# egc - Doramagic AI Context Pack

> 定位：安装前体验与判断资产。它帮助宿主 AI 有一个好的开始，但不代表已经安装、执行或验证目标项目。

## 充分原则

- **充分原则，不是压缩原则**：AI Context Pack 应该充分到让宿主 AI 在开工前理解项目价值、能力边界、使用入口、风险和证据来源；它可以分层组织，但不以最短摘要为目标。
- **压缩策略**：只压缩噪声和重复内容，不压缩会影响判断和开工质量的上下文。

## 给宿主 AI 的使用方式

你正在读取 Doramagic 为 egc 编译的 AI Context Pack。请把它当作开工前上下文：帮助用户理解适合谁、能做什么、如何开始、哪些必须安装后验证、风险在哪里。不要声称你已经安装、运行或执行了目标项目。

## Claim 消费规则

- **事实来源**：Repo Evidence + Claim/Evidence Graph；Human Wiki 只提供显著性、术语和叙事结构。
- **事实最低状态**：`supported`
- `supported`：可以作为项目事实使用，但回答中必须引用 claim_id 和证据路径。
- `weak`：只能作为低置信度线索，必须要求用户继续核实。
- `inferred`：只能用于风险提示或待确认问题，不能包装成项目事实。
- `unverified`：不得作为事实使用，应明确说证据不足。
- `contradicted`：必须展示冲突来源，不得替用户强行选择一个版本。

## 它最适合谁

- **正在使用 Claude/Codex/Cursor/Gemini 等宿主 AI 的开发者**：README 或插件配置提到多个宿主 AI。 证据：`README.md` Claim：`clm_0004` supported 0.86
- **希望把专业流程带进宿主 AI 的用户**：仓库包含 Skill 文档。 证据：`.agents/skills/agent-introspection-debugging/SKILL.md`, `.agents/skills/agent-sort/SKILL.md`, `.agents/skills/api-design/SKILL.md`, `.agents/skills/article-writing/SKILL.md` 等 Claim：`clm_0005` supported 0.86

## 它能做什么

- **AI Skill / Agent 指令资产库**（可做安装前预览）：项目包含可被宿主 AI 读取的 Skill 或 Agent 指令文件，可用于把专业流程带入 Claude、Codex、Cursor 等宿主。 证据：`.agents/skills/agent-introspection-debugging/SKILL.md`, `.agents/skills/agent-sort/SKILL.md`, `.agents/skills/api-design/SKILL.md`, `.agents/skills/article-writing/SKILL.md` 等 Claim：`clm_0001` supported 0.86
- **多宿主安装与分发**（需要安装后验证）：项目包含插件或 marketplace 配置，说明它面向一个或多个 AI 宿主的安装和分发。 证据：`.agents/plugins/marketplace.json`, `.codex-plugin/plugin.json`, `.gemini-plugin/marketplace.json`, `.gemini-plugin/plugin.json` Claim：`clm_0002` unverified 0.25
- **命令行启动或安装流程**（需要安装后验证）：项目文档中存在可执行命令，真实使用需要在本地或宿主环境中运行这些命令。 证据：`README.md` Claim：`clm_0003` supported 0.86

## 怎么开始

- `npm install -g @egchq/egc && egc install` 证据：`README.md` Claim：`clm_0006` supported 0.86
- `npx @egchq/egc install` 证据：`README.md` Claim：`clm_0007` supported 0.86
- `npm install -g @egchq/egc` 证据：`README.md` Claim：`clm_0006` supported 0.86, `clm_0008` supported 0.86

## 继续前判断卡

- **当前建议**：先做流程沙盒试用
- **为什么**：这个项目会改变宿主 AI 的开发流程和规则，值得继续试，但不要直接装进主力 Claude/Cursor/Codex；先用临时宿主或隔离目录验证。

### 30 秒判断

- **现在怎么做**：先做流程沙盒试用
- **最小安全下一步**：先用 Prompt Preview 感受流程约束；满意后再临时宿主试装
- **先别相信**：这套流程是否适合你的工作方式不能直接相信。
- **继续会触碰**：宿主行为改变、命令执行、宿主 AI 配置

### 现在可以相信

- **适合人群线索：正在使用 Claude/Codex/Cursor/Gemini 等宿主 AI 的开发者**（supported）：有 supported claim 或项目证据支撑，但仍不等于真实安装效果。 证据：`README.md` Claim：`clm_0004` supported 0.86
- **适合人群线索：希望把专业流程带进宿主 AI 的用户**（supported）：有 supported claim 或项目证据支撑，但仍不等于真实安装效果。 证据：`.agents/skills/agent-introspection-debugging/SKILL.md`, `.agents/skills/agent-sort/SKILL.md`, `.agents/skills/api-design/SKILL.md`, `.agents/skills/article-writing/SKILL.md` 等 Claim：`clm_0005` supported 0.86
- **能力存在：AI Skill / Agent 指令资产库**（supported）：可以相信项目包含这类能力线索；是否适合你的具体任务仍要试用或安装后验证。 证据：`.agents/skills/agent-introspection-debugging/SKILL.md`, `.agents/skills/agent-sort/SKILL.md`, `.agents/skills/api-design/SKILL.md`, `.agents/skills/article-writing/SKILL.md` 等 Claim：`clm_0001` supported 0.86
- **能力存在：命令行启动或安装流程**（supported）：可以相信项目包含这类能力线索；是否适合你的具体任务仍要试用或安装后验证。 证据：`README.md` Claim：`clm_0003` supported 0.86
- **存在 Quick Start / 安装命令线索**（supported）：可以相信项目文档出现过启动或安装入口；不要因此直接在主力环境运行。 证据：`README.md` Claim：`clm_0006` supported 0.86

### 现在还不能相信

- **这套流程是否适合你的工作方式不能直接相信。**（unverified）：流程型 Skill 可能强约束 AI 行为；它能提升纪律，也可能拖慢你当前任务节奏。 证据：`.agents/plugins/marketplace.json`, `.agents/skills/agent-introspection-debugging/SKILL.md`, `.agents/skills/agent-sort/SKILL.md`, `.agents/skills/api-design/SKILL.md` 等
- **不会和你已有 Claude/Cursor/Codex 规则冲突，不能直接相信。**（inferred）：开发流程 Skill 会改变澄清、计划、测试、验证等默认行为，必须在临时宿主里试。 证据：`.agents/plugins/marketplace.json`, `.agents/skills/agent-introspection-debugging/SKILL.md`, `.agents/skills/agent-sort/SKILL.md`, `.agents/skills/api-design/SKILL.md` 等
- **真实输出质量不能在安装前相信。**（unverified）：Prompt Preview 只能展示引导方式，不能证明真实项目中的结果质量。
- **宿主 AI 版本兼容性不能在安装前相信。**（unverified）：Claude、Cursor、Codex、Gemini 等宿主加载规则和版本差异必须在真实环境验证。
- **不会污染现有宿主 AI 行为，不能直接相信。**（inferred）：Skill、plugin、AGENTS/CLAUDE/GEMINI 指令可能改变宿主 AI 的默认行为。 证据：`.agents/plugins/marketplace.json`, `.agents/skills/agent-introspection-debugging/SKILL.md`, `.agents/skills/agent-sort/SKILL.md`, `.agents/skills/api-design/SKILL.md` 等
- **可安全回滚不能默认相信。**（unverified）：除非项目明确提供卸载和恢复说明，否则必须先在隔离环境验证。
- **真实安装后是否与用户当前宿主 AI 版本兼容？**（unverified）：兼容性只能通过实际宿主环境验证。 证据：`.agents/plugins/marketplace.json`, `.codex-plugin/plugin.json`, `.gemini-plugin/marketplace.json`, `.gemini-plugin/plugin.json`
- **项目输出质量是否满足用户具体任务？**（unverified）：安装前预览只能展示流程和边界，不能替代真实评测。

### 继续会触碰什么

- **宿主行为改变**：澄清、计划、TDD、验证、收尾等默认开发节奏。 原因：这类 Skill 的价值和风险都来自强约束流程；必须先确认你愿意被它改变工作方式。 证据：`.agents/plugins/marketplace.json`, `.agents/skills/agent-introspection-debugging/SKILL.md`, `.agents/skills/agent-sort/SKILL.md`, `.agents/skills/api-design/SKILL.md` 等
- **命令执行**：包管理器、网络下载、本地插件目录、项目配置或用户主目录。 原因：运行第一条命令就可能产生环境改动；必须先判断是否值得跑。 证据：`README.md`
- **宿主 AI 配置**：Claude/Codex/Cursor/Gemini/OpenCode 等宿主的 plugin、Skill 或规则加载配置。 原因：宿主配置会改变 AI 后续工作方式，可能和用户已有规则冲突。 证据：`.agents/plugins/marketplace.json`, `.agents/skills/agent-introspection-debugging/SKILL.md`, `.agents/skills/agent-sort/SKILL.md`, `.agents/skills/api-design/SKILL.md` 等
- **本地环境或项目文件**：安装结果、插件缓存、项目配置或本地依赖目录。 原因：安装前无法证明写入范围和回滚方式，需要隔离验证。 证据：`.agents/plugins/marketplace.json`, `.codex-plugin/plugin.json`, `.gemini-plugin/marketplace.json`, `.gemini-plugin/plugin.json` 等
- **宿主 AI 上下文**：AI Context Pack、Prompt Preview、Skill 路由、风险规则和项目事实。 原因：导入上下文会影响宿主 AI 后续判断，必须避免把未验证项包装成事实。

### 最小安全下一步

- **先跑 Prompt Preview**：先感受它会怎样改变 AI 的开发节奏，再决定是否让它进入真实宿主。（适用：任何项目都适用，尤其是输出质量未知时。）
- **只在隔离目录或测试账号试装**：避免安装命令污染主力宿主 AI、真实项目或用户主目录。（适用：存在命令执行、插件配置或本地写入线索时。）
- **先备份宿主 AI 配置**：Skill、plugin、规则文件可能改变 Claude/Cursor/Codex 的默认行为。（适用：存在插件 manifest、Skill 或宿主规则入口时。）
- **安装后只验证一个最小任务**：先验证加载、兼容、输出质量和回滚，再决定是否深用。（适用：准备从试用进入真实工作流时。）

### 退出方式

- **保留安装前状态**：记录原始宿主配置和项目状态，后续才能判断是否可恢复。
- **准备移除宿主 plugin / Skill / 规则入口**：如果试装后行为异常，可以把宿主 AI 恢复到试装前状态。
- **记录安装命令和写入路径**：没有明确卸载说明时，至少要知道哪些目录或配置需要手动清理。
- **如果没有回滚路径，不进入主力环境**：不可回滚是继续前阻断项，不应靠信任或运气继续。

## 哪些只能预览

- 解释项目适合谁和能做什么
- 基于项目文档演示典型对话流程
- 帮助用户判断是否值得安装或继续研究

## 哪些必须安装后验证

- 真实安装 Skill、插件或 CLI
- 执行脚本、修改本地文件或访问外部服务
- 验证真实输出质量、性能和兼容性

## 边界与风险判断卡

- **把安装前预览误认为真实运行**：用户可能高估项目已经完成的配置、权限和兼容性验证。 处理方式：明确区分 prompt_preview_can_do 与 runtime_required。 Claim：`clm_0009` inferred 0.45
- **宿主 AI 插件或 Skill 规则冲突**：新规则可能改变用户现有宿主 AI 的工作方式。 处理方式：安装前先检查插件 manifest 和 Skill 文件，必要时隔离测试。 证据：`.agents/plugins/marketplace.json`, `.codex-plugin/plugin.json`, `.gemini-plugin/marketplace.json`, `.gemini-plugin/plugin.json` Claim：`clm_0010` inferred 0.45
- **命令执行会修改本地环境**：安装命令可能写入用户主目录、宿主插件目录或项目配置。 处理方式：先在隔离环境或测试账号中运行。 证据：`README.md` Claim：`clm_0011` supported 0.86, `clm_0012` supported 0.86, `clm_0026` contradicted 0.20
- **源文档冲突：skill_count**：项目文档中存在数量表述不一致，AI Context Pack 必须提示用户不要把单一数字当作已验证事实。 处理方式：在 Human Manual 和 AI Context Pack 中共同标记为待核实，而不是强行选择一个数字。 证据：`skills/general/configure-egc/SKILL.md`, `manifests/install-components.json`, `.codex-plugin/README.md`, `docs/installation.md` 等 Claim：`clm_0011` supported 0.86, `clm_0012` supported 0.86, `clm_0026` contradicted 0.20
- **源文件冲突 skill_count**：发现多个值 `45, 81, 182, 229, 230`，应在真实使用前核实。
- **待确认**：真实安装后是否与用户当前宿主 AI 版本兼容？。原因：兼容性只能通过实际宿主环境验证。
- **待确认**：项目输出质量是否满足用户具体任务？。原因：安装前预览只能展示流程和边界，不能替代真实评测。
- **待确认**：安装命令是否需要网络、权限或全局写入？。原因：这影响企业环境和个人环境的安装风险。

## 开工前工作上下文

### 加载顺序

- 先读取 how_to_use.host_ai_instruction，建立安装前判断资产的边界。
- 读取 claim_graph_summary，确认事实来自 Claim/Evidence Graph，而不是 Human Wiki 叙事。
- 再读取 intended_users、capabilities 和 quick_start_candidates，判断用户是否匹配。
- 需要执行具体任务时，优先查 role_skill_index，再查 evidence_index。
- 遇到真实安装、文件修改、网络访问、性能或兼容性问题时，转入 risk_card 和 boundaries.runtime_required。

### 任务路由

- **AI Skill / Agent 指令资产库**：先基于 role_skill_index / evidence_index 帮用户挑选可用角色、Skill 或工作流。 边界：可做安装前 Prompt 体验。 证据：`.agents/skills/agent-introspection-debugging/SKILL.md`, `.agents/skills/agent-sort/SKILL.md`, `.agents/skills/api-design/SKILL.md`, `.agents/skills/article-writing/SKILL.md` 等 Claim：`clm_0001` supported 0.86
- **多宿主安装与分发**：先说明这是安装后验证能力，再给出安装前检查清单。 边界：必须真实安装或运行后验证。 证据：`.agents/plugins/marketplace.json`, `.codex-plugin/plugin.json`, `.gemini-plugin/marketplace.json`, `.gemini-plugin/plugin.json`
- **命令行启动或安装流程**：先说明这是安装后验证能力，再给出安装前检查清单。 边界：必须真实安装或运行后验证。 证据：`README.md` Claim：`clm_0003` supported 0.86

### 上下文规模

- 文件总数：1459
- 重要文件覆盖：40/1459
- 证据索引条目：180
- 角色 / Skill 条目：260

### 证据不足时的处理

- **missing_evidence**：说明证据不足，要求用户提供目标文件、README 段落或安装后验证记录；不要补全事实。
- **out_of_scope_request**：说明该任务超出当前 AI Context Pack 证据范围，并建议用户先查看 Human Manual 或真实安装后验证。
- **runtime_request**：给出安装前检查清单和命令来源，但不要替用户执行命令或声称已执行。
- **source_conflict**：同时展示冲突来源，标记为待核实，不要强行选择一个版本。

## Prompt Recipes

### 适配判断

- 目标：判断这个项目是否适合用户当前任务。
- 预期输出：适配结论、关键理由、证据引用、安装前可预览内容、必须安装后验证内容、下一步建议。

```text
请基于 egc 的 AI Context Pack，先问我 3 个必要问题，然后判断它是否适合我的任务。回答必须包含：适合谁、能做什么、不能做什么、是否值得安装、证据来自哪里。所有项目事实必须引用 evidence_refs、source_paths 或 claim_id。
```

### 安装前体验

- 目标：让用户在安装前感受核心工作流，同时避免把预览包装成真实能力或营销承诺。
- 预期输出：一段带边界标签的体验剧本、安装后验证清单和谨慎建议；不含真实运行承诺或强营销表述。

```text
请把 egc 当作安装前体验资产，而不是已安装工具或真实运行环境。

请严格输出四段：
1. 先问我 3 个必要问题。
2. 给出一段“体验剧本”：用 [安装前可预览]、[必须安装后验证]、[证据不足] 三种标签展示它可能如何引导工作流。
3. 给出安装后验证清单：列出哪些能力只有真实安装、真实宿主加载、真实项目运行后才能确认。
4. 给出谨慎建议：只能说“值得继续研究/试装”“先补充信息后再判断”或“不建议继续”，不得替项目背书。

硬性边界：
- 不要声称已经安装、运行、执行测试、修改文件或产生真实结果。
- 不要写“自动适配”“确保通过”“完美适配”“强烈建议安装”等承诺性表达。
- 如果描述安装后的工作方式，必须使用“如果安装成功且宿主正确加载 Skill，它可能会……”这种条件句。
- 体验剧本只能写成“示例台词/假设流程”：使用“可能会询问/可能会建议/可能会展示”，不要写“已写入、已生成、已通过、正在运行、正在生成”。
- Prompt Preview 不负责给安装命令；如用户准备试装，只能提示先阅读 Quick Start 和 Risk Card，并在隔离环境验证。
- 所有项目事实必须来自 supported claim、evidence_refs 或 source_paths；inferred/unverified 只能作风险或待确认项。

```

### 角色 / Skill 选择

- 目标：从项目里的角色或 Skill 中挑选最匹配的资产。
- 预期输出：候选角色或 Skill 列表，每项包含适用场景、证据路径、风险边界和是否需要安装后验证。

```text
请读取 role_skill_index，根据我的目标任务推荐 3-5 个最相关的角色或 Skill。每个推荐都要说明适用场景、可能输出、风险边界和 evidence_refs。
```

### 风险预检

- 目标：安装或引入前识别环境、权限、规则冲突和质量风险。
- 预期输出：环境、权限、依赖、许可、宿主冲突、质量风险和未知项的检查清单。

```text
请基于 risk_card、boundaries 和 quick_start_candidates，给我一份安装前风险预检清单。不要替我执行命令，只说明我应该检查什么、为什么检查、失败会有什么影响。
```

### 宿主 AI 开工指令

- 目标：把项目上下文转成一次对话开始前的宿主 AI 指令。
- 预期输出：一段边界明确、证据引用明确、适合复制给宿主 AI 的开工前指令。

```text
请基于 egc 的 AI Context Pack，生成一段我可以粘贴给宿主 AI 的开工前指令。这段指令必须遵守 not_runtime=true，不能声称项目已经安装、运行或产生真实结果。
```

## 角色 / Skill 索引

- 共索引 260 个角色 / Skill / 项目文档条目。

- **agent-introspection-debugging**（skill）：Structured self-debugging workflow for AI agent failures using capture, diagnosis, contained recovery, and introspection reports. 激活提示：当用户任务与“agent-introspection-debugging”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/agent-introspection-debugging/SKILL.md`
- **agent-sort**（skill）：Build an evidence-backed EGC install plan for a specific repo by sorting skills, commands, rules, hooks, and extras into DAILY vs LIBRARY buckets using parallel repo-aware review passes. Use when EGC should be trimmed to what a project actually needs instead of loading the full bundle. 激活提示：当用户任务与“agent-sort”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/agent-sort/SKILL.md`
- **api-design**（skill）：REST API design patterns including resource naming, status codes, pagination, filtering, error responses, versioning, and rate limiting for production APIs. 激活提示：当用户任务与“api-design”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/api-design/SKILL.md`
- **article-writing**（skill）：Write articles, guides, blog posts, tutorials, newsletter issues, and other long-form content in a distinctive voice derived from supplied examples or brand guidance. Use when the user wants polished written content longer than a paragraph, especially when voice consistency, structure, and credibility matter. 激活提示：当用户任务与“article-writing”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/article-writing/SKILL.md`
- **backend-patterns**（skill）：Backend architecture patterns, API design, database optimization, and server-side best practices for Node.js, Express, and Next.js API routes. 激活提示：当用户任务与“backend-patterns”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/backend-patterns/SKILL.md`
- **brand-voice**（skill）：Build a source-derived writing style profile from real posts, essays, launch notes, docs, or site copy, then reuse that profile across content, outreach, and social workflows. Use when the user wants voice consistency without generic AI writing tropes. 激活提示：当用户任务与“brand-voice”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/brand-voice/SKILL.md`
- **bun-runtime**（skill）：Bun as runtime, package manager, bundler, and test runner. When to choose Bun vs Node, migration notes, and Vercel support. 激活提示：当用户任务与“bun-runtime”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/bun-runtime/SKILL.md`
- **coding-standards**（skill）：Baseline cross-project coding conventions for naming, readability, immutability, and code-quality review. Use detailed frontend or backend skills for framework-specific patterns. 激活提示：当用户任务与“coding-standards”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/coding-standards/SKILL.md`
- **content-engine**（skill）：Create platform-native content systems for X, LinkedIn, TikTok, YouTube, newsletters, and repurposed multi-platform campaigns. Use when the user wants social posts, threads, scripts, content calendars, or one source asset adapted cleanly across platforms. 激活提示：当用户任务与“content-engine”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/content-engine/SKILL.md`
- **crosspost**（skill）：Multi-platform content distribution across X, LinkedIn, Threads, and Bluesky. Adapts content per platform using content-engine patterns. Never posts identical content cross-platform. Use when the user wants to distribute content across social platforms. 激活提示：当用户任务与“crosspost”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/crosspost/SKILL.md`
- **deep-research**（skill）：Multi-source deep research using firecrawl and exa MCPs. Searches the web, synthesizes findings, and delivers cited reports with source attribution. Use when the user wants thorough research on any topic with evidence and citations. 激活提示：当用户任务与“deep-research”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/deep-research/SKILL.md`
- **dmux-workflows**（skill）：Multi-agent orchestration using dmux tmux pane manager for AI agents . Patterns for parallel agent workflows across Gemini Code, Codex, OpenCode, and other harnesses. Use when running multiple agent sessions in parallel or coordinating multi-agent development workflows. 激活提示：当用户任务与“dmux-workflows”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/dmux-workflows/SKILL.md`
- **documentation-lookup**（skill）：Use up-to-date library and framework docs via Context7 MCP instead of training data. Activates for setup questions, API references, code examples, or when the user names a framework e.g. React, Next.js, Prisma . 激活提示：当用户任务与“documentation-lookup”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/documentation-lookup/SKILL.md`
- **e2e-testing**（skill）：Playwright E2E testing patterns, Page Object Model, configuration, CI/CD integration, artifact management, and flaky test strategies. 激活提示：当用户任务与“e2e-testing”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/e2e-testing/SKILL.md`
- **egc**（skill）：Development conventions and patterns for EGC - Extended Global Context. JavaScript project with conventional commits. 激活提示：当用户任务与“egc”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/egc/SKILL.md`
- **eval-harness**（skill）：Formal evaluation framework for Gemini Code sessions implementing eval-driven development EDD principles 激活提示：当用户任务与“eval-harness”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/eval-harness/SKILL.md`
- **exa-search**（skill）：Neural search via Exa MCP for web, code, and company research. Use when the user needs web search, code examples, company intel, people lookup, or AI-powered deep research with Exa's neural search engine. 激活提示：当用户任务与“exa-search”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/exa-search/SKILL.md`
- **fal-ai-media**（skill）：Unified media generation via fal.ai MCP: image, video, and audio. Covers text-to-image Nano Banana , text/image-to-video Seedance, Kling, Veo 3 , text-to-speech CSM-1B , and video-to-audio ThinkSound . Use when the user wants to generate images, videos, or audio with AI. 激活提示：当用户任务与“fal-ai-media”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/fal-ai-media/SKILL.md`
- **frontend-patterns**（skill）：Frontend development patterns for React, Next.js, state management, performance optimization, and UI best practices. 激活提示：当用户任务与“frontend-patterns”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/frontend-patterns/SKILL.md`
- **frontend-slides**（skill）：Create stunning, animation-rich HTML presentations from scratch or by converting PowerPoint files. Use when the user wants to build a presentation, convert a PPT/PPTX to web, or create slides for a talk/pitch. Helps non-designers discover their aesthetic through visual exploration rather than abstract choices. 激活提示：当用户任务与“frontend-slides”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/frontend-slides/SKILL.md`
- **investor-materials**（skill）：Create and update pitch decks, one-pagers, investor memos, accelerator applications, financial models, and fundraising materials. Use when the user needs investor-facing documents, projections, use-of-funds tables, milestone plans, or materials that must stay internally consistent across multiple fundraising assets. 激活提示：当用户任务与“investor-materials”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/investor-materials/SKILL.md`
- **investor-outreach**（skill）：Draft cold emails, warm intro blurbs, follow-ups, update emails, and investor communications for fundraising. Use when the user wants outreach to angels, VCs, strategic investors, or accelerators and needs concise, personalized, investor-facing messaging. 激活提示：当用户任务与“investor-outreach”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/investor-outreach/SKILL.md`
- **market-research**（skill）：Conduct market research, competitive analysis, investor due diligence, and industry intelligence with source attribution and decision-oriented summaries. Use when the user wants market sizing, competitor comparisons, fund research, technology scans, or research that informs business decisions. 激活提示：当用户任务与“market-research”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/market-research/SKILL.md`
- **mcp-server-patterns**（skill）：Build MCP servers with Node/TypeScript SDK: tools, resources, prompts, Zod validation, stdio vs Streamable HTTP. Use Context7 or official MCP docs for latest API. 激活提示：当用户任务与“mcp-server-patterns”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/mcp-server-patterns/SKILL.md`
- **nextjs-turbopack**（skill）：Next.js 16+ and Turbopack: incremental bundling, FS caching, dev speed, and when to use Turbopack vs webpack. 激活提示：当用户任务与“nextjs-turbopack”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/nextjs-turbopack/SKILL.md`
- **product-capability**（skill）：Translate PRD intent, roadmap asks, or product discussions into an implementation-ready capability plan that exposes constraints, invariants, interfaces, and unresolved decisions before multi-service work starts. Use when the user needs an EGC-native PRD-to-SRS lane instead of vague planning prose. 激活提示：当用户任务与“product-capability”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/product-capability/SKILL.md`
- **security-review**（skill）：Use this skill when adding authentication, handling user input, working with secrets, creating API endpoints, or implementing payment/sensitive features. Provides comprehensive security checklist and patterns. 激活提示：当用户任务与“security-review”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/security-review/SKILL.md`
- **strategic-compact**（skill）：Suggests manual context compaction at logical intervals to preserve context through task phases rather than arbitrary auto-compaction. 激活提示：当用户任务与“strategic-compact”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/strategic-compact/SKILL.md`
- **tdd-workflow**（skill）：Use this skill when writing new features, fixing bugs, or refactoring code. Enforces test-driven development with 80%+ coverage including unit, integration, and E2E tests. 激活提示：当用户任务与“tdd-workflow”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/tdd-workflow/SKILL.md`
- **verification-loop**（skill）：A comprehensive verification system for Gemini Code sessions. 激活提示：当用户任务与“verification-loop”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/verification-loop/SKILL.md`
- **video-editing**（skill）：AI-assisted video editing workflows for cutting, structuring, and augmenting real footage. Covers the full pipeline from raw capture through FFmpeg, Remotion, ElevenLabs, fal.ai, and final polish in Descript or CapCut. Use when the user wants to edit video, cut footage, create vlogs, or build video content. 激活提示：当用户任务与“video-editing”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/video-editing/SKILL.md`
- **x-api**（skill）：X/Twitter API integration for posting tweets, threads, reading timelines, search, and analytics. Covers OAuth auth patterns, rate limits, and platform-native content posting. Use when the user wants to interact with X programmatically. 激活提示：当用户任务与“x-api”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.agents/skills/x-api/SKILL.md`
- **article-writing**（skill）：Write articles, guides, blog posts, tutorials, newsletter issues, and other long-form content in a distinctive voice derived from supplied examples or brand guidance. Use when the user wants polished written content longer than a paragraph, especially when voice consistency, structure, and credibility matter. 激活提示：当用户任务与“article-writing”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.cursor/skills/article-writing/SKILL.md`
- **bun-runtime**（skill）：Bun as runtime, package manager, bundler, and test runner. When to choose Bun vs Node, migration notes, and Vercel support. 激活提示：当用户任务与“bun-runtime”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.cursor/skills/bun-runtime/SKILL.md`
- **content-engine**（skill）：Create platform-native content systems for X, LinkedIn, TikTok, YouTube, newsletters, and repurposed multi-platform campaigns. Use when the user wants social posts, threads, scripts, content calendars, or one source asset adapted cleanly across platforms. 激活提示：当用户任务与“content-engine”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.cursor/skills/content-engine/SKILL.md`
- **documentation-lookup**（skill）：Use up-to-date library and framework docs via Context7 MCP instead of training data. Activates for setup questions, API references, code examples, or when the user names a framework e.g. React, Next.js, Prisma . 激活提示：当用户任务与“documentation-lookup”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.cursor/skills/documentation-lookup/SKILL.md`
- **frontend-slides**（skill）：Create stunning, animation-rich HTML presentations from scratch or by converting PowerPoint files. Use when the user wants to build a presentation, convert a PPT/PPTX to web, or create slides for a talk/pitch. Helps non-designers discover their aesthetic through visual exploration rather than abstract choices. 激活提示：当用户任务与“frontend-slides”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.cursor/skills/frontend-slides/SKILL.md`
- **investor-materials**（skill）：Create and update pitch decks, one-pagers, investor memos, accelerator applications, financial models, and fundraising materials. Use when the user needs investor-facing documents, projections, use-of-funds tables, milestone plans, or materials that must stay internally consistent across multiple fundraising assets. 激活提示：当用户任务与“investor-materials”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.cursor/skills/investor-materials/SKILL.md`
- **investor-outreach**（skill）：Draft cold emails, warm intro blurbs, follow-ups, update emails, and investor communications for fundraising. Use when the user wants outreach to angels, VCs, strategic investors, or accelerators and needs concise, personalized, investor-facing messaging. 激活提示：当用户任务与“investor-outreach”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.cursor/skills/investor-outreach/SKILL.md`
- **market-research**（skill）：Conduct market research, competitive analysis, investor due diligence, and industry intelligence with source attribution and decision-oriented summaries. Use when the user wants market sizing, competitor comparisons, fund research, technology scans, or research that informs business decisions. 激活提示：当用户任务与“market-research”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.cursor/skills/market-research/SKILL.md`
- **mcp-server-patterns**（skill）：Build MCP servers with Node/TypeScript SDK: tools, resources, prompts, Zod validation, stdio vs Streamable HTTP. Use Context7 or official MCP docs for latest API. 激活提示：当用户任务与“mcp-server-patterns”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.cursor/skills/mcp-server-patterns/SKILL.md`
- **nextjs-turbopack**（skill）：Next.js 16+ and Turbopack: incremental bundling, FS caching, dev speed, and when to use Turbopack vs webpack. 激活提示：当用户任务与“nextjs-turbopack”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.cursor/skills/nextjs-turbopack/SKILL.md`
- **agentic-engineering**（skill）： 激活提示：当用户任务与“agentic-engineering”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.kiro/skills/agentic-engineering/SKILL.md`
- **api-design**（skill）： 激活提示：当用户任务与“api-design”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.kiro/skills/api-design/SKILL.md`
- **backend-patterns**（skill）： 激活提示：当用户任务与“backend-patterns”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.kiro/skills/backend-patterns/SKILL.md`
- **coding-standards**（skill）： 激活提示：当用户任务与“coding-standards”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.kiro/skills/coding-standards/SKILL.md`
- **database-migrations**（skill）： 激活提示：当用户任务与“database-migrations”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.kiro/skills/database-migrations/SKILL.md`
- **deployment-patterns**（skill）： 激活提示：当用户任务与“deployment-patterns”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.kiro/skills/deployment-patterns/SKILL.md`
- **e2e-testing**（skill）： 激活提示：当用户任务与“e2e-testing”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.kiro/skills/e2e-testing/SKILL.md`
- **frontend-patterns**（skill）： 激活提示：当用户任务与“frontend-patterns”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.kiro/skills/frontend-patterns/SKILL.md`
- **golang-patterns**（skill）： 激活提示：当用户任务与“golang-patterns”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.kiro/skills/golang-patterns/SKILL.md`
- **golang-testing**（skill）： 激活提示：当用户任务与“golang-testing”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.kiro/skills/golang-testing/SKILL.md`
- **postgres-patterns**（skill）： 激活提示：当用户任务与“postgres-patterns”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.kiro/skills/postgres-patterns/SKILL.md`
- **python-patterns**（skill）： 激活提示：当用户任务与“python-patterns”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.kiro/skills/python-patterns/SKILL.md`
- **python-testing**（skill）： 激活提示：当用户任务与“python-testing”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.kiro/skills/python-testing/SKILL.md`
- **search-first**（skill）： 激活提示：当用户任务与“search-first”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.kiro/skills/search-first/SKILL.md`
- **security-review**（skill）： 激活提示：当用户任务与“security-review”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.kiro/skills/security-review/SKILL.md`
- **tdd-workflow**（skill）： 激活提示：当用户任务与“tdd-workflow”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.kiro/skills/tdd-workflow/SKILL.md`
- **verification-loop**（skill）： 激活提示：当用户任务与“verification-loop”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`.kiro/skills/verification-loop/SKILL.md`
- **agent-eval**（skill）：Head-to-head comparison of coding agents Gemini Code, Aider, Codex, etc. on custom tasks with pass rate, cost, time, and consistency metrics 激活提示：当用户任务与“agent-eval”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/ai/agent-eval/SKILL.md`
- **agent-harness-construction**（skill）：Design and optimize AI agent action spaces, tool definitions, and observation formatting for higher completion rates. 激活提示：当用户任务与“agent-harness-construction”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/ai/agent-harness-construction/SKILL.md`
- **agent-introspection-debugging**（skill）：Structured self-debugging workflow for AI agent failures using capture, diagnosis, contained recovery, and introspection reports. 激活提示：当用户任务与“agent-introspection-debugging”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/ai/agent-introspection-debugging/SKILL.md`
- **agent-sort**（skill）：Build an evidence-backed EGC install plan for a specific repo by sorting skills, commands, rules, hooks, and extras into DAILY vs LIBRARY buckets using parallel repo-aware review passes. Use when EGC should be trimmed to what a project actually needs instead of loading the full bundle. 激活提示：当用户任务与“agent-sort”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/ai/agent-sort/SKILL.md`
- **agentic-engineering**（skill）：Operate as an agentic engineer using eval-first execution, decomposition, and cost-aware model routing. 激活提示：当用户任务与“agentic-engineering”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/ai/agentic-engineering/SKILL.md`
- **ai-first-engineering**（skill）：Engineering operating model for teams where AI agents generate a large share of implementation output. 激活提示：当用户任务与“ai-first-engineering”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/ai/ai-first-engineering/SKILL.md`
- **autonomous-agent-harness**（skill）：Transform Gemini Code into a fully autonomous agent system with persistent memory, scheduled operations, computer use, and task queuing. Replaces standalone agent frameworks Hermes, AutoGPT by leveraging Gemini Code's native crons, dispatch, MCP tools, and memory. Use when the user wants continuous autonomous operation, scheduled tasks, or a self-directing agent loop. 激活提示：当用户任务与“autonomous-agent-harness”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/ai/autonomous-agent-harness/SKILL.md`
- **autonomous-lesson-learning**（skill）：Autonomous loop that learns while it runs: recalls stored lessons before each iteration, saves new lessons on failures and wins, and reinforces known lessons instead of repeating mistakes. Orchestrates continuous-agent-loop patterns with the egc-memory lesson tools. 激活提示：当用户任务与“autonomous-lesson-learning”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/ai/autonomous-lesson-learning/SKILL.md`
- **autonomous-loops**（skill）：Patterns and architectures for autonomous Gemini Code loops: from simple sequential pipelines to RFC-driven multi-agent DAG systems. 激活提示：当用户任务与“autonomous-loops”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/ai/autonomous-loops/SKILL.md`
- **continuous-learning-v2**（skill）：Instinct-based learning system that observes sessions via hooks, creates atomic instincts with confidence scoring, and evolves them into skills/commands/agents. v2.1 adds project-scoped instincts to prevent cross-project contamination. 激活提示：当用户任务与“continuous-learning-v2”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/ai/continuous-learning-v2/SKILL.md`
- **continuous-learning**（skill）：Automatically extract reusable patterns from Gemini Code sessions and save them as learned skills for future use. 激活提示：当用户任务与“continuous-learning”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/ai/continuous-learning/SKILL.md`
- **cost-aware-llm-pipeline**（skill）：Cost optimization patterns for LLM API usage: model routing by task complexity, budget tracking, retry logic, and prompt caching. 激活提示：当用户任务与“cost-aware-llm-pipeline”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/ai/cost-aware-llm-pipeline/SKILL.md`
- **deep-research**（skill）：Multi-source deep research using firecrawl and exa MCPs. Searches the web, synthesizes findings, and delivers cited reports with source attribution. Use when the user wants thorough research on any topic with evidence and citations. 激活提示：当用户任务与“deep-research”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/ai/deep-research/SKILL.md`
- **exa-search**（skill）：Neural search via Exa MCP for web, code, and company research. Use when the user needs web search, code examples, company intel, people lookup, or AI-powered deep research with Exa's neural search engine. 激活提示：当用户任务与“exa-search”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/ai/exa-search/SKILL.md`
- **fal-ai-media**（skill）：Unified media generation via fal.ai MCP: image, video, and audio. Covers text-to-image Nano Banana , text/image-to-video Seedance, Kling, Veo 3 , text-to-speech CSM-1B , and video-to-audio ThinkSound . Use when the user wants to generate images, videos, or audio with AI. 激活提示：当用户任务与“fal-ai-media”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/ai/fal-ai-media/SKILL.md`
- **prompt-optimizer**（skill）：- 激活提示：当用户任务与“prompt-optimizer”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/ai/prompt-optimizer/SKILL.md`
- **regex-vs-llm-structured-text**（skill）：Decision framework for choosing between regex and LLM when parsing structured text: start with regex, add LLM only for low-confidence edge cases. 激活提示：当用户任务与“regex-vs-llm-structured-text”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/ai/regex-vs-llm-structured-text/SKILL.md`
- **token-budget-advisor**（skill）：- 激活提示：当用户任务与“token-budget-advisor”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/ai/token-budget-advisor/SKILL.md`
- **agent-architecture-audit**（skill）：Full-stack diagnostic for agent and LLM applications. Audits the 12-layer agent stack for wrapper regression, memory pollution, tool discipline failures, hidden repair loops, and rendering corruption. Produces severity-ranked findings with code-first fixes. Essential for developers building agent applications, autonomous loops, or any LLM-powered feature. 激活提示：当用户任务与“agent-architecture-audit”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/architecture/agent-architecture-audit/SKILL.md`
- **agentic-os**（skill）：Build persistent multi-agent operating systems on Gemini CLI. Covers kernel architecture, specialist agents, slash commands, file-based memory, scheduled automation, and state management without external databases. 激活提示：当用户任务与“agentic-os”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/architecture/agentic-os/SKILL.md`
- **android-clean-architecture**（skill）：Clean Architecture patterns for Android and Kotlin Multiplatform projects: module structure, dependency rules, UseCases, Repositories, and data layer patterns. 激活提示：当用户任务与“android-clean-architecture”描述的流程高度相关时，先用它做安装前体验，再决定是否安装。 证据：`skills/architecture/android-clean-architecture/SKILL.md`
- 其余 180 个条目见 `AI_CONTEXT_PACK.json`。

## 证据索引

- 共索引 180 条证据。

- **EGC Documentation**（documentation）：This directory is the canonical entry point for everything that is not the user-facing README.md at the repository root. 证据：`docs/README.md`
- **EGC Architecture**（documentation）：EGC ships one production runtime alongside an exploratory kernel direction kept under architecture/ for research and ecosystem-evolution work. This page is the index: read it first, then drill into the specific documents below. 证据：`docs/architecture/README.md`
- **EGC Governance**（documentation）：Documents that define how subsystems are classified, how skills and agents are placed and adapted, and how capability surfaces are selected. Read SUBSYSTEM-MAP.md first. 证据：`docs/governance/README.md`
- **EGC Guides**（documentation）：Operational and contributor walk-throughs. These are tutorial-shaped documents; for architecture or governance see ../architecture/ and ../governance/ . 证据：`docs/guides/README.md`
- **VEX Documents**（documentation）：This directory contains Vulnerability Exploitability eXchange VEX documents for EGC. 证据：`docs/security/vex/README.md`
- **EGC Specification**（documentation）：The EGC specification is executable. It lives in JSON Schemas under schemas/ , in install manifests under scripts/lib/ , and in tests under tests/spec/ . This document is the index that ties them together. 证据：`docs/spec/README.md`
- **Documentacao do EGC**（documentation）：Este diretorio e o ponto de entrada canonico para tudo que nao e o README.md voltado ao usuario na raiz do repositorio. 证据：`translations/pt/docs/README.md`
- **Arquitetura do EGC**（documentation）：O EGC possui um runtime de producao alem de uma direcao exploratoria de kernel mantida em architecture/ para pesquisa e evolucao do ecossistema. Esta pagina e o indice: leia primeiro, depois aprofunde-se nos documentos especificos abaixo. 证据：`translations/pt/docs/architecture/README.md`
- **Governanca do EGC**（documentation）：Documentos que definem como os subsistemas sao classificados, como skills e agentes sao posicionados e adaptados, e como as superficies de capacidade sao selecionadas. Leia SUBSYSTEM-MAP.md primeiro. 证据：`translations/pt/docs/governance/README.md`
- **Guias do EGC**（documentation）：Guias operacionais e tutoriais para contribuidores. Estes sao documentos em formato tutorial; para arquitetura ou governanca, consulte ../architecture/ e ../governance/ . 证据：`translations/pt/docs/guides/README.md`
- **Documentos VEX**（documentation）：Este diretorio contem documentos de Vulnerability Exploitability eXchange VEX para o EGC. 证据：`translations/pt/docs/security/vex/README.md`
- **Especificacao do EGC**（documentation）：A especificacao do EGC e executavel. Ela vive em JSON Schemas em schemas/ , em manifestos de instalacao em scripts/lib/ , e em testes em tests/spec/ . Este documento e o indice que os une. 证据：`translations/pt/docs/spec/README.md`
- **Article Writing**（skill_instruction）：Write long-form content that sounds like an actual person with a point of view, not an LLM smoothing itself into paste. 证据：`skills/docs/article-writing/SKILL.md`
- **Brand Voice**（skill_instruction）：Build a durable voice profile from real source material, then use that profile everywhere instead of re-deriving style from scratch or defaulting to generic AI copy. 证据：`skills/docs/brand-voice/SKILL.md`
- **Content Engine**（skill_instruction）：Build platform-native content without flattening the author's real voice into platform slop. 证据：`skills/docs/content-engine/SKILL.md`
- **Documentation Lookup Context7**（skill_instruction）：When the user asks about libraries, frameworks, or APIs, fetch current documentation via the Context7 MCP tools resolve-library-id and query-docs instead of relying on training data. 证据：`skills/docs/documentation-lookup/SKILL.md`
- **Investor Materials**（skill_instruction）：Build investor-facing materials that are consistent, credible, and easy to defend. 证据：`skills/docs/investor-materials/SKILL.md`
- **Investor Outreach**（skill_instruction）：Write investor communication that is short, concrete, and easy to act on. 证据：`skills/docs/investor-outreach/SKILL.md`
- **Market Research**（skill_instruction）：Produce research that supports decisions, not research theater. 证据：`skills/docs/market-research/SKILL.md`
- **Machine Learning Engineering Workflow**（skill_instruction）：Machine Learning Engineering Workflow 证据：`skills/docs/mle-workflow/SKILL.md`
- **PubMed Database**（skill_instruction）：Use this skill when a task needs biomedical literature from PubMed rather than general web search. 证据：`skills/docs/scientific-db-pubmed-database/SKILL.md`
- **USPTO Database**（skill_instruction）：Use this skill when a task needs official United States patent or trademark records from USPTO systems. 证据：`skills/docs/scientific-db-uspto-database/SKILL.md`
- **gget**（skill_instruction）：Use this skill when a task needs quick bioinformatics lookup across genomic reference databases with the gget CLI or Python package. 证据：`skills/docs/scientific-pkg-gget/SKILL.md`
- **Literature Review**（skill_instruction）：Use this skill when the task is to find, screen, synthesize, and cite a body of academic or technical literature. 证据：`skills/docs/scientific-thinking-literature-review/SKILL.md`
- **Scholar Evaluation**（skill_instruction）：Use this skill to evaluate academic or scientific work with a repeatable rubric. 证据：`skills/docs/scientific-thinking-scholar-evaluation/SKILL.md`
- **SEO**（skill_instruction）：Improve search visibility through technical correctness, performance, and content relevance, not gimmicks. 证据：`skills/docs/seo/SKILL.md`
- **EGC - Extended Global Context for CodeBuddy**（documentation）：EGC - Extended Global Context for CodeBuddy 证据：`.codebuddy/README.md`
- **.codex-plugin: Codex Native Plugin for EGC**（documentation）：.codex-plugin: Codex Native Plugin for EGC 证据：`.codex-plugin/README.md`
- **Plugin Manifest Gotchas**（documentation）：If you plan to edit .gemini-plugin/plugin.json , be aware that the Gemini plugin validator enforces several undocumented but strict constraints that can cause installs to fail with vague errors for example, agents: Invalid input . In particular, component fields must be arrays, agents is not a supported manifest field and must not be included in plugin.json, and a version field is required for reliable validation and installation. 证据：`.gemini-plugin/README.md`
- **EGC - Extended Global Context for Kiro**（documentation）：EGC - Extended Global Context for Kiro 证据：`.kiro/README.md`
- **OpenCode EGC Plugin**（documentation）：WARNING: This README is specific to OpenCode usage. If you installed EGC via npm e.g. npm install opencode-egc , refer to the root README instead. 证据：`.opencode/README.md`
- **EGC - Extended Global Context for Trae**（documentation）：EGC - Extended Global Context for Trae 证据：`.trae/README.md`
- **EGC - Extended Global Context**（documentation）：🌐 English · العربية translations/ar/README.md · Español translations/es/README.md · हिन्दी translations/hi/README.md · 日本語 translations/ja/README.md · 한국어 translations/ko/README.md · Português Brasil translations/pt/README.md · Русский translations/ru/README.md 证据：`README.md`
- **Hooks**（documentation）：Hooks are event-driven automations that fire before or after AI tool executions. They enforce code quality, catch mistakes early, and automate repetitive checks. 证据：`hooks/README.md`
- **Rules**（documentation）：Rules are organized into a common layer plus language-specific directories: 证据：`rules/README.md`
- **Hooks in Kiro**（documentation）：1. IDE Hooks this directory - Standalone .kiro.hook files that work in the Kiro IDE 2. CLI Hooks - Embedded in agent configuration files for CLI usage 证据：`.kiro/hooks/README.md`
- **GAN-Style Harness Examples**（documentation）：Examples showing how to use the Generator-Evaluator harness for different project types. 证据：`examples/gan-harness/README.md`
- **规则**（documentation）：- common/ 包含通用原则: 无语言特定的代码示例。 - zh/ 包含 common 目录的中文翻译版本。 - 语言目录 扩展通用规则，包含框架特定的模式、工具和代码示例。每个文件引用其对应的通用版本。 证据：`rules/zh/README.md`
- **scripts/runtime**（documentation）：router.js , mount-all.js , unmount-all.js , and activator.js implement a dynamic skill router that would materialize skills into .agents/skills/ . 证据：`scripts/runtime/README.md`
- **Visa Document Translator**（documentation）：Automatically translate visa application documents from images to professional English PDFs. 证据：`skills/security/visa-doc-translate/README.md`
- **src/llm: Python LLM Bridge**（documentation）：Python bridge invoked by egc prompt via scripts/gemini.js . 证据：`src/llm/README.md`
- **EGC - السياق العالمي الممتد Extended Global Context**（documentation）：🌐 English ../../README.md · العربية · Español ../es/README.md · हिन्दी ../hi/README.md · 日本語 ../ja/README.md · 한국어 ../ko/README.md · Português Brasil ../pt/README.md · Русский ../ru/README.md 证据：`translations/ar/README.md`
- **EGC - Contexto Global Extendido**（documentation）：🌐 English ../../README.md · العربية ../ar/README.md · Español · हिन्दी ../hi/README.md · 日本語 ../ja/README.md · 한국어 ../ko/README.md · Português Brasil ../pt/README.md · Русский ../ru/README.md 证据：`translations/es/README.md`
- **EGC - Extended Global Context**（documentation）：🌐 English ../../README.md · العربية ../ar/README.md · Español ../es/README.md · हिन्दी · 日本語 ../ja/README.md · 한국어 ../ko/README.md · Português Brasil ../pt/README.md · Русский ../ru/README.md 证据：`translations/hi/README.md`
- **EGC - Extended Global Context**（documentation）：🌐 English ../../README.md · العربية ../ar/README.md · Español ../es/README.md · हिन्दी ../hi/README.md · 日本語 · 한국어 ../ko/README.md · Português Brasil ../pt/README.md · Русский ../ru/README.md 证据：`translations/ja/README.md`
- **EGC - 확장된 전역 컨텍스트 Extended Global Context**（documentation）：🌐 English ../../README.md · العربية ../ar/README.md · Español ../es/README.md · हिन्दी ../hi/README.md · 日本語 ../ja/README.md · 한국어 · Português Brasil ../pt/README.md · Русский ../ru/README.md 证据：`translations/ko/README.md`
- **EGC - Extended Global Context**（documentation）：🌐 English ../../README.md · العربية ../ar/README.md · Español ../es/README.md · हिन्दी ../hi/README.md · 日本語 ../ja/README.md · 한국어 ../ko/README.md · Português Brasil · Русский ../ru/README.md 证据：`translations/pt/README.md`
- **EGC - Расширенный глобальный контекст Extended Global Context**（documentation）：🌐 English ../../README.md · العربية ../ar/README.md · Español ../es/README.md · हिन्दी ../hi/README.md · 日本語 ../ja/README.md · 한국어 ../ko/README.md · Português Brasil ../pt/README.md · Русский 证据：`translations/ru/README.md`
- **Package**（package_manifest）：{ "name": "egc-universal", "version": "1.1.11", "description": "Extended Global Context EGC plugin for OpenCode - agents, commands, hooks, and skills", "main": "dist/index.js", "types": "dist/index.d.ts", "type": "module", "exports": { ".": { "types": "./dist/index.d.ts", "import": "./dist/index.js" }, "./plugins": { "types": "./dist/plugins/index.d.ts", "import": "./dist/plugins/index.js" }, "./tools": { "types": "./dist/tools/index.d.ts", "import": "./dist/tools/index.js" } }, "files": "dist", "commands", "prompts", "instructions", "opencode.json", "README.md" , "scripts": { "build": "tsc", "clean": "rm -rf dist", "prepublishOnly": "npm run build" }, "keywords": "opencode", "plugin", "egc… 证据：`.opencode/package.json`
- **Package**（package_manifest）：{ "name": "egc-dashboard", "version": "1.0.0", "private": true, "description": "EGC real-time agent monitoring dashboard", "main": "server.js", "scripts": { "start": "node server.js" }, "dependencies": { "express": "^4.18.2", "ws": "^8.16.0" } } 证据：`dashboard/package.json`
- **Package**（package_manifest）：{ "name": "mcp-proxy-pin", "version": "1.0.0", "private": true, "dependencies": { "mcp-proxy": "6.4.3" } } 证据：`docker/package.json`
- **Package**（package_manifest）：{ "name": "egc-fuzz", "version": "1.0.0", "private": true, "description": "Fuzzing harness for egc-guardian validator", "dependencies": { "jsfuzz": "1.0.15" }, "overrides": { "cross-spawn": "^7.0.6", "uuid": "^11.1.1", "js-yaml": "^4.2.0" } } 证据：`fuzz/package.json`
- **Package**（package_manifest）：{ "name": "@egchq/egc", "version": "1.1.11", "description": "EGC - Extended Global Context. Persistent memory and shared context for AI coding tools.", "author": "Felipe Marzochi https://github.com/Fmarzochi ", "license": "MIT", "repository": { "type": "git", "url": "git+https://github.com/Fmarzochi/EGC.git" }, "homepage": "https://github.com/Fmarzochi/EGC", "bugs": { "url": "https://github.com/Fmarzochi/EGC/issues" }, "engines": { "node": " =20.0.0" }, "main": "scripts/egc.js", "bin": { "egc": "scripts/egc.js", "egc-install": "scripts/install-apply.js", "egc-memory": "mcp/servers/egc-memory/build/index.js", "egc-guardian": "mcp/servers/egc-guardian/build/index.js" }, "scripts": { "preinsta… 证据：`package.json`
- **Package**（package_manifest）：{ "name": "egc-guardian", "version": "1.0.0", "type": "module", "description": "EGC Guardian MCP Server: validation, context reduction, and task orchestration", "main": "build/index.js", "scripts": { "build": "node ../../../scripts/build-skill-index.js && tsc && chmod +x build/index.js", "start": "node build/index.js" }, "dependencies": { "@modelcontextprotocol/sdk": "^1.4.1", "sqlite": "^5.1.1", "sqlite3": "^6.0.1", "zod": "^4.4.3" }, "devDependencies": { "@types/node": "^26.1.1", "typescript": "^7.0.2" }, "overrides": { "undici": "^6.27.0" } } 证据：`mcp/servers/egc-guardian/package.json`
- **Package**（package_manifest）：{ "name": "egc-memory", "version": "1.0.0", "description": "EGC Memory MCP Server: persistent cross-session project state and decision history", "main": "build/index.js", "scripts": { "build": "tsc && chmod +x build/index.js", "start": "node build/index.js" }, "dependencies": { "@modelcontextprotocol/sdk": "^1.4.1", "simple-git": "^3.36.0", "sqlite": "^5.1.1", "sqlite3": "^6.0.1", "zod": "^4.4.3" }, "devDependencies": { "@types/node": "^22.0.0", "typescript": "^7.0.2" }, "engines": { "node": " =18" }, "overrides": { "undici": "^6.27.0" } } 证据：`mcp/servers/egc-memory/package.json`
- **Agent Introspection Debugging**（skill_instruction）：Use this skill when an agent run is failing repeatedly, consuming tokens without progress, looping on the same tools, or drifting away from the intended task. 证据：`.agents/skills/agent-introspection-debugging/SKILL.md`
- **Agent Sort**（skill_instruction）：Use this skill when a repo needs a project-specific EGC surface instead of the default full install. 证据：`.agents/skills/agent-sort/SKILL.md`
- **API Design Patterns**（skill_instruction）：Conventions and best practices for designing consistent, developer-friendly REST APIs. 证据：`.agents/skills/api-design/SKILL.md`
- **Article Writing**（skill_instruction）：Write long-form content that sounds like an actual person with a point of view, not an LLM smoothing itself into paste. 证据：`.agents/skills/article-writing/SKILL.md`
- **Backend Development Patterns**（skill_instruction）：Backend architecture patterns and best practices for scalable server-side applications. 证据：`.agents/skills/backend-patterns/SKILL.md`
- 其余 120 条证据见 `AI_CONTEXT_PACK.json` 或 `EVIDENCE_INDEX.json`。

## 宿主 AI 必须遵守的规则

- **把本资产当作开工前上下文，而不是运行环境。**：AI Context Pack 只包含证据化项目理解，不包含目标项目的可执行状态。 证据：`docs/README.md`, `docs/architecture/README.md`, `docs/governance/README.md`
- **回答用户时区分可预览内容与必须安装后才能验证的内容。**：安装前体验的消费者价值来自降低误装和误判，而不是伪装成真实运行。 证据：`docs/README.md`, `docs/architecture/README.md`, `docs/governance/README.md`

## 用户开工前应该回答的问题

- 你准备在哪个宿主 AI 或本地环境中使用它？
- 你只是想先体验工作流，还是准备真实安装？
- 你最在意的是安装成本、输出质量、还是和现有规则的冲突？

## 验收标准

- 所有能力声明都能回指到 evidence_refs 中的文件路径。
- AI_CONTEXT_PACK.md 没有把预览包装成真实运行。
- 用户能在 3 分钟内看懂适合谁、能做什么、如何开始和风险边界。

---

## Doramagic Context Augmentation

下面内容用于强化 Repomix/AI Context Pack 主体。Human Manual 只提供阅读骨架；踩坑日志会被转成宿主 AI 必须遵守的工作约束。

## Human Manual 骨架

使用规则：这里只是项目阅读路线和显著性信号，不是事实权威。具体事实仍必须回到 repo evidence / Claim Graph。

宿主 AI 硬性规则：
- 不得把页标题、章节顺序、摘要或 importance 当作项目事实证据。
- 解释 Human Manual 骨架时，必须明确说它只是阅读路线/显著性信号。
- 能力、安装、兼容性、运行状态和风险判断必须引用 repo evidence、source path 或 Claim Graph。

- **项目概述、安装与多工具集成目标**：importance `high`
  - source_paths: README.md, package.json, install.sh, install.ps1, docs/installation.md
- **egc-memory 持久记忆系统**：importance `high`
  - source_paths: mcp/servers/egc-memory/src/index.ts, mcp/servers/egc-memory/src/working-memory.ts, mcp/servers/egc-memory/src/branch-state.ts, mcp/servers/egc-memory/src/patterns.ts, mcp/servers/egc-memory/src/compress.ts
- **egc-guardian 安全守护层与 LLM 调度**：importance `high`
  - source_paths: mcp/servers/egc-guardian/src/index.ts, mcp/servers/egc-guardian/src/validator.ts, mcp/servers/egc-guardian/src/guardian-cli.ts, mcp/servers/egc-guardian/src/llm-router.ts, mcp/servers/egc-guardian/src/headroom-client.ts
- **提示词库：智能体、技能、命令与规则**：importance `high`
  - source_paths: agents/code-reviewer.md, agents/architect.md, agents/security-reviewer.md, agents/planner.md, commands/plan.md

## Repo Inspection Evidence / 源码检查证据

- repo_clone_verified: true
- repo_inspection_verified: true
- repo_commit: `94288cd6395050a9640a9a01d5d171d641d12d8d`
- inspected_files: `Dockerfile`, `README.md`, `package.json`, `requirements.txt`, `docs/MAINTAINERS.md`, `docs/README.md`, `docs/ROADMAP.md`, `docs/RULES.md`, `docs/TROUBLESHOOTING.md`, `docs/architecture/ARCHITECTURE-IMPROVEMENTS.md`, `docs/architecture/EGC_2.0_BLUEPRINT.md`, `docs/architecture/EGC_2.0_TECHNICAL_DESIGN.md`, `docs/architecture/README.md`, `docs/architecture/SELECTIVE-INSTALL-ARCHITECTURE.md`, `docs/architecture/SELECTIVE-INSTALL-DESIGN.md`, `docs/architecture/SINGLE-AGENT-OPERATIONAL-MODEL.md`, `docs/architecture/continuous-learning-v2-spec.md`, `docs/architecture/cross-harness.md`, `docs/governance/README.md`, `docs/governance/SKILL-PLACEMENT-POLICY.md`

宿主 AI 硬性规则：
- 没有 repo_clone_verified=true 时，不得声称已经读过源码。
- 没有 repo_inspection_verified=true 时，不得把 README/docs/package 文件判断写成事实。
- 没有 quick_start_verified=true 时，不得声称 Quick Start 已跑通。

## Doramagic Pitfall Constraints / 踩坑约束

这些规则来自 Doramagic 发现、验证或编译过程中的项目专属坑点。宿主 AI 必须把它们当作工作约束，而不是普通说明文字。

### Constraint 1: 失败模式：installation: feat(install): add native Cline install target

- Trigger: Developers should check this installation risk before relying on the project: feat(install): add native Cline install target
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: feat(install): add native Cline install target. Context: Observed when using node, python
- Why it matters: Developers may fail before the first successful local run: feat(install): add native Cline install target
- Evidence: failure_mode_cluster:github_issue | https://github.com/Fmarzochi/EGC/issues/846 | feat(install): add native Cline install target
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 2: 失败模式：installation: feat(install): add native JetBrains Junie install target

- Trigger: Developers should check this installation risk before relying on the project: feat(install): add native JetBrains Junie install target
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: feat(install): add native JetBrains Junie install target. Context: Observed when using node, python
- Why it matters: Developers may fail before the first successful local run: feat(install): add native JetBrains Junie install target
- Evidence: failure_mode_cluster:github_issue | https://github.com/Fmarzochi/EGC/issues/847 | feat(install): add native JetBrains Junie install target
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 3: 失败模式：installation: feat(install): add native Qwen Code install target

- Trigger: Developers should check this installation risk before relying on the project: feat(install): add native Qwen Code install target
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: feat(install): add native Qwen Code install target. Context: Observed when using node, python
- Why it matters: Developers may fail before the first successful local run: feat(install): add native Qwen Code install target
- Evidence: failure_mode_cluster:github_issue | https://github.com/Fmarzochi/EGC/issues/845 | feat(install): add native Qwen Code install target
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 4: 来源证据：feat(install): add native Cline install target

- Trigger: GitHub 社区证据显示该项目存在一个安装相关的待验证问题：feat(install): add native Cline install target
- Why it matters: 可能增加新用户试用和生产接入成本。
- Evidence: community_evidence:github | https://github.com/Fmarzochi/EGC/issues/846 | 来源讨论提到 python 相关条件，需在安装/试用前复核。
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 5: 来源证据：feat(install): add native JetBrains Junie install target

- Trigger: GitHub 社区证据显示该项目存在一个安装相关的待验证问题：feat(install): add native JetBrains Junie install target
- Host AI rule: 来源显示可能已有修复、规避或版本变化，说明书中必须标注适用版本。
- Why it matters: 可能增加新用户试用和生产接入成本。
- Evidence: community_evidence:github | https://github.com/Fmarzochi/EGC/issues/847 | 来源讨论提到 python 相关条件，需在安装/试用前复核。
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 6: 来源证据：feat(install): add native Qwen Code install target

- Trigger: GitHub 社区证据显示该项目存在一个安装相关的待验证问题：feat(install): add native Qwen Code install target
- Why it matters: 可能增加新用户试用和生产接入成本。
- Evidence: community_evidence:github | https://github.com/Fmarzochi/EGC/issues/845 | 来源讨论提到 python 相关条件，需在安装/试用前复核。
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 7: 来源证据：feat(install): add native Roo Code install target

- Trigger: GitHub 社区证据显示该项目存在一个安装相关的待验证问题：feat(install): add native Roo Code install target
- Why it matters: 可能增加新用户试用和生产接入成本。
- Evidence: community_evidence:github | https://github.com/Fmarzochi/EGC/issues/852 | 来源讨论提到 python 相关条件，需在安装/试用前复核。
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 8: 可能修改宿主 AI 配置

- Trigger: 项目面向 Claude/Cursor/Codex/Gemini/OpenCode 等宿主，或安装命令涉及用户配置目录。
- Host AI rule: 列出会写入的配置文件、目录和卸载/回滚步骤。
- Why it matters: 安装可能改变本机 AI 工具行为，用户需要知道写入位置和回滚方法。
- Evidence: capability.host_targets | https://github.com/Fmarzochi/EGC | host_targets=mcp_host, claude_code, claude, cursor, gemini_cli
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 9: 失败模式：configuration: v1.1.10

- Trigger: Developers should check this configuration risk before relying on the project: v1.1.10
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: v1.1.10. Context: Source discussion did not expose a precise runtime context.
- Why it matters: Upgrade or migration may change expected behavior: v1.1.10
- Evidence: failure_mode_cluster:github_release | https://github.com/Fmarzochi/EGC/releases/tag/v1.1.10 | v1.1.10
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 10: 失败模式：configuration: v1.1.11

- Trigger: Developers should check this configuration risk before relying on the project: v1.1.11
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: v1.1.11. Context: Observed when using node
- Why it matters: Upgrade or migration may change expected behavior: v1.1.11
- Evidence: failure_mode_cluster:github_release | https://github.com/Fmarzochi/EGC/releases/tag/v1.1.11 | v1.1.11
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。