are intentional producer output used by capable editors, and changing them would weaken every consumer while leaving old session logs misleading in the TUI.
+
+**Extend `FileDiff` with persisted per-line tags.** Rejected: the tags can be derived deterministically from the existing before/after pair; persisting them would widen the cross-package and session-log contract solely for one renderer.
+
+**Match equal lines by position without a diff algorithm.** Rejected: insertions and deletions shift subsequent context, so positional pairing would misclassify valid hunks.
+
+**Run every comparison to completion.** Rejected: pending tool views can contain unrestricted model-authored old and new strings, and an unbounded Myers comparison can block the synchronous terminal renderer.
+
+## Consequences
+
+TUI diff cards distinguish evidence-bearing context from the mutation itself, and an exact `+A -R` footer reports the actual line delta. Replaying an existing contextual diff gains the corrected rendering without a migration. Result-time filesystem hunks are context-bounded; unrestricted pending views either complete within the configured edit-length budget or degrade to an explicitly approximate linear rendering.
+
+The focused TUI tests cover neutral context, exact totals, an empty create, bounded fallback, result-time cache invalidation, and redraw cache reuse. The assembled `advanced-cards` terminal snapshots pin the neutral context style, semantic change colors, exact footer, and approximate fallback through collapsed and expanded card states.

.agents/notes/implemented/bug-fix/2026-07-31-tui-diff-context-line-accounting.zh.md

@@ -0,0 +1,33 @@
+# Agent Note: TUI diff 上下文行保持中性
+
+Status: implemented
+
+[English](2026-07-31-tui-diff-context-line-accounting.md) | 中文
+
+## 问题
+
+文件系统 diff 返回结果时，每个 `FileDiff.oldText` 和 `FileDiff.newText` 都会包含已应用的变更及其前后各 3 行上下文。TUI 将旧侧的每一行都渲染为删除行，将新侧的每一行都渲染为新增行，其中包括两侧相同的上下文。因此，一行编辑会显示为删除 7 行并新增 7 行，页脚还会重复这些虚高的合计值。
+
+## 决策
+
+TUI 会比较每个变更前后文本均可用的 `FileDiff`。新增行和删除行仍分别使用绿色 `+` 和红色 `-` 标记；相同的上下文行则使用弱化的正文色调，并带有由两个空格构成的中性前缀。页脚只汇总归类为新增或删除的行。`maxDiffEditLength` 以新增行与删除行的合计数为精确比较设置上限，默认值为 1000。超过上限时，TUI 会把完整旧侧渲染为删除内容、把完整新侧渲染为新增内容，将页脚标记为近似结果，并缓存该结果，避免后续重绘重复比较。工具结果会在派生已结算视图前清除待处理视图缓存，即使 presenter 修改并复用同一个视图对象也一样。
+
+当 `oldText` 为 `null` 时，渲染器无法区分文件创建、待处理覆写，以及旧文本不可用的参数回退。因此，它会把新侧的每个非空行显示并计作新增行，但不会声称这些行原先不存在于已有文件中。新内容为空时，不会渲染虚构的新增行。
+
+该行为仍然只是消费方对现有 `FileDiff` 契约的解释。文件系统工具仍会持久化带上下文的变更前后片段，因此其他消费方仍能获得定位上下文，已有会话日志在回放时也会采用修正后的 TUI 呈现。TUI 与 `dsh-tool-fs` 共用同一个受维护的 `diff` 包（package），无需引入第二套逐行 diff 实现。
+
+## 考虑过的替代方案
+
+**从文件系统结果元数据中移除上下文。** 不予采纳：带上下文的已应用 hunk 是有意保留的生产方输出，供具备相应能力的编辑器使用；更改这些内容会让所有消费方丢失信息，同时旧会话日志在 TUI 中仍会产生误导。
+
+**为 `FileDiff` 扩展持久化的逐行标签。** 不予采纳：这些标签可以根据现有的变更前后文本对确定性派生；仅为一个渲染器持久化标签，会扩大跨包契约和会话日志契约。
+
+**不使用 diff 算法，按位置匹配相同行。** 不予采纳：插入和删除会使后续上下文发生位移，因此按位置配对会把有效 hunk 错误分类。
+
+**让所有比较都运行至完成。** 不予采纳：待处理工具视图可能包含由模型生成且长度不受限制的新旧字符串，无界的 Myers 比较可能阻塞同步终端渲染器。
+
+## 后果
+
+TUI diff 卡片会区分用于佐证的上下文与变更本身，精确的 `+A -R` 页脚会报告实际的行变更量。回放已有的上下文 diff 无需迁移即可获得修正后的渲染。结果时刻的文件系统 hunk 受上下文范围限制；不受限制的待处理视图要么在配置的编辑长度预算内完成比较，要么降级为明确标注为近似结果的线性渲染。
+
+聚焦的 TUI 测试覆盖中性上下文、精确合计值、空文件创建、有界回退、结果到达时的缓存失效和重绘缓存复用。组装后的 `advanced-cards` 终端快照在卡片折叠和展开状态下固定了中性上下文样式、变更行的语义色彩、精确结果页脚和近似回退。

.agents/notes/implemented/feature/2026-07-24-tui-question-dialog-multiline.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/2026-07-24-tui-question-dialog-multiline.md
+2026-07-24-tui-question-dialog-multiline.md: fc6e9bceeee4abc46a69a23124d09fcd4f3c7224
+2026-07-24-tui-question-dialog-multiline.zh.md: a56821921bad1016009687bde63eae5f4d893cdf

.agents/notes/implemented/feature/2026-07-24-tui-question-dialog-multiline.md

@@ -0,0 +1,41 @@
+# Agent Note: TUI QuestionDialog renders options across multiple lines
+
+Status: implemented
+
+English | [中文](2026-07-24-tui-question-dialog-multiline.zh.md)
+
+## Problem
+
+`ctx.userInteraction.ask()` must keep question text, supporting `detail`, option labels, descriptions, validation, and controls readable inside configured width and height bounds. The question panel also belongs directly above the editor: placing it at the terminal edge separates the pending decision from both the transcript that prompted it and the input that follows it.
+
+## Decision
+
+The TUI renders a pending question as an inline modal between the transcript/status area and the editor while retaining the shared FIFO with model and plugin overlays:
+
+- `InlineModalComponent` applies `questionDialogWidth` and `questionDialogMaxHeight` inside the normal component flow. The effective question height is additionally clamped to the current viewport after reserving the editor, so the editor remains below the question during resize.
+- `renderOptionBlock` wraps each label beneath its cursor/number prefix and renders the muted description on separately wrapped, equally indented lines. The progress header, question, custom-answer hint, validation text, and final rows are width-bounded as well; the final ellipsis clamp is only a safety boundary for prefixes or other indivisible content. The explicit `↑ N lines hidden` fallback is reserved for a viewport below the configured minimum, where the whole semantic layout cannot fit.
+- When question text or `detail` exceeds the header allocation, the header becomes a paged line viewport with its own `… lines A-B/N • PgUp/PgDn` status row. Page Up and Page Down traverse both line viewports: forward navigation exhausts the header/detail pages before entering oversized selected-option pages, and backward navigation reverses that order. This keeps plan-review detail reachable rather than leaving it behind the height clamp.
+- The option-line budget subtracts padding, header, position, and footer rows before `windowBlocks` runs. The window obeys both `maxQuestionOptions` and the remaining row budget, keeps the selected option visible, and renders omitted options as `↑ N more` / `↓ N more` markers. If fixed chrome would leave fewer than four option rows, the compact header becomes the line pager so selected content, paging status, and both option markers still fit.
+- When one selected block exceeds its allocation, it becomes a line viewport with a `lines A-B/N • PgUp/PgDn` status row. Page Up and Page Down expose every wrapped line without allowing the block to hide the option markers, validation, or controls.
+
+Package tests pin count and height bounds, header and selected-block paging order, narrow-width wrapping, selection behavior, and placement relative to retained editor input. Semantic TUI snapshots pin the assembled terminal layout, header/detail and selected-option page transitions, and validation state.
+
+## Alternatives considered
+
+**Ellipsis-only horizontal truncation.** Keeping one option per row would signal lost text without making the description readable and would not address vertical bounds. The implementation wraps readable content and retains an ellipsis only as a final safety boundary.
+
+**Wrap the combined label and description.** A composite row couples their widths, so either side can starve the other. Separate lines keep both widths predictable.
+
+**Keep the question as a bottom-edge overlay.** A terminal-edge anchor can place the panel after the editor or cover lower chrome, depending on transcript and viewport height. The inline modal preserves ordering while the modal manager retains focus and FIFO ownership.
+
+**Push the bounds into pi-tui.** Generic overlay slicing cannot identify option boundaries, selected content, controls, or the inline editor relationship. The owning dialog therefore applies semantic count, row, and paging rules.
+
+**Use only the option-count cap.** `maxQuestionOptions` remains a public count bound, but it cannot contain wrapped blocks by itself. The dialog enforces the count and row bounds together.
+
+## Consequences
+
+- Descriptions consume additional rows, so fewer options can be visible than `maxQuestionOptions`; markers state the omitted option counts.
+- Long question text and plan-review detail remain reachable inside a height-bounded panel, at the cost of sharing Page Up and Page Down with selected-option paging.
+- An oversized selected block reserves one status row and requires Page Up or Page Down to read beyond the current line page.
+- The inline question can displace older transcript rows from a short viewport. Below the configured minimum height, the final fallback can collapse upper rows behind an explicit hidden-line marker so the input controls and editor remain available.
+- The model-facing schema, selected labels, abort/cancel behavior, and ACP elicitation path are unchanged.

.agents/notes/implemented/feature/2026-07-24-tui-question-dialog-multiline.zh.md

@@ -0,0 +1,41 @@
+# Agent Note: TUI QuestionDialog 以多行方式渲染选项
+
+Status: implemented
+
+[English](2026-07-24-tui-question-dialog-multiline.md) | 中文
+
+## 问题
+
+`ctx.userInteraction.ask()` 必须确保问题正文、`detail` 补充内容、选项标签、描述、校验信息和控件在已配置的宽度与高度边界内均可读。问题面板也直接位于编辑器上方：若将其置于终端边缘，待处理决策就会同时脱离触发该决策的 transcript（文本记录）和后续输入。
+
+## 决策
+
+TUI 将待处理问题渲染为位于 transcript／状态区域与编辑器之间的内联模态框，同时仍与模型浮层和插件浮层共享 FIFO：
+
+- `InlineModalComponent` 在正常组件流内应用 `questionDialogWidth` 和 `questionDialogMaxHeight`。系统在为编辑器预留空间后，还会根据当前视口限制问题的实际高度，因此调整窗口大小时，编辑器仍位于问题下方。
+- `renderOptionBlock` 将每个标签换行到光标／编号前缀下方，并在另行换行且缩进相同的行上渲染弱化的描述。进度标题、问题、自定义答案提示、校验文本和末尾行也受宽度边界约束；最终的省略号截断仅作为前缀或其他不可拆分内容的安全边界。明确的 `↑ N lines hidden` 回退仅用于低于已配置最小值、无法容纳完整语义布局的视口。
+- 当问题正文或 `detail` 超出头部分配的空间时，头部会成为带有独立 `… lines A-B/N • PgUp/PgDn` 状态行的分页行视口。Page Up 和 Page Down 会遍历这两个行视口：向前导航先翻完问题正文／`detail` 页面，再进入超大选中选项页面；向后导航则采用相反顺序。这样可确保计划评审的 `detail` 内容始终可达，而不会被高度边界挡住。
+- 在 `windowBlocks` 运行前，选项行预算会扣除内边距、标题行、位置行和页脚行。窗口同时遵守 `maxQuestionOptions` 和剩余行预算，保持选中项可见，并将省略的选项渲染为 `↑ N more`／`↓ N more` 标记。若固定界面元素会使选项行少于四行，紧凑头部会转为行分页器，从而容纳选中内容、分页状态和上下两个选项标记。
+- 当一个选中块超出分配空间时，它会成为带有 `lines A-B/N • PgUp/PgDn` 状态行的行视口。Page Up 和 Page Down 可展示每一行已换行内容，同时防止该块遮住选项标记、校验信息或控件。
+
+包（package）测试固定数量和高度边界、头部与选中块的分页顺序、窄宽度换行、选择行为，以及问题相对于保留的编辑器输入的位置。语义 TUI 快照固定组装后的终端布局、头部／详情与选中选项的分页转换，以及校验状态。
+
+## 备选方案
+
+**仅用省略号进行横向截断。** 保持每个选项占一行，只能提示文本有所丢失，无法使描述变得可读，也无法处理纵向边界。该实现会对可读内容换行，仅将省略号保留为最终安全边界。
+
+**将标签与描述合并后换行。** 组合行会将两者的宽度耦合在一起，任一方都可能挤占另一方的空间。分行渲染可使二者的宽度保持可预测。
+
+**将问题保留为终端底边浮层。** 根据 transcript 和视口高度，锚定在终端边缘的面板可能出现在编辑器之后，也可能遮盖下方界面元素。内联模态框可保留顺序，同时由模态管理器继续负责焦点和 FIFO 所有权。
+
+**将边界处理下推至 pi-tui。** 通用浮层切片无法识别选项边界、选中内容、控件或内联编辑器关系。因此，负责该语义的对话框会应用数量、行数和分页规则。
+
+**仅使用选项数量上限。** `maxQuestionOptions` 仍是公开的数量边界，但仅靠它无法容纳已换行的块。对话框会同时执行数量边界和行数边界。
+
+## 后果
+
+- 描述会占用额外行，因此可见选项数可能少于 `maxQuestionOptions`；标记会说明省略的选项数量。
+- 较长的问题正文和计划评审 `detail` 在受高度约束的面板内仍然可达，代价是 Page Up 和 Page Down 需要与选中选项分页共用。
+- 超出空间的选中块会预留一行状态信息；若要阅读当前页面之外的行，必须使用 Page Up 或 Page Down。
+- 在较矮的视口内，内联问题可能将较早的 transcript 行挤出可见区域。低于已配置最小高度时，最终回退可能将上部行折叠到明确的隐藏行标记之后，从而让输入控件和编辑器仍然可用。
+- 面向模型的 schema、选中的标签、中止／取消行为，以及 ACP（Agent Client Protocol）的 elicitation 路径均保持不变。

.agents/notes/implemented/feature/2026-07-30-versioned-gui-welcome-onboarding.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/2026-07-30-versioned-gui-welcome-onboarding.md
-2026-07-30-versioned-gui-welcome-onboarding.md: 8155838f3b6b50f3474ef6c30065ad0d79e6f8af
-2026-07-30-versioned-gui-welcome-onboarding.zh.md: c221a6d663af60b03757f135045961bcbcdd0da7
+2026-07-30-versioned-gui-welcome-onboarding.md: 4707769d4fa9fbf184e09a2e73087dfd326070be
+2026-07-30-versioned-gui-welcome-onboarding.zh.md: c9d2e6274c476c59abc077aa3255ffe57a8a48bc

.agents/notes/implemented/feature/2026-07-30-versioned-gui-welcome-onboarding.md

@@ -14,15 +14,15 @@ The GUI's credential onboarding begins with a DeepSeek-specific readiness check,
 
 **Ownerless product onboarding belongs to `ui-settings-general`.** `src/onboarding-copy.ts` is the single editable source for the complete notice, the Continue label, and `WELCOME_NOTICE_VERSION`; both supported GUI locales intentionally render the same Chinese owner copy. Runtime locale dictionaries derive their welcome values from that file, and tests import the same owner instead of repeating paragraph text. The notice is browser UI only: it creates no Session event and contributes no model-visible content. The notice identifies `DSH_TELEMETRY_DISABLED=1` as the telemetry opt-out.
 
-**Acknowledgement is durable per Harness profile.** The Host half registers a `ui-onboarding` section in the user-settings seam, stored under the active `$DSH_HOME/settings.yaml`. The browser shows the notice unless `welcomeNoticeVersion` equals the owner constant exactly. Continue applies one path mutation with the current version and calls `complete()` only after the Host commits it; a failed write leaves the notice open, and closing the page or process writes nothing. Bumping the constant intentionally makes every profile acknowledge the revised copy once.
+**Loopback acknowledgement is durable per Harness profile.** The Host half registers a `ui-onboarding` section in the user-settings seam, stored under the active `$DSH_HOME/settings.yaml`. The connection plugin publishes whether the current page uses a loopback authority as `ctx.connection.isLoopback`; hostname classification remains internal to the connection package, and other client plugins consume the service state instead of importing its implementation. A loopback browser shows the notice unless `welcomeNoticeVersion` equals the owner constant exactly. Continue applies one path mutation with the current version and calls `complete()` only after the Host commits it; a failed write leaves the notice open, and closing the page or process writes nothing. Bumping the constant intentionally makes every profile acknowledge the revised copy once. A non-loopback browser must not call the loopback-only settings API. It presents the same notice, but explicit Continue completes the step only in the current browser process; reload or a new process presents it again.
 
-**Concurrent views converge without stale replacement.** The acknowledgement write omits `expectedRevision` deliberately: every tab writes the same version to one path, so the operation is idempotent and preserves sibling fields instead of rebuilding the section. `settings/document-updated` becomes `host/settings-changed`; an already mounted tab refetches and advances when another tab or an external editor commits the current version. The API proxy exposes this one product namespace through a closed allowlist beside configurable-provider namespaces, without treating its changes as model-catalog invalidations.
+**Concurrent loopback views converge without stale replacement.** The acknowledgement write omits `expectedRevision` deliberately: every loopback tab writes the same version to one path, so the operation is idempotent and preserves sibling fields instead of rebuilding the section. `settings/document-updated` becomes `host/settings-changed`; an already mounted loopback tab refetches and advances when another tab or an external editor commits the current version. The API proxy exposes this one product namespace through a closed allowlist beside configurable-provider namespaces, without treating its changes as model-catalog invalidations.
 
 **Onboarding temporarily owns the viewport as one continuous stage.** A solid product surface replaces the complete application view through a body-level portal and marks the underlying app root inert; the exact required mask remains mounted behind that surface with `position:absolute`, zero left/right/bottom offsets, `top:80px`, `rgba(0, 0, 0, 0.24)`, and `backdrop-filter: blur(2px)`. Welcome and conditional credential setup render as successive pages in this stage instead of independent modals. Both pages reuse the Web UI's black `BrandWordmark`. The welcome page preserves the four authored paragraphs verbatim under the `内测声明` title; every paragraph uses one 16/28 body scale, and only the requested action clause inside the final paragraph receives a subtle 500 weight. A short staggered opacity/vertical entrance supplies pacing without blocking interaction and disappears under reduced motion. The title receives initial focus, Continue is the sole button, and no close, Escape, or mask-click path exists.
 
 ## Alternatives considered
 
-**Browser local storage** — rejected because acknowledgement would follow one browser profile rather than `$DSH_HOME`; a fresh Harness profile could incorrectly inherit a prior acknowledgement, and external profile edits would have no authoritative update stream.
+**Browser local storage** — rejected because acknowledgement would follow one browser profile rather than `$DSH_HOME`; a fresh Harness profile could incorrectly inherit a prior acknowledgement, and external profile edits would have no authoritative update stream. Non-loopback fallback therefore remains process-local rather than browser-profile-local.
 
 **A second independent modal in `ui-settings-general`** — rejected because list registrants would still stack whenever welcome and credential readiness were both true. Ordered ownership belongs to the shell that declares and renders the list.
 
@@ -32,4 +32,4 @@ The GUI's credential onboarding begins with a DeepSeek-specific readiness check,
 
 ## Consequences
 
-A fresh profile always sees the welcome notice before provider-specific onboarding; an already configured credential skips only the later DeepSeek step. Reloading after Continue stays past the acknowledged version, changing the owner version presents it again, and closing before Continue leaves the next launch unchanged. Focused store and React tests pin exact-version comparison, write failure, sole-action behavior, no-dismiss paths, coordinator ordering, conditional DeepSeek transfer, and HMR cleanup. The real Chromium scenario boots the shipped Web composition with an isolated harness home, verifies the exact mask geometry and computed styles, reloads before and after acknowledgement, continues into missing-credential setup, confirms an acknowledged-version mismatch returns while the credential is configured, and checks the browser console.
+A fresh profile always sees the welcome notice before provider-specific onboarding; an already configured credential skips only the later DeepSeek step. On loopback, reloading after Continue stays past the acknowledged version, changing the owner version presents it again, and closing before Continue leaves the next launch unchanged. On non-loopback, Continue advances the live process without a privileged settings request and reload presents the notice again. Focused store and React tests pin both persistence modes, exact-version comparison, write failure, sole-action behavior, no-dismiss paths, coordinator ordering, conditional DeepSeek transfer, and HMR cleanup. The real Chromium scenario boots the shipped Web composition with an isolated harness home, verifies the exact mask geometry and computed styles, reloads before and after acknowledgement, continues into missing-credential setup, confirms an acknowledged-version mismatch returns while the credential is configured, and checks the browser console.

.agents/notes/implemented/feature/2026-07-30-versioned-gui-welcome-onboarding.zh.md

@@ -14,15 +14,15 @@ GUI 的凭据引导从 DeepSeek 专用的就绪状态检查开始，但内部测
 
 **不属于单一功能的产品引导由 `ui-settings-general` 持有。** `src/onboarding-copy.ts` 是完整通知、「继续」按钮文案和 `WELCOME_NOTICE_VERSION` 的唯一可编辑来源；GUI 支持的两种 locale 都有意渲染同一份中文所有者文案。运行时 locale 字典从该文件派生欢迎文案，测试也导入同一个所有者，而不重复段落文本。该通知只存在于浏览器 UI：它不会创建会话事件，也不会贡献任何模型可见内容。通知明确以 `DSH_TELEMETRY_DISABLED=1` 作为遥测关闭方式。
 
-**确认状态按 Harness profile 持久化。** 宿主端在 user-settings seam 中注册 `ui-onboarding` 分节，并存入当前 `$DSH_HOME/settings.yaml`。除非 `welcomeNoticeVersion` 与文案所有者文件中的常量精确相等，否则浏览器会显示通知。「继续」会以当前版本执行一次路径变更，并且仅在宿主端提交成功后调用 `complete()`；写入失败时通知保持打开，关闭页面或进程则不会写入任何内容。提升该常量会有意要求每个 profile 对修订后的文案重新确认一次。
+**loopback 确认状态按 Harness profile 持久化。** 宿主端在 user-settings seam 中注册 `ui-onboarding` 分节，并存入当前 `$DSH_HOME/settings.yaml`。connection 插件通过 `ctx.connection.isLoopback` 统一发布当前页面是否使用 loopback authority；hostname 判定函数留在 connection 包内，其他客户端插件只消费服务状态，不跨插件导入实现函数。除非 `welcomeNoticeVersion` 与文案所有者文件中的常量精确相等，否则 loopback 浏览器会显示通知。「继续」会以当前版本执行一次路径变更，并且仅在宿主端提交成功后调用 `complete()`；写入失败时通知保持打开，关闭页面或进程则不会写入任何内容。提升该常量会有意要求每个 profile 对修订后的文案重新确认一次。非 loopback 浏览器不能调用仅限 loopback 的 settings API；它仍显示同一通知，但显式点击「继续」只会在当前浏览器进程中完成该步骤，重新加载或新进程会再次显示通知。
 
-**并发视图无需陈旧的整体替换即可收敛。** 确认写入有意省略 `expectedRevision`：每个标签页都向同一路径写入相同版本，因此该操作是幂等的，并会保留同级字段，而不是重建整个分节。`settings/document-updated` 会转为 `host/settings-changed`；另一个标签页或外部编辑器提交当前版本后，已挂载的标签页会重新拉取状态并推进。API 网关在可配置提供方 namespace 之外，通过封闭的允许列表暴露这一个产品 namespace，同时不会把它的变更视为模型目录失效事件。
+**并发 loopback 视图无需陈旧的整体替换即可收敛。** 确认写入有意省略 `expectedRevision`：每个 loopback 标签页都向同一路径写入相同版本，因此该操作是幂等的，并会保留同级字段，而不是重建整个分节。`settings/document-updated` 会转为 `host/settings-changed`；另一个标签页或外部编辑器提交当前版本后，已挂载的 loopback 标签页会重新拉取状态并推进。API 网关在可配置提供方 namespace 之外，通过封闭的允许列表暴露这一个产品 namespace，同时不会把它的变更视为模型目录失效事件。
 
 **引导流程会暂时接管视口，形成一个连续阶段。** 纯色产品界面通过挂载到 `body` 的 portal 取代完整的应用视图，并将底层应用根节点标记为 inert；严格符合要求的遮罩仍挂载在该界面后方，并保留 `position:absolute`、left/right/bottom 偏移量为零、`top:80px`、`rgba(0, 0, 0, 0.24)` 和 `backdrop-filter: blur(2px)`。欢迎页和按条件显示的凭据设置页在这一阶段中依次呈现，而不是各自作为独立的模态窗口。两个页面都复用 Web UI 的黑色 `BrandWordmark`。欢迎页在 `内测声明` 标题下逐字保留既定的四段文案；所有段落统一采用 16/28 的正文字号与行高，只有最后一段中指定的行动语句使用较为克制的 500 字重。短暂的错落式透明度与纵向位移动画营造出舒缓节奏，但不会阻碍交互，并会在用户启用减少动态效果时禁用。初始焦点落在标题上，「继续」是唯一按钮，且不存在关闭、Escape 或点击遮罩的退出路径。
 
 ## 曾考虑的替代方案
 
-**浏览器本地存储**：不予采用，因为确认状态会跟随某个浏览器 profile，而不是 `$DSH_HOME`；全新的 Harness profile 可能错误继承此前的确认状态，外部 profile 编辑也没有权威更新流。
+**浏览器本地存储**：不予采用，因为确认状态会跟随某个浏览器 profile，而不是 `$DSH_HOME`；全新的 Harness profile 可能错误继承此前的确认状态，外部 profile 编辑也没有权威更新流。因此，非 loopback 的回退保持为进程内状态，而不是浏览器 profile 状态。
 
 **在 `ui-settings-general` 中再增加一个独立模态窗口**：不予采用，因为欢迎通知和凭据就绪状态同时为真时，list 注册方仍会堆叠。声明并渲染该 list 的外壳应当持有有序所有权。
 
@@ -32,4 +32,4 @@ GUI 的凭据引导从 DeepSeek 专用的就绪状态检查开始，但内部测
 
 ## 后果
 
-全新 profile 始终会在提供方专用引导之前看到欢迎通知；凭据已经配置时，只会跳过后续 DeepSeek 步骤。点击「继续」后重新加载不会再次显示已确认版本，更改文案所有者文件中的版本值会让通知重新出现，而确认前关闭窗口不会改变下次启动。针对性的 store 与 React 测试固化了精确版本比较、写入失败、单一操作、不可关闭路径、协调器顺序、按条件移交 DeepSeek 步骤和 HMR（热模块替换）清理行为。真实 Chromium 场景会使用隔离的 harness 家目录启动随产品提供的 Web 组合，验证遮罩的精确几何尺寸和计算样式，在确认前后分别重新加载，继续进入凭据缺失设置流程，确认凭据已配置时确认版本不匹配仍会使通知重新出现，并检查浏览器控制台。
+全新 profile 始终会在提供方专用引导之前看到欢迎通知；凭据已经配置时，只会跳过后续 DeepSeek 步骤。在 loopback 上，点击「继续」后重新加载不会再次显示已确认版本，更改文案所有者文件中的版本值会让通知重新出现，而确认前关闭窗口不会改变下次启动。在非 loopback 上，「继续」会在不发起受保护 settings 请求的情况下推进当前进程，重新加载则再次显示通知。针对性的 store 与 React 测试固化了两种持久化模式、精确版本比较、写入失败、单一操作、不可关闭路径、协调器顺序、按条件移交 DeepSeek 步骤和 HMR（热模块替换）清理行为。真实 Chromium 场景会使用隔离的 harness 家目录启动随产品提供的 Web 组合，验证遮罩的精确几何尺寸和计算样式，在确认前后分别重新加载，继续进入凭据缺失设置流程，确认凭据已配置时确认版本不匹配仍会使通知重新出现，并检查浏览器控制台。

.agents/notes/implemented/feature/2026-07-30-web-diff-card.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/2026-07-30-web-diff-card.md
-2026-07-30-web-diff-card.md: 396bdbc2843c1bbed5c6a913be436d8b9e96a81c
-2026-07-30-web-diff-card.zh.md: afdeafa6e94b46b4f0fbd4a065afdac8a93ac57d
+2026-07-30-web-diff-card.md: eb43e09d6173ca2270df97cecaab6da36c70a679
+2026-07-30-web-diff-card.zh.md: 669cd49abc8eba8637705cd7c9f331cc51607755

.agents/notes/implemented/feature/2026-07-30-web-diff-card.md

@@ -16,13 +16,13 @@ This is the [terminal card](2026-07-28-web-terminal-card.md) done for the `diff`
 
 `DiffBlock` is a `ui-primitives` component that renders a file mutation as an inline diff surface, and both Web render sites for a write/edit call consume the diff render intent through it: the chat tool row's body and the details panel's Output section. `ui-conversation/src/client/contract/diff-card-model.ts` is the single place that turns the snapshot's `callView`/`resultView` pair into the component's props, so the two sites cannot disagree about a change. It returns null — the generic path — whenever neither side declares `card: 'diff'`, including a `card` value this client version does not know, and whenever a settled call's result view is generic, which is how write/edit keep their execution errors on the generic path. The result side is authoritative once the call settles: the applied hunks replace the call-time diff derived from the arguments alone. A paging window that drops the call head still renders, because the result view carries the whole change.
 
-The component's contract follows the TUI's `diffLines` (`packages/ui/tui/src/components/transcript.ts`) so a diff reads the same shape across front ends:
+The component shares the TUI's single-column framing, line-terminator rule, and distinct-path file count. Line classification differs: Web renders the complete old and new sides, while the TUI derives neutral context and exact changed rows when its bounded comparison completes and labels its whole-side fallback approximate.
 
-- **One path header per file.** A new file opens a bold path header; a same-file second hunk (a scattered edit, or a `replace_all`) opens with a `⋯` gap instead of repeating the path. The `N file(s)` footer counts DISTINCT paths on both front ends — this PR moved the TUI footer off `diffs.length` onto the distinct-path count, so two hunks in one file read as `1 file` in both.
-- **The change in the diff's own colors.** A removed line is `- ` on the error token, an added line is `+ ` on the success token, drawn verbatim with `white-space: pre` inside a horizontally scrolling box — a source line is read by its indentation, so it scrolls rather than folds. A create (`oldText: null`) has no removed side.
+- **Path grouping.** A new file opens a bold path header; a same-file second hunk (a scattered edit, or a `replace_all`) opens with a `⋯` gap instead of repeating the path. The TUI keeps a path header on every hunk, but both front ends count distinct paths in the `N file(s)` footer, so two hunks in one file read as `1 file`.
+- **Whole-side change colors.** Every old-side line is `- ` on the error token and every new-side line is `+ ` on the success token, drawn verbatim with `white-space: pre` inside a horizontally scrolling box — a source line is read by its indentation, so it scrolls rather than folds. A create (`oldText: null`) has no removed side.
 - **Height cap with an expand control.** A diff longer than `DEFAULT_DIFF_MAX_LINES` (16) shows `ceil(max/2)` head rows plus the remaining tail rows, with a button between reporting the hidden count. The split arithmetic matches `TerminalBlock` and the TUI's collapsed card, so a long diff's head and tail slices agree across front ends.
-- **Line terminator.** A side's content splits on `\n` under the terminator rule `TerminalBlock` uses: empty text is zero lines (a full deletion's `newText`, a create's absent `oldText` side), a single trailing newline terminates its last line rather than adding a phantom empty one, and an interior blank line survives. This PR applied the same rule to the TUI diff branch, so the `+A -R` footer counts agree on both front ends for the newline-terminated content real write/edit calls carry.
-- **Footer and copy.** A dim `└ +A -R · N file(s)` footer summarizes the change; `+A -R` are the added/removed line counts, the same per-side counts the TUI footer draws. The copy control copies the prefixed diff text (path headers, `- `/`+ ` lines, the `⋯` gap), so a multi-file copy stays attributable.
+- **Line terminator.** A side's content splits on `\n` under the terminator rule `TerminalBlock` and the TUI use: empty text is zero lines (a full deletion's `newText`, a create's absent `oldText` side), a single trailing newline terminates its last line rather than adding a phantom empty one, and an interior blank line survives.
+- **Footer and copy.** A dim `└ +A -R · N file(s)` footer reports the Web card's complete new- and old-side line counts. The TUI footer instead reports exact changed rows when available and marks a bounded whole-side fallback approximate; both use the same distinct-path file count. The copy control copies the prefixed Web diff text (path headers, `- `/`+ ` lines, the `⋯` gap), so a multi-file copy stays attributable.
 
 Geometry, radius, and fonts mirror `CodeBlock`/`TerminalBlock` so a diff card, a terminal card, and a fenced block read as one family; `white-space: pre` plus horizontal scroll is the deliberate divergence. The copy control floats in the card's top-right corner rather than on a banner row of its own, because a banner carrying only a copy button drew an empty band above the first diff line — the TUI diff card has no banner either, only the footer.
 

.agents/notes/implemented/feature/2026-07-30-web-diff-card.zh.md

@@ -16,13 +16,13 @@ Web 客户端忽略了它。write/edit 调用落到 `GenericToolCard`，其行
 
 `DiffBlock` 是一个 `ui-primitives` 组件，把文件改动渲染为内联 diff 表面，write/edit 调用的两个 Web 渲染点都通过它消费 diff 渲染意图：chat 工具行的行体和详情面板的 Output 区。`ui-conversation/src/client/contract/diff-card-model.ts` 是唯一把快照的 `callView`/`resultView` 对转成组件 props 的地方，因此两个渲染点不会对一次改动产生分歧。当两侧都未声明 `card: 'diff'` 时它返回 null —— 走通用路径 —— 包括本客户端版本不认识的 `card` 值，以及已结算调用的 result view 是 generic 的情况（write/edit 的执行错误正是这样留在通用路径上的）。调用结算后 result 侧是权威：已应用的 hunk 替换仅从参数推导的 call 时 diff。分页窗口丢弃了 call 头也仍能渲染，因为 result view 携带完整改动。
 
-组件的契约遵循 TUI 的 `diffLines`（`packages/ui/tui/src/components/transcript.ts`），使 diff 在两个前端读起来是同一形态：
+该组件与 TUI 共用单栏框架、行终止符规则和去重路径计数。两者的行分类不同：Web 渲染完整的变更前后两侧，而 TUI 会在有界比较完成时派生中性上下文和精确变更行，并把整侧回退标记为近似结果。
 
-- **每个文件一个路径头。** 新文件开启一个粗体路径头；同文件的第二个 hunk（分散编辑，或 `replace_all`）以一个 `⋯` gap 开启，而非重复路径。`N file(s)` 页脚在两个前端都统计**去重后的路径数** —— 本 PR 把 TUI 页脚从 `diffs.length` 改为去重路径计数，因此同文件两个 hunk 在两端都读作 `1 file`。
-- **改动用 diff 自身的颜色。** 删除行是 error token 上的 `- `，新增行是 success token 上的 `+ `，在横向滚动的盒子里以 `white-space: pre` 逐字绘制 —— 源码行靠缩进阅读，所以滚动而不折行。新建（`oldText: null`）没有删除侧。
+- **路径分组。** 新文件开启一个粗体路径头；同文件的第二个 hunk（分散编辑，或 `replace_all`）以一个 `⋯` gap 开启，而非重复路径。TUI 在每个 hunk 上都保留路径头，但两个前端的 `N file(s)` 页脚都按去重路径计数，因此同文件两个 hunk 在两端都读作 `1 file`。
+- **整侧改动配色。** 旧侧每一行都以 error token 上的 `- ` 显示，新侧每一行都以 success token 上的 `+ ` 显示，并在横向滚动的盒子里以 `white-space: pre` 逐字绘制：源码行靠缩进阅读，因此滚动而不折行。新建（`oldText: null`）没有删除侧。
 - **高度上限带展开控件。** 长于 `DEFAULT_DIFF_MAX_LINES`（16）的 diff 显示 `ceil(max/2)` 个头部行加剩余尾部行，中间一个按钮报告隐藏行数。分割算术与 `TerminalBlock` 和 TUI 的折叠卡片一致，因此长 diff 的头尾切片在两个前端一致。
-- **行终止符。** 每一侧的内容按 `TerminalBlock` 的终止符规则在 `\n` 上切分：空文本是零行（整文件删除的 `newText`、新建缺失的 `oldText` 侧），单个结尾换行终止其最后一行而非新增一条幻影空行，内部空行保留。本 PR 把同一规则应用到了 TUI diff 分支，因此对于真实 write/edit 调用携带的以换行结尾的内容，两个前端的 `+A -R` 页脚计数一致。
-- **页脚与复制。** 暗色 `└ +A -R · N file(s)` 页脚概括改动；`+A -R` 是新增/删除行数，与 TUI 页脚绘制的每侧计数相同。复制控件复制带前缀的 diff 文本（路径头、`- `/`+ ` 行、`⋯` gap），使多文件复制保持可归属。
+- **行终止符。** 每一侧的内容按 `TerminalBlock` 与 TUI 共用的终止符规则在 `\n` 上切分：空文本是零行（整文件删除的 `newText`、新建缺失的 `oldText` 侧），单个结尾换行终止其最后一行而非新增一条幻影空行，内部空行保留。
+- **页脚与复制。** 暗色 `└ +A -R · N file(s)` 页脚报告 Web 卡片完整新侧与旧侧的行数。TUI 页脚则在可用时报告精确变更行数，并把有界整侧回退标记为近似结果；两者使用相同的去重路径计数。复制控件复制带前缀的 Web diff 文本（路径头、`- `/`+ ` 行、`⋯` gap），使多文件复制保持可辨别归属。
 
 几何、圆角、字体镜像 `CodeBlock`/`TerminalBlock`，使 diff 卡片、terminal 卡片、代码块读起来是一家；`white-space: pre` 加横向滚动是刻意的分歧。复制控件浮在卡片右上角，而非占据自己的 banner 行，因为只放一个复制按钮的 banner 会在第一行 diff 上方画出一条空带 —— TUI 的 diff 卡片也没有 banner，只有页脚。
 
@@ -32,7 +32,7 @@ chat 行把 diff 常驻渲染在路径链接摘要之下，上限 `CHAT_DIFF_MAX
 
 **并排（双栏）diff。** owner 目前拒绝：它更密但不适合狭窄的 chat 行，目标是与 TUI 单栏统一形式对齐。详情面板里的双栏模式是后续的 props 改动，不是重设计。
 
-**git 式行号槽。** `FileDiff` 契约只携带 `{ path, oldText, newText }` —— `structuredPatch` 的 hunk 起始行在 `diff.ts` 里被丢弃，所以没有行号抵达客户端。渲染行号槽需要后端契约改动（携带 `oldStart`/`newStart`）并同步升级 TUI 以保持一致；推迟，使本 PR 保持为对既有契约的纯 Web 消费。
+**git 式行号槽。** `FileDiff` 契约只携带 `{ path, oldText, newText }` —— `structuredPatch` 的 hunk 起始行在 `diff.ts` 里被丢弃，所以没有行号抵达客户端。渲染行号槽需要后端契约改动（携带 `oldStart`/`newStart`）并同步升级 TUI 以保持一致；推迟，使本 PR（Pull Request）保持为对既有契约的纯 Web 消费。
 
 **复用 `CodeBlock`。** 因与 terminal 卡片相同的理由拒绝：`CodeBlock` 会折行，且没有每行 `+`/`-` 角色、没有路径头、没有页脚。两者共享几何与字体 token，那是唯一一处一个实现对两者都正确的部分。
 

.agents/notes/implemented/feature/2026-07-31-even-out-shipped-tool-rosters.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/2026-07-31-even-out-shipped-tool-rosters.md
-2026-07-31-even-out-shipped-tool-rosters.md: fe2ed54a70934918b739ac466dc4b0e8f4a93115
-2026-07-31-even-out-shipped-tool-rosters.zh.md: 14cf6e891aa368bcaee8d977fbf5263f36a39dc6
+2026-07-31-even-out-shipped-tool-rosters.md: e325f4614f8d7305ce2c6199a25afd56b51fad61
+2026-07-31-even-out-shipped-tool-rosters.zh.md: a9b49c454d78387583aa7dd9e25f5d5c850a15ae

.agents/notes/implemented/feature/2026-07-31-even-out-shipped-tool-rosters.md

@@ -12,7 +12,7 @@ The result was a user-visible difference nobody had decided: the same model, ask
 
 ## Decision
 
-The rows that are not surface-specific move into [`base.cordis.yml`](../../../../apps/cli/config/base.cordis.yml), and three more join them: `tool-session-query`, `tool-str-replace-editor`, and `repeat-tool-guard`. Web search moves there too; its [deployment decision](2026-07-31-web-default-search.md) owns the security boundary while the shared base owns its surface-neutral mount. Both surfaces assemble the same roster: twenty tools on every host, plus `glob` and `grep` when ripgrep is available. `tool-session-query` joined and then left again — the [session-search-not-shipped-default decision](2026-08-02-session-search-not-shipped-default.md) keeps the model-facing consumer opt-in — while the rest of this roster stands.
+The rows that are not surface-specific move into [`base.cordis.yml`](../../../../apps/cli/config/base.cordis.yml), and three more join them: `tool-session-query`, `tool-str-replace-editor`, and `repeat-tool-guard`. Web search moves there too; its [deployment decision](2026-07-31-web-default-search.md) owns the security boundary while the shared base owns its surface-neutral mount. Both surfaces assemble the same roster: twenty-two tools on every host — the twenty shared rows plus `glob` and `grep`, which are fixed members because `dsh-tool-fs-search` spawns the [packaged ripgrep binary](../architecture/2026-08-01-packaged-ripgrep-search.md). `tool-session-query` joined and then left again — the [session-search-not-shipped-default decision](2026-08-02-session-search-not-shipped-default.md) keeps the model-facing consumer opt-in — while the rest of this roster stands.
 
 Two rows stay surface-specific. `tmux-context` is TUI-only because a browser surface has no terminal multiplexer to describe. `session-reference` is TUI-only because it drives the shared session-query index from the launcher's process-local path, and the browser sidebar reconciles that index on its own first search.
 
@@ -46,7 +46,7 @@ The same smoke also pins the TUI execution posture from the same artifact. Those
 
 [`apps/web/tests/shipped-composition.e2e.ts`](../../../../apps/web/tests/shipped-composition.e2e.ts) covers the Web surface in the built lane, asserting its catalog, that its access default is untouched, and that `workspace-write`'s writable roots include the temp directories — a trap that makes sandbox tests lie when the workspace sits under `/tmp` ([`roots.ts`](../../../../packages/sandbox/sandbox/src/roots.ts)).
 
-`glob` and `grep` are asserted as an all-or-nothing pair rather than fixed members: `dsh-tool-fs-search` probes `command -v rg` at load and registers neither tool without ripgrep, which is a host dependency.
+`glob` and `grep` are asserted as fixed members rather than a host-dependent pair: `dsh-tool-fs-search` spawns the packaged ripgrep binary and registers both tools unconditionally, so the pair is always present.
 
 Beyond the committed tests, both surfaces were driven against a real key from the built `apps/cli/lib/bin.js` under plain Node. Every mounted tool executed successfully, including `ralph` and `web_search`; the model never reached `cordis_*` or `mcp_*`, fell back to `grep` when asked for LSP navigation, and used a background `bash` task when asked for a persistent terminal.
 
@@ -62,7 +62,7 @@ Beyond the committed tests, both surfaces were driven against a real key from th
 
 ## Consequences
 
-The same model gets the same tools on both surfaces, and the difference that existed for no recorded reason is gone. The tests assert the twenty unconditional names exactly and require the ripgrep-dependent pair to be either present together or absent together on both sides, so a later change that alters only one surface fails a check instead of shipping quietly; the [session-search-not-shipped-default decision](2026-08-02-session-search-not-shipped-default.md) is exactly such a later change, and both tests moved with it.
+The same model gets the same tools on both surfaces, and the difference that existed for no recorded reason is gone. The tests assert the twenty unconditional names exactly and pin `glob` and `grep` as fixed members on both sides, so a later change that alters only one surface fails a check instead of shipping quietly; the [session-search-not-shipped-default decision](2026-08-02-session-search-not-shipped-default.md) is exactly such a later change, and both tests moved with it.
 
 `apps/cli` gained five workspace dependencies: four the shipped tree mounted, plus `dsh-mcp-client`, which it does not mount and which exists so an installed `dsh` can. Four remain — the [session-search-not-shipped-default decision](2026-08-02-session-search-not-shipped-default.md) removed `@deepseek-ai/dsh-tool-session-query` along with its row.
 

.agents/notes/implemented/feature/2026-07-31-even-out-shipped-tool-rosters.zh.md

@@ -12,7 +12,7 @@ Status: implemented
 
 ## 决策
 
-那些并非 surface 专属的行移入 [`base.cordis.yml`](../../../../apps/cli/config/base.cordis.yml),另有三行加入:`tool-session-query`、`tool-str-replace-editor` 和 `repeat-tool-guard`。Web 搜索也一并移入；其[部署决策](2026-07-31-web-default-search.md)负责安全边界，共享 base 则负责与 surface 无关的挂载。两个 surface 组装同一份清单：每台宿主上都有二十个工具，ripgrep 可用时再加上 `glob` 和 `grep`。`tool-session-query` 加入后又退出了——[session-search-not-shipped-default 决策](2026-08-02-session-search-not-shipped-default.md)让面向模型的消费方保持需显式启用——而这份清单的其余部分保持不变。
+那些并非 surface 专属的行移入 [`base.cordis.yml`](../../../../apps/cli/config/base.cordis.yml),另有三行加入:`tool-session-query`、`tool-str-replace-editor` 和 `repeat-tool-guard`。Web 搜索也一并移入；其[部署决策](2026-07-31-web-default-search.md)负责安全边界，共享 base 则负责与 surface 无关的挂载。两个 surface 组装同一份清单：每台宿主上都有二十二个工具——二十个共享行加上 `glob` 和 `grep`，它们成为固定成员，因为 `dsh-tool-fs-search` 直接 spawn [打包的 ripgrep 二进制](../architecture/2026-08-01-packaged-ripgrep-search.md)。`tool-session-query` 加入后又退出了——[session-search-not-shipped-default 决策](2026-08-02-session-search-not-shipped-default.md)让面向模型的消费方保持需显式启用——而这份清单的其余部分保持不变。
 
 有两行仍是 surface 专属。`tmux-context` 只在 TUI,因为浏览器 surface 没有终端复用器可描述。`session-reference` 只在 TUI,因为它以 launcher 的进程本地路径驱动共享的 session-query 索引,而浏览器侧边栏会在自己的首次搜索里重建该索引。
 
@@ -46,7 +46,7 @@ Status: implemented
 
 [`apps/web/tests/shipped-composition.e2e.ts`](../../../../apps/web/tests/shipped-composition.e2e.ts) 在构建产物 lane 中覆盖 Web surface,断言它的工具目录、它的访问默认值未被触碰,以及 `workspace-write` 的可写根包含临时目录——一个会让沙箱测试说谎的陷阱,当工作区落在 `/tmp` 下时([`roots.ts`](../../../../packages/sandbox/sandbox/src/roots.ts))。
 
-`glob` 与 `grep` 被作为全有或全无的一对断言,而不是固定成员:`dsh-tool-fs-search` 在加载时探测 `command -v rg`,没有 ripgrep 就两个工具都不注册,这是宿主依赖。
+`glob` 与 `grep` 被作为固定成员断言，而不是一对宿主依赖：`dsh-tool-fs-search` spawn 打包的 ripgrep 二进制并无条件注册两个工具，因此这一对始终在场。
 
 除入库测试外,两个 surface 都以 plain Node 从构建产物 `apps/cli/lib/bin.js` 出发、用真实密钥驱动过。每一个已挂载的工具都执行成功,包括 `ralph` 与 `web_search`;模型从未触达 `cordis_*` 或 `mcp_*`,被要求做 LSP 跳转时退化到 `grep`,被要求开持久终端时用了后台 `bash` 任务。
 
@@ -62,7 +62,7 @@ Status: implemented
 
 ## 后果
 
-同一个模型在两个 surface 上拿到同样的工具,那处没有记录理由的差异消失了。测试会精确断言二十个无条件提供的名称，并要求依赖 ripgrep 的一对工具在两侧要么同时存在、要么同时缺席，因此日后只改一个 surface 都会让检查失败而不是悄悄发出去；[session-search-not-shipped-default 决策](2026-08-02-session-search-not-shipped-default.md)正是这样一次后来的改动，两个测试也随之移动。
+同一个模型在两个 surface 上拿到同样的工具,那处没有记录理由的差异消失了。测试会精确断言二十个无条件提供的名称，并把 `glob` 与 `grep` 作为固定成员钉在两侧，因此日后只改一个 surface 都会让检查失败而不是悄悄发出去；[session-search-not-shipped-default 决策](2026-08-02-session-search-not-shipped-default.md)正是这样一次后来的改动，两个测试也随之移动。
 
 `apps/cli` 增加了五个 workspace 依赖:四个是交付树当时挂载的,外加 `dsh-mcp-client`——它并不被挂载,存在的意义是让已安装的 `dsh` 能挂。四个保留了下来——[session-search-not-shipped-default 决策](2026-08-02-session-search-not-shipped-default.md)把 `@deepseek-ai/dsh-tool-session-query` 连同它的行一起移除了。
 

.agents/notes/implemented/simplification/2026-08-03-omit-invariants-from-shipped-config.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/simplification/2026-08-03-omit-invariants-from-shipped-config.md
+2026-08-03-omit-invariants-from-shipped-config.md: ff9a3b0ab2b4797ca4e96bea9e6b961b2e38501f
+2026-08-03-omit-invariants-from-shipped-config.zh.md: 526ce0756e69b36b6f54b46b23e1814322d6a1fd

.agents/notes/implemented/simplification/2026-08-03-omit-invariants-from-shipped-config.md

@@ -0,0 +1,30 @@
+# Agent Note: Omit runtime invariants from shipped dsh config
+
+Status: implemented
+
+English | [中文](2026-08-03-omit-invariants-from-shipped-config.zh.md)
+
+## Problem
+
+`@deepseek-ai/dsh-invariants` and package-owned `./invariant` companions are optional development diagnostics. The shipped TUI mounted the service and four stateful companions while the shipped Web tree omitted them, so the two product surfaces had different diagnostic cost and failure behavior. A relational assertion failure could terminate an ordinary TUI run even though the always-on product boundary remained responsible for session validation and immutable history.
+
+## Decision
+
+The shipped `dsh` configuration trees under `apps/cli/config/` mount neither `@deepseek-ai/dsh-invariants` nor any package-owned `./invariant` companion. The CLI package therefore carries no direct dependency on the invariant service.
+
+Invariant support remains available for focused tests, example bundles, generated SDK compositions, and custom deployments that opt into diagnostics explicitly. Session validation, snapshotting, freezing, and provenance remain always on and do not depend on the optional service, as defined by the [source-owned immutability decision](../architecture/2026-06-11-dev-invariants-over-deep-readonly.md).
+
+The built CLI config-dump test checks both shipped surfaces and rejects either the service entry or any `@deepseek-ai/dsh-*/invariant` entry.
+
+## Alternatives considered
+
+- **Mount the service with `enabled: false`.** Rejected because the shipped tree and CLI dependency would still carry diagnostics that install no checks.
+- **Keep the TUI-only mount.** Rejected because the shipped surfaces would retain different diagnostic and failure behavior.
+- **Remove invariant support from the repository.** Rejected because package-owned checks remain useful in tests, examples, generated SDKs, and explicit development compositions; only the default product config is out of scope.
+
+## Consequences
+
+- Ordinary `dsh` TUI and Web runs install no invariant listeners or trace state and cannot fail through `InvariantError`.
+- Development and custom compositions retain explicit access to the invariant service and companions.
+- The shipped config absence is verified from the built CLI's composed output for both surfaces.
+- Always-on session integrity remains unchanged.

.agents/notes/implemented/simplification/2026-08-03-omit-invariants-from-shipped-config.zh.md

@@ -0,0 +1,30 @@
+# Agent Note: 从交付的 dsh 配置中省略运行时不变式
+
+Status: implemented
+
+[English](2026-08-03-omit-invariants-from-shipped-config.md) | 中文
+
+## 问题
+
+`@deepseek-ai/dsh-invariants` 与各包（package）拥有的 `./invariant` 伴随插件是可选的开发诊断。交付的 TUI 挂载了该服务和四个有状态伴随插件，而交付的 Web 配置树省略了这些条目，导致两个产品 surface 的诊断成本和失败行为不同。即使始终启用的产品边界仍负责会话验证与不可变历史，关系断言失败也可能终止普通的 TUI 运行。
+
+## 决策
+
+`apps/cli/config/` 下交付的 `dsh` 配置树既不挂载 `@deepseek-ai/dsh-invariants`，也不挂载任何包拥有的 `./invariant` 伴随插件。因此，CLI 包不再直接依赖不变式服务。
+
+不变式支持仍可供聚焦测试、示例组合包、生成的 SDK 组合，以及显式选择诊断的自定义部署使用。会话验证、快照、冻结和 provenance 始终启用，且不依赖可选服务，具体由[源端拥有的不可变性决策](../architecture/2026-06-11-dev-invariants-over-deep-readonly.md)规定。
+
+构建后 CLI 的配置转储测试会检查两个交付的 surface，并拒绝服务条目或任何 `@deepseek-ai/dsh-*/invariant` 条目。
+
+## 已考虑的替代方案
+
+- **挂载服务并设置 `enabled: false`。** 不予采纳，因为交付的配置树和 CLI 依赖仍会携带不安装任何检查的诊断。
+- **保留仅由 TUI 挂载的方案。** 不予采纳，因为两个交付的 surface 仍会保留不同的诊断和失败行为。
+- **从仓库中移除不变式支持。** 不予采纳，因为包拥有的检查在测试、示例、生成的 SDK 及显式开发组合中仍然有用；只有默认产品配置不在其范围内。
+
+## 后果
+
+- 普通的 `dsh` TUI 与 Web 运行不安装不变式监听器或 trace 状态，也不会因 `InvariantError` 失败。
+- 开发和自定义组合仍可显式使用不变式服务及伴随插件。
+- 构建后 CLI 的组合输出会验证两个 surface 的交付配置中均不存在这些条目。
+- 始终启用的会话完整性保持不变。