dsh-embedded-workbench 0.8.12 → 0.9.1

package/README.en-US.md

@@ -8,7 +8,7 @@ Embedded C/C++ firmware toolbox — 4 agents, 8 skills covering FreeRTOS, ISR, N
 storage, Keil MDK (AC5/AC6), ARMCLANG, HardFault triage, state machines,
 architecture principles, LVGL patterns, and claim fact-checking.
 
-**Cross-platform** — works with Claude Code, Codex CLI, Cursor, Kimi CLI, OpenCode, and ZCode. Built on the [Agent Skills](https://agentskills.io) open standard.
+**Cross-platform** — works with Claude Code, Codex CLI, Cursor, Kimi CLI, OpenCode, ZCode, and DeepSeek Harness (dsh). Built on the [Agent Skills](https://agentskills.io) open standard.
 
 ## Components
 
@@ -32,8 +32,9 @@ architecture principles, LVGL patterns, and claim fact-checking.
 | `c-cpp-dev` | Code generation, style, memory layout, refactoring for C/C++ |
 | `state-machine-design` | State models, retries, timeouts, transition gates, implementation patterns |
 | `hardfault-triage` | Processor exception triage — fault registers, stack frames, PC-to-source, root-cause classification |
+| `fact-check` | Claim-check fallback: verifies API names, file paths, enum values, counts, and mechanism feasibility against the codebase; used by the Plan Verification Gate when logicprobe is not installed |
 
-`logicprobe` (design-doc & plan claim verification) was **split out into its own plugin** — see [Other Plugins Recommended](#other-plugins-recommended).
+`logicprobe` (design-doc & plan claim verification) was **split out into its own plugin** — see [Other Plugins Recommended](#other-plugins-recommended). When it is not installed, the Plan Verification Gate falls back to this plugin's built-in `fact-check` skill; only behavioral/model claims degrade to manual confirmation.
 
 > The skill content is mostly distilled from the author's personal embedded/firmware engineering experience and code-cleanliness discipline, based on real-world pitfalls and engineering constraints.
 
@@ -84,7 +85,7 @@ Then enable in `~/.claude/settings.json`:
 Native dsh support ships as a cordis plugin bundle at the repository root (the root `package.json` declares `dsh.bundle`):
 
 - The skills are discovered as-is by dsh's `skill-filesystem` provider (Agent Skills open standard) — zero code.
-- The bundle injects the first-model-step gate (1% Rule / Red Flags / Plan Verification Gate) into the first model step of every agent session — the dsh-native counterpart of the Claude `SessionStart` hook. It also registers a model-visible catalog entry (`cordis_inspect`).
+- The bundle folds a **trimmed** first-model-step gate (the Plan Verification Gate plus a context-budget rule) into the first model step of every agent session — the dsh-native counterpart of the Claude `SessionStart` hook — and always registers the model-visible catalog entry (`cordis_inspect`). For why the 1% Rule and the Red Flags table left the payload, see [Design trade-offs and feedback](#design-trade-offs-and-feedback).
 - The 4 custom agents are intentionally not ported — dsh's native subagent tooling covers parallel multi-agent work.
 
 Install (native bundle, recommended):
@@ -104,13 +105,32 @@ Restart the profile, then run `dsh --profile web --dump-config`: the `id: embedd
 
 ## Usage
 
-The plugin auto-injects a capability notification into the first model step with a skill table, 1% Rule, and Red Flags reinforcement. Skills are loaded on demand:
+Skills load on demand and do not depend on injection:
 
-- Say "use Multi-Agent Workflow" or invoke `Skill("embedded-workbench")` for the full workflow system
+- Invoke `Skill("embedded-workbench")` for the workflow and engineering policies — it picks a light or full path by risk, and does not force fixed stages
 - Domain skills activate automatically when their `Use when` description matches your task — NOT clauses prevent false triggers (e.g., formatting-only won't load c-cpp-dev)
 - The agent proactively suggests verification, adversarial probing, and parallel subagents when it detects state machines, behavioral claims, or multi-module tasks
 - No manual CLAUDE.md configuration required
 
+The plugin also folds a **trimmed** gate (about 400 tokens) into the first model step, carrying just two things: the Plan Verification Gate, and a context-budget rule (never guess a readout you cannot see; when a large step shows no signal, ask the user to decide). Set `enabled: false` to drop it entirely.
+
+## Design trade-offs and feedback
+
+This revision walks back an earlier decision on the evidence, and the reasoning is below — challenge it.
+
+**Background.** We checked the official documentation for all 8 supported harnesses one by one (and read the source for Codex CLI). One assumption did not survive: **7 of the 8 expose no context-budget readout to the model at all** (Claude Code, Copilot CLI, Cursor, OpenCode, Kimi CLI, ZCode and dsh show token figures only in the user's interface; only Codex has a `get_context_remaining` tool, and it is off by default). Asking the model to judge "do I have budget to delegate?" therefore had nothing to stand on.
+
+**So we changed two things.**
+
+1. **Cost is now managed by trimming, not by switching injection off.** The first-step gate carries only two things: the **Plan Verification Gate** (verify, or tell the user you did not) and a **context-budget rule** (never guess a readout; when a large step shows no signal, ask the user to decide). The payload went from ~1,400 tokens to ~400 (Claude side −72%, dsh side −58%, measured) and it is on by default.
+2. **The verification gate stays; the enforcement scaffolding goes.** The 1% Rule and the 9-row Red Flags table left the **injected payload** because they are enforcement, and reported experience shows capable models follow that kind of prompt pressure literally — producing rigid phases, unnecessary questions, and six or seven agents on a five-line task at 10–15× overhead (see [obra/superpowers#1120](https://github.com/obra/superpowers/issues/1120), [openai/codex#22005](https://github.com/openai/codex/issues/22005), [#20366](https://github.com/openai/codex/issues/20366)). The full table still lives in `Skill("embedded-workbench")`: the discipline is available on request rather than applied to everyone by default. Workflow selection likewise moved from a fixed agent chain to **risk-proportional** paths.
+
+**Deliberately kept.** The Plan Verification Gate is intact (logicprobe → the built-in `fact-check` when it is not installed → tell the user if you used neither). Following [Superpowers Lite](https://github.com/BB-84C/superpowers-lite), safety, permission and **verification** gates are the kind to keep; process ceremony is the kind to scale back.
+
+**Known uncertainty.** These budget interfaces change fast and we checked once, on 2026-09-25. Every cell a vendor does not document is marked `UNVERIFIED` in [`platform-tool-mapping.md`](skills/embedded-workbench/references/platform-tool-mapping.md) rather than filled in by analogy.
+
+**Disagree?** These are judgement calls, not settled facts — especially "the Red Flags table leaves the payload" and "the gate is on by default". Open an [issue](https://github.com/AmethystLuna/embedded-workbench/issues) with the model tier, harness and counter-example you are working with; we would rather adjust on evidence.
+
 ## Codex CLI
 
 This plugin also supports OpenAI Codex CLI. Skills follow the Agent Skills standard and work identically across both platforms. Agents are provided in Codex TOML format under `.codex/agents/`.
@@ -188,7 +208,8 @@ Skills are invoked with `$skill-name`. ZCode also auto-discovers from `.claude/s
 ## Requirements
 
 - Claude Code v2.1+ / Codex CLI latest / Cursor 2.5+ / Kimi CLI latest / OpenCode latest / ZCode 3.0+
-- DeepSeek Harness (dsh): dev preview — verified on mainline 2026-08-14 (gate bundle loaded and injected in-session)
+- DeepSeek Harness (dsh): dev preview — supports `>= 0.1.0-rc.7` (the standing declaration; this round re-measured 0.1.5-rc.2 / 0.1.5-rc.3 / 0.1.6-alpha.2 / 0.1.7-alpha.1 / 0.1.7-alpha.2 / 0.1.7-rc.1 / 0.1.7-rc.2 / 0.2.0-rc.1 / 0.2.0-rc.2 — per-release evidence in [DSH-COMPATIBILITY.md](DSH-COMPATIBILITY.md))
+- The Web Plugins-page "Gate injection" switch requires **dsh ≥ 0.1.7-alpha.1** — its settings service must project live fields. On older dsh the plugin and its 8 skills still load and still inject, with the switch simply absent and **no error**: below schemastery 3.18.3 the field degrades to an ordinary boolean, and a settings service without `whileServed` makes the client half register nothing.
 - No external dependencies
 
 ## Configuration
@@ -197,10 +218,12 @@ In DeepSeek Harness, the bundle accepts a small configuration object:
 
 | Key | Type | Default | Description |
 |---|---|---|---|
-| `enabled` | boolean | `true` | Set to `false` to disable the session-start gate injection. |
+| `enabled` | boolean | `true` | Set to `false` to drop the first-step gate injection entirely; skill registration is unaffected. |
 | `gateContent` | string | built-in gate text | Override the text injected into the first model step. |
 
-To change it, override the row by id in your profile's `cordis.patch.yml`:
+The switch is editable live in the dsh Web GUI: sidebar **Plugins** → this plugin's card → "Gate injection". It takes effect without a profile restart and controls only the injected text — turning it off leaves all eight skills registered. The same card also carries a coarser row switch: turning that off unmounts the whole row (the skills and this switch go with it). Persistent overrides still go through the profile patch below.
+
+To change it, override the row by id in your profile's `cordis.patch.yml` (the example below customises the gate text):
 
 ```yaml
 - insert:
@@ -257,7 +280,7 @@ To report a security vulnerability, do **not** open a public issue. Use the priv
 
 | Plugin | Description |
 |--------|-------------|
-| [logicprobe](https://github.com/AmethystLuna/logicprobe) | Claim-verification skill: checks every verifiable claim in design docs, architecture specs, and refactoring plans against the codebase, escalating behavioral claims to executable-model verification. Split out of this plugin; the Plan Verification Gate requires it. |
+| [logicprobe](https://github.com/AmethystLuna/logicprobe) | Claim-verification skill: checks every verifiable claim in design docs, architecture specs, and refactoring plans against the codebase, escalating behavioral claims to executable-model verification. Split out of this plugin; the Plan Verification Gate prefers it and falls back to the built-in `fact-check` skill when it is not installed. Install with `claude plugin install logicprobe@logicprobe` (on dsh: `dsh plugin --profile <name> add dsh-logicprobe`). |
 | [superpowers](https://github.com/obra/superpowers) | The original agent discipline engine — skill loading enforcement, Red Flags, subagent-driven development. Many of this plugin's agent-compliance patterns (1% Rule, Red Flags, `<SUBAGENT-STOP>`, instruction priority) were adapted from Superpowers. |
 
 ## Acknowledgments

package/README.md

@@ -5,7 +5,7 @@
 嵌入式 C/C++ 固件开发工具箱 — 4 个代理、8 个技能，覆盖 FreeRTOS、中断、NVM 存储、Keil
 MDK（AC5/AC6）、ARMCLANG、HardFault 分析、状态机、架构原则、LVGL 陷阱。
 
-**跨平台** — 支持 Claude Code、Codex CLI、Cursor、Kimi CLI、OpenCode、ZCode。基于 [Agent Skills](https://agentskills.io) 开放标准构建。
+**跨平台** — 支持 Claude Code、Codex CLI、Cursor、Kimi CLI、OpenCode、ZCode、DeepSeek Harness (dsh)。基于 [Agent Skills](https://agentskills.io) 开放标准构建。
 
 ## 组件
 
@@ -29,8 +29,9 @@ MDK（AC5/AC6）、ARMCLANG、HardFault 分析、状态机、架构原则、LVGL
 | `c-cpp-dev` | C/C++ 代码生成、风格、内存布局、重构 |
 | `state-machine-design` | 状态模型、重试、超时、转换门控、实现模式 |
 | `hardfault-triage` | 处理器异常分类 — 故障寄存器、栈帧、PC 定位源码、根因分类 |
+| `fact-check` | 声称核查回退：逐条对照代码库核实 API 名、文件路径、枚举值、数量与机制可行性；logicprobe 未安装时由 Plan Verification Gate 使用 |
 
-`logicprobe`（文档与计划声称核查技能）**已拆分为独立插件** — 见下方[其他插件推荐](#其他插件推荐)。
+`logicprobe`（文档与计划声称核查技能）**已拆分为独立插件** — 见下方[其他插件推荐](#其他插件推荐)。未安装时，Plan Verification Gate 回退到本插件自带的 `fact-check` 技能，只有行为/模型类声称降级为人工确认。
 
 > 技能内容大多来自作者个人嵌入式/固件开发工作经验和代码洁癖，按实际工程踩坑与约束沉淀，而非泛泛的模型生成内容。
 
@@ -81,7 +82,7 @@ git clone https://github.com/AmethystLuna/embedded-workbench.git ~/.claude/plugi
 原生 dsh 支持以 cordis 插件 bundle 的形式提供，位于**仓库根**（根 `package.json` 声明了 `dsh.bundle`）：
 
 - 技能遵循 Agent Skills 开放标准，被 dsh 的 `skill-filesystem` provider 原样发现——零代码。
-- bundle 将首步门禁（1% Rule / Red Flags / Plan Verification Gate）注入每个 agent 会话的第一个模型步骤——是 Claude `SessionStart` hook 在 dsh 的原生对应物，并注册了模型可见的目录条目（`cordis_inspect`）。
+- bundle 把一段**精简后**的首步门禁（Plan Verification Gate + 上下文预算规则）注入每个 agent 会话的第一个模型步骤（`enabled`，默认开启）——是 Claude `SessionStart` hook 在 dsh 的原生对应物；模型可见的目录条目（`cordis_inspect`）始终注册。为什么载荷里不再有 1% Rule / Red Flags，见[设计取舍与反馈](#设计取舍与反馈)。
 - 4 个自定义 agent 有意不移植——dsh 原生 subagent 工具已覆盖并行多 agent 工作。
 
 安装（原生 bundle，推荐）：
@@ -101,13 +102,32 @@ npx -p @deepseek-ai/dsh dsh plugin --profile web add dsh-embedded-workbench
 
 ## 使用
 
-插件在会话首个模型步骤自动注入能力通知（含技能表、1% Rule、Red Flags 强化）。技能按需加载：
+技能按需加载，不依赖注入：
 
-- 说"用 Multi-Agent Workflow"或调用 `Skill("embedded-workbench")` 加载完整工作流系统
+- 调用 `Skill("embedded-workbench")` 加载工作流与工程策略——技能内部按风险比例选择轻量或完整路径，不强制固定阶段
 - 领域技能在任务匹配其 `Use when` 描述时自动激活——NOT 子句防止误触发（如纯格式化不会加载 c-cpp-dev）
 - Agent 在检测到状态机、行为声称或多模块任务时，主动建议验证、对抗探测和并行子代理
 - 无需手动配置 CLAUDE.md
 
+插件还会在会话首个模型步骤注入一段**精简**门禁（约 400 token），只含两件事：Plan Verification Gate，以及上下文预算规则（看不到读数就不要猜；工作量大且无信号时问用户，由用户决断）。设 `enabled: false` 可完全关闭。
+
+## 设计取舍与反馈
+
+这一版做了一次**基于证据的回退**，理由写在下面，欢迎质疑。
+
+**背景**：我们逐个核对了 8 个受支持 harness 的官方文档（Codex 还对照了源码），发现一个此前想当然的前提是错的——**8 家里有 7 家根本不向模型暴露任何上下文预算读数**（Claude Code、Copilot CLI、Cursor、OpenCode、Kimi CLI、ZCode、dsh 都只把 token 数给用户界面；只有 Codex 有一个 `get_context_remaining` 工具，且默认关闭）。也就是说，让模型"按预算自行判断要不要委派/切窗口"，本身没有依据。
+
+**因此分两步处理**：
+
+1. **不再用"关掉注入"控制成本，改为"瘦身"**。首步门禁现在只保留两块：**Plan Verification Gate**（未核查就必须告诉用户，不许静默跳过）和**上下文预算规则**（看不到读数就不要猜；工作量大且无信号时问用户）。载荷从约 1,400 token 降到约 400 token（Claude 侧 −72%，dsh 侧 −58%，实测值），并且默认开启。
+2. **验证门禁保留，强制执行脚手架退场**。原来的 1% Rule 与 9 行 Red Flags 表从**注入载荷**中移除：它们属于"强制纪律"，而社区实证显示这类提示会被能力较强的模型字面执行，产生僵硬阶段、多余提问，以及五行任务拉起六七个 agent 的 10–15× 开销（见 [obra/superpowers#1120](https://github.com/obra/superpowers/issues/1120)、[openai/codex#22005](https://github.com/openai/codex/issues/22005)、[#20366](https://github.com/openai/codex/issues/20366)）。完整表格仍保留在 `Skill("embedded-workbench")` 正文里——需要纪律时纪律还在，只是不再对所有人默认施压。工作流选择也从固定 agent 串场改为**按风险比例**。
+
+**有意保留的**：Plan Verification Gate 的语义没有削弱（logicprobe → 未安装则内置 `fact-check` → 两者都没用就必须告知用户）。按 [Superpowers Lite](https://github.com/BB-84C/superpowers-lite) 的原则，安全、权限与**验证**门禁是应当保留的一类，按比例裁掉的应该是流程仪式。
+
+**已知的不确定**：各 harness 的预算接口变动很快，我们只在 2026-09-25 核对过一次；[`platform-tool-mapping.md`](skills/embedded-workbench/references/platform-tool-mapping.md) 里凡厂商未公开的格子都明确标为 `UNVERIFIED`，没有靠类比填空。
+
+**有不同意见？** 这些取舍（尤其"Red Flags 从载荷退场"和"门禁默认开"）是可讨论的判断，不是定论。欢迎到 [Issues](https://github.com/AmethystLuna/embedded-workbench/issues) 提出——写清你用的模型档位、harness 和反例，我们倾向按证据调整。
+
 ## Codex CLI
 
 本插件同样支持 OpenAI Codex CLI。技能遵循 Agent Skills 标准，跨平台行为一致。代理以 Codex TOML 格式提供于 `.codex/agents/`。
@@ -185,7 +205,8 @@ cp -r embedded-workbench/skills/* .zcode/skills/
 ## 依赖
 
 - Claude Code v2.1+ / Codex CLI 最新版 / Cursor 2.5+ / Kimi CLI 最新版 / OpenCode 最新版 / ZCode 3.0+
-- DeepSeek Harness (dsh): dev preview — 已实测 mainline 2026-08-14（gate bundle 加载并注入会话成功）
+- DeepSeek Harness (dsh): dev preview — 支持 `>= 0.1.0-rc.7`（沿用既有声明；本轮实测覆盖 0.1.5-rc.2 / 0.1.5-rc.3 / 0.1.6-alpha.2 / 0.1.7-alpha.1 / 0.1.7-alpha.2 / 0.1.7-rc.1 / 0.1.7-rc.2 / 0.2.0-rc.1 / 0.2.0-rc.2，逐版本证据见 [DSH-COMPATIBILITY.md](DSH-COMPATIBILITY.md)）
+- Web 端的「Gate 注入」开关需要 **dsh ≥ 0.1.7-alpha.1**（设置服务必须能投影即时字段）。更早的 dsh 上插件与 8 个 skill 照常加载、照常注入，只是开关不出现、**也不报错**：schemastery 早于 3.18.3 时该字段退化为普通布尔值；设置服务没有 `whileServed` 时客户端半侧不注册任何东西。
 - 无外部依赖
 
 ## 配置
@@ -194,10 +215,12 @@ cp -r embedded-workbench/skills/* .zcode/skills/
 
 | 键 | 类型 | 默认值 | 说明 |
 |---|---|---|---|
-| `enabled` | boolean | `true` | 设为 `false` 可关闭首步 Gate 注入。 |
+| `enabled` | boolean | `true` | 设为 `false` 可完全关闭首步 Gate 注入；技能注册不受影响。 |
 | `gateContent` | string | 内置 gate 文本 | 覆盖注入到首轮模型上下文中的文本。 |
 
-在 profile 的 `cordis.patch.yml` 中按 row id 覆盖：
+在 dsh Web GUI 里可以直接改这个开关：侧边栏 **插件** → 本插件卡片 → 「Gate 注入」。它实时生效，不必重启 profile，而且只管注入的那段文本——关掉后 8 个技能照常注册。同一张卡片上还有一个更粗粒度的行开关：关掉它会整行卸载插件（技能和这个开关一起消失）。要持久化覆盖，仍按下面的 profile patch 写。
+
+在 profile 的 `cordis.patch.yml` 中按 row id 覆盖（下面的例子自定义 Gate 文本）：
 
 ```yaml
 - insert:
@@ -254,7 +277,7 @@ bash tests/skill-triggering/run-all.sh
 
 | 插件 | 简介 |
 |------|------|
-| [logicprobe](https://github.com/AmethystLuna/logicprobe) | 声称核查技能：逐条核验设计文档、架构规格、重构计划中的可验证声称与代码库是否一致，行为类声称升级为可执行模型验证。自本插件拆分；Plan Verification Gate 依赖它。 |
+| [logicprobe](https://github.com/AmethystLuna/logicprobe) | 声称核查技能：逐条核验设计文档、架构规格、重构计划中的可验证声称与代码库是否一致，行为类声称升级为可执行模型验证。自本插件拆分；Plan Verification Gate 优先使用它，未安装时回退到内置 `fact-check` 技能。安装：`claude plugin install logicprobe@logicprobe`（dsh：`dsh plugin --profile <name> add dsh-logicprobe`）。 |
 | [superpowers](https://github.com/obra/superpowers) | 原始 agent 纪律引擎——技能加载强制、Red Flags、子代理驱动开发。本插件的多项 agent 合规模式（1% Rule、Red Flags、`<SUBAGENT-STOP>`、指令优先级）均借鉴自 Superpowers。 |
 
 ## 致谢

package/cordis.patch.yml

@@ -3,6 +3,10 @@
 # Row ids are stable identity in the config tree; later layers (the user's
 # profile cordis.patch.yml, $DSH_HOME/cordis.patch.yml, --patch overlays)
 # override a row by id, replacing the whole `config` (no deep merge).
+#
+# The first-model-step gate is ON by default and deliberately small — it carries
+# the Plan Verification Gate and the context-budget rule, not the 1% Rule /
+# Red Flags enforcement scaffolding. Set `enabled: false` per profile to drop it.
 
 - insert:
     - id: embedded-workbench

package/lib/client.js

@@ -0,0 +1,214 @@
+/**
+ * embedded-workbench — browser half: the gate-injection switch on the dsh Web
+ * client's Plugins page.
+ *
+ * The Plugins page (`@deepseek-ai/dsh-client-ui-plugin-manager`) owns the
+ * sidebar **Plugins** entry and declares the slots a bundle's own configuration
+ * registers into. This module contributes one `plugins.bundle.config` entry,
+ * keyed by this package's npm name, so the switch renders on
+ * embedded-workbench's own page between its description and its rows.
+ *
+ * Why the switch writes through `configForms` rather than reaching for the
+ * profile file: dsh's settings service exposes only the Config fields declared
+ * `.volatile()`, and it rejects a write to any other path. `enabled` is such a
+ * field (see `src/index.ts`), so flipping the switch is an ordinary
+ * revision-fenced settings write that the running Host picks up in place — the
+ * injection there re-reads the reference on every model step.
+ *
+ * Shape: this is a prebuilt module-system bundle, not a source module. It calls
+ * `window.__ModuleLoader__.load({ id, factory })` with this package's resolved
+ * npm name, and `factory` returns the cordis plugin face. Only the client
+ * baseline is requested (`react` and
+ * `@deepseek-ai/dsh-client-ui-primitives`); every other capability arrives
+ * through cordis `inject`. `scripts/build-client.mjs` publishes this file
+ * verbatim as `lib/client.js`.
+ *
+ * @module dsh-embedded-workbench/client
+ */
+
+window.__ModuleLoader__.load({
+  id: 'dsh-embedded-workbench',
+  factory: (require) => {
+    const React = require('react')
+    const { Button, Switch } = require('@deepseek-ai/dsh-client-ui-primitives')
+
+    /** Settings namespace: the Loader entry id this bundle's patch declares. */
+    const NS = 'embedded-workbench'
+    /** `plugins.bundle.config` key: the bundle's npm package name. */
+    const PACKAGE = 'dsh-embedded-workbench'
+    /** This page's dictionary namespace. */
+    const LOCALE_NS = 'embedded-workbench.plugins'
+    /** The Config field the switch writes inside the namespace's section. */
+    const FIELD = 'enabled'
+
+    /** English copy. */
+    const en = {
+      title: 'Gate injection',
+      label: 'Inject the gate text',
+      hint: 'Folds the Plan Verification Gate and the context-budget rule into the first model step of every session. Turning it off leaves all eight skills registered — only the injected text is dropped.',
+      overridden: 'Overridden',
+      reset: 'Reset to default',
+      readOnly: 'This deployment stores settings read-only.',
+      unavailable: 'This plugin is not loaded, so it cannot be configured right now.',
+      saveFailed: 'The deployment did not accept that value; the switch shows what is stored.',
+    }
+    /** Simplified Chinese copy. */
+    const zh = {
+      title: 'Gate 注入',
+      label: '注入 gate 文本',
+      hint: '把 Plan Verification Gate 与上下文预算规则折进每个会话的第一个模型步。关掉后 8 个技能仍然注册，只是不再注入那段提示文本。',
+      overridden: '已覆盖',
+      reset: '恢复默认',
+      readOnly: '本部署的设置为只读。',
+      unavailable: '该插件当前未加载，暂时无法配置。',
+      saveFailed: '本部署没有接受这个值，开关显示的是已存下的状态。',
+    }
+
+    /** Required cordis services. */
+    const inject = ['slots', 'locale', 'configForms']
+
+    const GROUP = { display: 'flex', flexDirection: 'column', gap: '8px' }
+    const TITLE = { margin: 0, fontSize: '14px', fontWeight: '500', lineHeight: '22px' }
+    const ROW = { display: 'flex', alignItems: 'center', justifyContent: 'space-between', gap: '16px' }
+    const LABEL = { fontSize: '13px', lineHeight: '20px' }
+    const NOTE = { margin: 0, fontSize: '12px', lineHeight: '18px', color: 'var(--dsw-alias-label-tertiary)' }
+    const FAILED = { margin: 0, fontSize: '12px', lineHeight: '18px', color: 'var(--dsw-alias-state-error-primary)' }
+
+    /**
+     * Whether a settings-layer value carries this field, which is what marks it
+     * overridden: an override equal to the default is still an override.
+     * @param layer - the raw user layer the form snapshot carries.
+     * @returns whether the layer holds the field.
+     */
+    function carries(layer) {
+      return layer !== null && typeof layer === 'object' && Object.prototype.hasOwnProperty.call(layer, FIELD)
+    }
+
+    /**
+     * Render the gate-injection switch, or the note saying why it cannot render.
+     * @param props - the page's `t` seat, the bound form snapshot hook, and the write actions.
+     * @returns the body of this bundle's configuration section.
+     */
+    function InjectionCard(props) {
+      const t = props.t
+      const state = props.useInjectionForm((snapshot) => snapshot)
+      const [pending, setPending] = React.useState(false)
+      const [failed, setFailed] = React.useState(false)
+
+      /** Run one settings write and report a refusal or a transport failure. */
+      const write = (run) => {
+        setPending(true)
+        setFailed(false)
+        Promise.resolve(run()).then(
+          (accepted) => {
+            setPending(false)
+            setFailed(accepted === false)
+          },
+          () => {
+            setPending(false)
+            setFailed(true)
+          },
+        )
+      }
+
+      if (state.status !== 'ready') {
+        return React.createElement('p', { style: NOTE }, t('unavailable'))
+      }
+
+      const section = state.value !== null && typeof state.value === 'object' ? state.value : {}
+      // The schema default is `true`; only an explicit false means off.
+      const checked = section[FIELD] !== false
+      const overridden = carries(state.user)
+      const locked = state.writable !== true || pending
+
+      const children = [
+        React.createElement('h4', { key: 'title', style: TITLE }, t('title')),
+        React.createElement('div', { key: 'row', style: ROW }, [
+          React.createElement('span', { key: 'label', style: LABEL }, t('label')),
+          React.createElement(Switch, {
+            key: 'switch',
+            checked,
+            disabled: locked,
+            label: t('label'),
+            onChange: (next) => write(() => props.setEnabled(next)),
+          }),
+        ]),
+        React.createElement('p', { key: 'hint', style: NOTE }, state.writable === true ? t('hint') : t('readOnly')),
+      ]
+
+      if (overridden) {
+        children.push(
+          React.createElement('div', { key: 'overridden', style: ROW }, [
+            React.createElement('span', { key: 'badge', style: NOTE }, t('overridden')),
+            React.createElement(
+              Button,
+              {
+                key: 'reset',
+                variant: 'outline',
+                size: 'sm',
+                disabled: locked,
+                onClick: () => write(() => props.resetEnabled()),
+              },
+              t('reset'),
+            ),
+          ]),
+        )
+      }
+
+      if (failed) {
+        children.push(React.createElement('p', { key: 'failed', style: FAILED, role: 'alert' }, t('saveFailed')))
+      }
+
+      return React.createElement('div', { style: GROUP }, children)
+    }
+
+    /**
+     * Mount the switch while the Host serves embedded-workbench's settings
+     * namespace.
+     * @param ctx - the browser plugin context.
+     */
+    function apply(ctx) {
+      ctx.effect(() => ctx.locale.register(LOCALE_NS, { zh, en }), 'dsh-embedded-workbench: dictionaries')
+      // `whileServed` is the registration barrier that keeps this page alive only
+      // while the Host serves the namespace. Hosts predating it (measured: dsh
+      // 0.1.5-rc.3 and 0.1.6-alpha.2) have no live settings field to offer at all,
+      // so there is nothing to register — and calling it there would throw during
+      // this plugin's own activation, which the Web boot audit then reports as a
+      // failed client entry. Degrade to no page instead.
+      if (typeof ctx.configForms.whileServed !== 'function') return
+      // The page renders the section only for a bundle whose package name is in
+      // its configuration ledger, and the ledger follows this registration. The
+      // registration in turn waits for the namespace to be served, so a profile
+      // whose embedded-workbench row is switched off shows no trace of the
+      // switch.
+      ctx.effect(
+        () =>
+          ctx.configForms.whileServed([NS], () => {
+            const form = ctx.configForms.get(NS)
+            const source = {
+              getSnapshot: () => form.getSnapshot(),
+              subscribe: (listener) => form.subscribe(listener),
+            }
+            return ctx.slots.inject('plugins.bundle.config', () =>
+              ctx.slots.register(
+                {
+                  name: 'plugins.bundle.config',
+                  key: PACKAGE,
+                  locale: LOCALE_NS,
+                  inject: () => ({
+                    hooks: { injectionForm: source },
+                    setEnabled: (next) => form.set(FIELD, next),
+                    resetEnabled: () => form.unset(FIELD),
+                  }),
+                },
+                InjectionCard,
+              ),
+            )
+          }),
+        'dsh-embedded-workbench: gate-injection switch',
+      )
+    }
+
+    return { inject, apply }
+  },
+})