double-coding-lab · liguochuan00 · Aug 14, 2026 · Aug 14, 2026
diff --git a/.Knowledge/index.md b/.Knowledge/index.md
@@ -38,6 +38,7 @@
 | f2s-kb-distill | `.Knowledge/topics/f2s-kb-distill.md` | Q&A 驱动知识提取；自动判断创建新 topic 或追加现有 topic | 技能：skills/f2s-kb-distill/SKILL.mdc |
 | flow2spec-init-defaults | `.Knowledge/topics/f2s-init-defaults.md` | `flow2spec init` 字段默认值、四处落点一致性、老项目缺字段补写、`init` 不动 stock/req/topics/matchers、manifest 两个版本字段（`projectRev` / `pkgRev`）对照 | 包源码：`lib/flow2specConfig.js` `DEFAULTS` / `CONFIG_FIELDS`；模板：`templates/{locale}/flow2spec.config.json` |
 | flow2spec-doctor | `.Knowledge/topics/flow2spec-doctor.md` | `flow2spec doctor` 只读检查环境、项目初始化、协作上下文与知识库健康 | stock：[Doctor 诊断命令](stock-docs/Flow2Spec-doctor诊断命令.md)；中英文命令文档：`docs/命令说明.md` / `docs/en/commands-reference.md` |
+| flow2spec-dsh-adapter | `.Knowledge/topics/flow2spec-dsh-adapter.md` | `flow2spec init dsh`、DeepSeek Harness 项目技能发现与 `.dsh/` 目录适配 | 用户文档：`docs/使用说明.md` / `docs/en/usage-guide.md`；实现：`lib/dshAgentsAdapter.js` |
 
 每主题保留 **1–3 条** 可点击摘要链接；全量路径对照写入 `.Knowledge/migration-report.md`（迁移场景）。  
 其中 **`implement-tech-design`**、**`f2s-doc-routing`**、**`config-precheck`**、**`f2s-task`** 在 `topics/` 内为**路由摘要**；执行长文见配置根 **`rules/f2s-*.md(c)`**；使用 Codex 时见 **`.codex/AGENTS.md`**、**`.codex/topics/f2s-*.md`**（`f2s-config-check` 与 `AGENTS` 前置同源，按需打开）。**`f2s-knowledge-preflight`** 与 **`f2s-kb-feedback-closing`** 是普通问答首读 / 源码补答收口门禁，作为配置根规则 / Codex 专题长文生效，不写入 `topicPaths` 或 `taskToTopicRules`；**`f2s-kb-feedback-closing`** 在 cases 1–3 建议执行 **`f2s-kb-distill`**。

diff --git a/.Knowledge/manifest-routing.json b/.Knowledge/manifest-routing.json
@@ -33,7 +33,8 @@
     "f2s-kb-distill": ".Knowledge/topics/f2s-kb-distill.md",
     "flow2spec-init-defaults": ".Knowledge/topics/f2s-init-defaults.md",
     "flow2spec-collaboration": ".Knowledge/topics/flow2spec-collaboration.md",
-    "flow2spec-doctor": ".Knowledge/topics/flow2spec-doctor.md"
+    "flow2spec-doctor": ".Knowledge/topics/flow2spec-doctor.md",
+    "flow2spec-dsh-adapter": ".Knowledge/topics/flow2spec-dsh-adapter.md"
   },
   "taskToTopicRules": [
     {
@@ -156,6 +157,14 @@
       "topics": [
         "flow2spec-doctor"
       ]
+    },
+    {
+      "task": "flow2spec-dsh-adapter",
+      "matcherId": "m-flow2spec-dsh-adapter",
+      "matcherPath": ".Knowledge/matchers/m-flow2spec-dsh-adapter.json",
+      "topics": [
+        "flow2spec-dsh-adapter"
+      ]
     }
   ],
   "topicMetadata": {
@@ -234,6 +243,13 @@
       "tags": [
         "module"
       ]
+    },
+    "flow2spec-dsh-adapter": {
+      "primary": "feature",
+      "confidence": "inferred",
+      "tags": [
+        "module"
+      ]
     }
   }
 }
diff --git a/.Knowledge/matchers/m-flow2spec-dsh-adapter.json b/.Knowledge/matchers/m-flow2spec-dsh-adapter.json
@@ -0,0 +1,16 @@
+{
+  "includeAny": [
+    "DeepSeek Harness",
+    "deepseek-harness",
+    "dsh",
+    "flow2spec init dsh",
+    ".dsh/skills",
+    ".dsh/topics",
+    "Cordis 插件",
+    "Harness 技能",
+    "Harness 适配"
+  ],
+  "id": "m-flow2spec-dsh-adapter",
+  "version": "1.0.0",
+  "schema": "flow2spec.matcher.v1"
+}
diff --git a/.Knowledge/topics/flow2spec-dsh-adapter.md b/.Knowledge/topics/flow2spec-dsh-adapter.md
@@ -0,0 +1,27 @@
+---
+id: flow2spec-dsh-adapter
+revision: 0
+summary: "DeepSeek Harness 项目级技能初始化与目录适配"
+primary: feature
+confidence: inferred
+tags: [module]
+---
+# DeepSeek Harness 适配
+
+## 适用场景
+
+用于 `flow2spec init dsh`、DeepSeek Harness 技能发现、`.dsh/skills`、`.dsh/topics` 和根 `AGENTS.md` 入口问题。
+
+## 已落地能力
+
+- `flow2spec init dsh` 将 Flow2Spec 技能复制到 `.dsh/skills/<skill-name>/SKILL.md`。
+- 将模板规则长文镜像到 `.dsh/topics/*.md`，并写入 `.dsh/AGENTS.md` 目录指针。
+- 根目录没有 `AGENTS.md` 时生成 Harness 兼容的完整入口；已有入口不会被覆盖。
+- 原生 Cordis 插件不在本适配范围内，属于后续路线图事项。
+
+## 维护边界
+
+- Agent 注册：`lib/agents.js`
+- 初始化适配：`lib/init.js`、`lib/dshAgentsAdapter.js`
+- 诊断与回归：`lib/doctor.js`、`scripts/test-dsh-init.js`
+- 用户文档：`docs/使用说明.md`、`docs/en/usage-guide.md`、`docs/目录与路径约定.md`
diff --git a/README.en.md b/README.en.md
@@ -5,15 +5,16 @@
 </p>
 
 <p align="center">
-  <strong>Give Cursor, Claude Code, and Codex the project facts they need before editing.</strong>
+  <strong>Give each initialized AI coding client the project facts it needs before editing.</strong>
 </p>
 
 <p align="center">
   <a href="./README.zh-CN.md">中文</a> ·
   <a href="https://double-coding-lab.github.io/Flow2Spec">Live demo</a> ·
   <a href="./docs/en/Flow2Spec-Introduction.md">Introduction</a> ·
   <a href="./docs/en/usage-guide.md">Usage guide</a> ·
-  <a href="./docs/en/commands-reference.md">Commands</a>
+  <a href="./docs/en/commands-reference.md">Commands</a> ·
+  <a href="https://github.com/orgs/double-coding-lab/projects/2/views/1">Roadmap</a>
 </p>
 
 <p align="center">
@@ -26,6 +27,8 @@ Flow2Spec adds a spec-driven workflow layer to AI coding agents. It creates a sm
 
 ```bash
 npx @double-coding/flow2spec@latest init
+# DeepSeek Harness projects:
+flow2spec init dsh
 ```
 
 ## Why it exists
@@ -47,7 +50,7 @@ Flow2Spec does not add documentation for its own sake. It keeps a small, machine
 | --- | --- | --- |
 | Knowledge routing | Maps a request to the few topics the agent needs to read. | `.Knowledge/manifest-routing.json`, `.Knowledge/matchers/*.json` |
 | Topic shards | Stores project facts such as APIs, limits, locks, data rules, and workflows. | `.Knowledge/topics/*.md` |
-| Agent entrypoints | Installs rules and skills for Cursor, Claude Code, and Codex. | `.cursor/`, `.claude/`, `.codex/`, `AGENTS.md` |
+| Agent entrypoints | Installs rules and skills for the selected AI coding clients. | client configuration roots, `.dsh/`, `AGENTS.md` |
 | Skill workflows | Clarifies requirements, writes specs, implements, fixes, syncs knowledge, and commits. | `f2s-*` skills |
 | Local task state | Keeps AI steps and user-side todos separate from product knowledge. | `.task/` |
 

diff --git a/README.md b/README.md
@@ -5,15 +5,16 @@
 </p>
 
 <p align="center">
-  <strong>Give Cursor, Claude Code, and Codex the project facts they need before editing.</strong>
+  <strong>Give each initialized AI coding client the project facts it needs before editing.</strong>
 </p>
 
 <p align="center">
   <a href="./README.zh-CN.md">中文</a> ·
   <a href="https://double-coding-lab.github.io/Flow2Spec">Live demo</a> ·
   <a href="./docs/en/Flow2Spec-Introduction.md">Introduction</a> ·
   <a href="./docs/en/usage-guide.md">Usage guide</a> ·
-  <a href="./docs/en/commands-reference.md">Commands</a>
+  <a href="./docs/en/commands-reference.md">Commands</a> ·
+  <a href="https://github.com/orgs/double-coding-lab/projects/2/views/1">Roadmap</a>
 </p>
 
 <p align="center">
@@ -26,6 +27,8 @@ Flow2Spec adds a spec-driven workflow layer to AI coding agents. It creates a sm
 
 ```bash
 npx @double-coding/flow2spec@latest init
+# DeepSeek Harness projects:
+flow2spec init dsh
 ```
 
 ## Why it exists
@@ -47,7 +50,7 @@ Flow2Spec does not add documentation for its own sake. It keeps a small, machine
 | --- | --- | --- |
 | Knowledge routing | Maps a request to the few topics the agent needs to read. | `.Knowledge/manifest-routing.json`, `.Knowledge/matchers/*.json` |
 | Topic shards | Stores project facts such as APIs, limits, locks, data rules, and workflows. | `.Knowledge/topics/*.md` |
-| Agent entrypoints | Installs rules and skills for Cursor, Claude Code, and Codex. | `.cursor/`, `.claude/`, `.codex/`, `AGENTS.md` |
+| Agent entrypoints | Installs rules and skills for the selected AI coding clients. | client configuration roots, `.dsh/`, `AGENTS.md` |
 | Skill workflows | Clarifies requirements, writes specs, implements, fixes, syncs knowledge, and commits. | `f2s-*` skills |
 | Team collaboration | Keeps each developer's task state local while merging reviewed knowledge through structured deltas and topic revisions. | `.task/<developerId>/`, `.Knowledge/` |
 

diff --git a/README.zh-CN.md b/README.zh-CN.md
@@ -5,15 +5,16 @@
 </p>
 
 <p align="center">
-  <strong>让 Cursor、Claude Code、Codex 在动手改代码前，先读到正确的项目事实。</strong>
+  <strong>让每个已初始化的 AI 编程客户端在动手改代码前，先读到正确的项目事实。</strong>
 </p>
 
 <p align="center">
   <a href="./README.md">English</a> ·
   <a href="https://double-coding-lab.github.io/Flow2Spec">在线演示</a> ·
   <a href="./docs/Flow2Spec基础介绍.md">基础介绍</a> ·
   <a href="./docs/使用说明.md">使用说明</a> ·
-  <a href="./docs/命令说明.md">命令说明</a>
+  <a href="./docs/命令说明.md">命令说明</a> ·
+  <a href="https://github.com/orgs/double-coding-lab/projects/2/views/1">路线图</a>
 </p>
 
 <p align="center">
@@ -26,6 +27,8 @@ Flow2Spec 是给 AI 编码工具使用的 Spec-driven 工作流层。它会在
 
 ```bash
 npx @double-coding/flow2spec@latest init
+# DeepSeek Harness 项目：
+flow2spec init dsh
 ```
 
 ## 为什么需要它
@@ -47,7 +50,7 @@ Flow2Spec 不是为了增加文档数量。它把项目事实保存在一层小
 | --- | --- | --- |
 | 知识路由 | 把一次需求映射到 agent 需要读取的少量 topics。 | `.Knowledge/manifest-routing.json`, `.Knowledge/matchers/*.json` |
 | 主题分片 | 保存 API、上限、锁、数据规则、业务流程等项目事实。 | `.Knowledge/topics/*.md` |
-| Agent 入口 | 为 Cursor、Claude Code、Codex 安装规则和技能。 | `.cursor/`, `.claude/`, `.codex/`, `AGENTS.md` |
+| Agent 入口 | 为选中的 AI 编程客户端安装规则和技能。 | 客户端配置根、`.dsh/`、`AGENTS.md` |
 | 技能工作流 | 澄清需求、编写方案、实现、修复、同步知识、提交。 | `f2s-*` skills |
 | 团队协作 | 每个人的任务现场留在本地，确认后的知识通过结构化 delta 与 topic revision 合入共享仓库。 | `.task/<developerId>/`, `.Knowledge/` |
 

diff --git a/cli.js b/cli.js
@@ -246,26 +246,21 @@ agent（可多个，空格分隔；省略时交互选择）：
 
 示例:
   flow2spec init                  # 交互选择工具和配置
-  flow2spec init claude           # 直接写入 .claude/，跳过工具选择
-  flow2spec init cursor claude    # 同时写入 .cursor/ 与 .claude/
-  flow2spec init codex --locale en-US  # 使用英文模板初始化 Codex
+  flow2spec init <agent>          # 直接写入指定客户端配置根，跳过工具选择
+  flow2spec init <agent> <agent>  # 同时初始化多个客户端
+  flow2spec init <agent> --locale en-US  # 使用英文模板初始化指定客户端
+  flow2spec init dsh                  # 初始化 DeepSeek Harness 项目技能
   flow2spec init --yes            # 跳过所有问答，使用默认值（适合 CI）
   flow2spec init --reset-knowledge  # 强制用模板覆盖 .Knowledge（谨慎）
 
 init 会:
-  1. 交互询问要初始化的 AI 工具（cursor / claude / codex，可多选）；已通过参数指定则跳过。
+  1. 交互询问要初始化的 AI 工具（见上方 agent 列表，可多选）；已通过参数指定则跳过。
      传 --yes 或非 TTY 环境时跳过问答，使用默认值。
   2. 对 ${CONFIG_FILENAME} 中缺失的配置字段逐项提问（已有字段不覆盖）。模板语言由 --locale、已有 locale 或默认 zh-CN 决定。
      传 --yes 时所有缺失字段使用各自默认值。
   3. 默认仅补齐 .Knowledge 缺失模板，并对路由清单做包级/结构增量对齐（manifest-routing + matcherPath 分片；关键词仅写在 matchers/*.json）；不替代 f2s-* 对业务文档与路由内容的写入。
      传 --reset-knowledge 时才会强制用模板覆盖 .Knowledge 中模板承载部分。
-  4. 在各 agent 配置根写入 rules、skills（Claude 规则自动转 .md；Codex 在仓库根写入完整 AGENTS.md，.codex/ 写入指针）。
-     Claude 额外写入 .claude/hooks/f2s-config-session.js、f2s-config-inject.js 与 .claude/settings.json：
-     SessionStart 注入一次配置摘要；PreToolUse 仅在调用 f2s-* Skill 时提示首步必须 Read 配置。
-     Cursor 额外写入 f2s-config-check.mdc（alwaysApply），强制在技能首步读取配置文件；
-     并写入 .cursor/hooks.json，在 sessionStart 自动检测知识库版本。
-     Codex：仓库根 AGENTS.md（完整条令）；.codex/AGENTS.md 为指针；
-     并写入 .codex/hooks.json，在 SessionStart 注入配置摘要并检测知识库版本。
+  4. 在各 agent 配置根写入对应的 rules、skills、入口和 hooks（若该客户端支持）；具体落盘方式以客户端适配器为准。
   5. 每次 init 将当前 locale 包模板 knowledge/index.md 复制到 .Knowledge/template/index.template.md，供 f2s-kb-upgrade 技能与 .Knowledge/index.md 对照；不自动改写 index.md。（「知识库升级」指 f2s-kb-upgrade 技能，init 本身不是升级命令。）
   6. 非破坏式补充 .gitignore：忽略 .task/ 与 .Knowledge/update-check.json 这类本地运行态。
   7. 规则与技能在各 agent 配置根加载；其他模版类文件在 .Knowledge/template/ 等目录。
@@ -795,6 +790,8 @@ if (sub === "init") {
         const { root, label } = AGENTS[id];
         if (id === "codex")
           return `  - ${root}/：（${label}）skills/、topics/、hooks/、hooks.json、AGENTS.md（指针）；仓库根 AGENTS.md（完整）`;
+        if (id === "dsh")
+          return `  - ${root}/：（${label}）skills/、topics/、AGENTS.md（指针）；根 AGENTS.md（缺少时生成）`;
         if (id === "claude") {
           const hookLine = claudeHooksResult?.settingsChanged
             ? "rules/、skills/、hooks/f2s-config-session.js、hooks/f2s-config-inject.js、settings.json（已写入 f2s SessionStart/PreToolUse hooks）"

diff --git a/docs/Flow2Spec基础介绍.md b/docs/Flow2Spec基础介绍.md
@@ -415,7 +415,7 @@ Flow2Spec 提供三个互补命令，按触发方式与粒度区分：
 
 <p><img src="./images/flow2spec-intro-13-quick-start.jpg" alt="适合什么项目与快速体验" style="max-width:720px;width:100%;" /></p>
 
-**最适合**：中大型业务项目、长期维护的代码仓库、多人协作规则很多的项目、经常使用 Cursor / Claude Code / Codex 的团队、希望 AI 不只是"看文档"而是参与维护项目知识的场景。
+**最适合**：中大型业务项目、长期维护的代码仓库、多人协作规则很多的项目、经常使用多个 AI 编程客户端的团队、希望 AI 不只是"看文档"而是参与维护项目知识的场景。
 
 **可能不适合**：一次性脚本或非常小的个人项目（\< 5000 行，一份 README 够用）。
 
@@ -442,7 +442,7 @@ npx @double-coding/flow2spec@latest init
 
 ```
 
-目前已支持 Cursor、Claude Code、Codex 三端初始化，也支持中文 / 英文模板。
+目前支持按客户端初始化，也支持中文 / 英文模板；新增客户端可以沿用同一套知识库与技能分发模型。
 
 ---
 

diff --git a/docs/en/Flow2Spec-Introduction.md b/docs/en/Flow2Spec-Introduction.md
@@ -431,7 +431,7 @@ In short: **single Q&A → distill (auto), new module bulk import → add, perio
 
 <p><img src="../images/flow2spec-intro-13-quick-start.jpg" alt="What projects and quick start" style="max-width:720px;width:100%;" /></p>
 
-**Best for**: medium to large business projects, long-lived codebases, multi-person teams with many rules, teams using Cursor / Claude Code / Codex regularly, and projects where you want AI to participate in maintaining project knowledge — not just read docs.
+**Best for**: medium to large business projects, long-lived codebases, multi-person teams with many rules, teams using multiple AI coding clients, and projects where you want AI to participate in maintaining project knowledge — not just read docs.
 
 **May not fit**: one-off scripts or very small personal projects (\< 5000 lines, a README is enough).
 
@@ -458,7 +458,7 @@ Common workflows:
 
 ```
 
-Currently supports initialization for Cursor, Claude Code, and Codex, with both Chinese and English templates.
+Initialization is client-selectable and supports both Chinese and English templates; new clients can reuse the same knowledge-base and skill-distribution model.
 
 ---
 

diff --git a/docs/en/architecture.md b/docs/en/architecture.md
@@ -23,7 +23,7 @@ Four rings in the repo (do not collapse rules + skills into a single "third ring
 | --- | --- | --- |
 | **Knowledge** | `.Knowledge/` | Routing, topics, stock/req docs (§2 layers) |
 | **Task** | `.task/` | `todo.json`, checklists, user todos |
-| **Rules** | `.cursor/.claude/.codex` rules, `AGENTS.md` | Read order, gap gates, execution constraints |
+| **Rules** | `.cursor/.claude/.codex/.dsh` rules/topics, `AGENTS.md` | Read order, gap gates, execution constraints |
 | **Skills** | config root `skills/*/SKILL.md` | `f2s-kb-feat/fix/sync`, etc. |
 
 ---
@@ -51,7 +51,7 @@ The pipeline `match → expand → verify → act` operates on L0–L2 (and L3 w
 | Layer | Location | Role |
 | --- | --- | --- |
 | Knowledge layer (knowledge ring) | `.Knowledge/` | Business docs, index, routing (§2 layers) |
-| Execution layer (rules + skills rings) | `.cursor/.claude/.codex` | Rules and skill entry points |
+| Execution layer (rules + skills rings) | `.cursor/.claude/.codex/.dsh` | Rules and skill entry points |
 
 ---
 

diff --git a/docs/en/design-principles.md b/docs/en/design-principles.md
@@ -556,9 +556,9 @@ Each SKILL.md's orchestration section reads:
 #### 1. Tools are pluggable: one knowledge base, any tool combination
 
 ```
-flow2spec init cursor claude codex   ← all three tools installed
-flow2spec init claude                ← only Claude
-flow2spec init cursor codex          ← skip Claude
+flow2spec init <selected-agents>     ← initialize the requested tools
+flow2spec init <one-agent>           ← initialize one tool
+flow2spec init <selected-agents>     ← add or skip tools as needed
 
 .Knowledge/ stays the same, tools can be added or removed at any time
 ```

diff --git a/docs/en/directory-conventions.md b/docs/en/directory-conventions.md
@@ -6,7 +6,7 @@
 
 - `.Knowledge/`: **Knowledge ring** — business docs and machine-readable routing ([architecture.md §2](./architecture.md))
 - `.task/`: **Task ring** — change tracking (not inside `.Knowledge/`)
-- `Config Root` (`.cursor/.claude/.codex`): **Rules ring + skills ring**
+- `Config Root` (`.cursor/.claude/.codex/.dsh`): **Rules ring + skills ring**
 
 See [architecture.md §1](./architecture.md) for Memory Coding four rings.
 
@@ -29,6 +29,8 @@ See [architecture.md §1](./architecture.md) for Memory Coding four rings.
 | `Config Root/skills/` | Skill definitions (`SKILL.md`) |
 | `Config Root/template/` | (Deprecated) No longer written to; historical directories may be cleaned up |
 | `.codex/AGENTS.md` | Codex unified entry point and loading instructions |
+| `.dsh/AGENTS.md` | DeepSeek Harness directory pointer; the complete project entry remains root `AGENTS.md` |
+| `.dsh/topics/` | Long-form rule mirrors loaded on demand by DeepSeek Harness |
 | `flow2spec.config.json` | Project root configuration, controls `subAgent`, `switchAgentVerification`, `changeTracking` (nested object with `feat` / `fix` / `implement` sub-items) |
 
 > See [Usage Guide Section 1](./usage-guide.md) for multi-platform references and path tables (detail maintained in a single table); **the authoritative source remains `Read(flow2spec.config.json)`**.