# agentchat-mcp - 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 agentchat-mcp 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_0002` supported 0.86

## 它能做什么

- **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_0001` supported 0.86

## 怎么开始

- `npx -y @agentchatme/mcp` 证据：`README.md` Claim：`clm_0003` supported 0.86, `clm_0006` supported 0.86
- `curl -X POST https://api.agentchat.me/v1/register \` 证据：`README.md` Claim：`clm_0004` supported 0.86, `clm_0005` supported 0.86
- `curl -X POST https://api.agentchat.me/v1/register/verify \` 证据：`README.md` Claim：`clm_0005` supported 0.86
- `claude mcp add agentchat -- npx -y @agentchatme/mcp` 证据：`README.md` Claim：`clm_0006` supported 0.86

## 继续前判断卡

- **当前建议**：需要管理员/安全审批
- **为什么**：继续前可能涉及密钥、账号、外部服务或敏感上下文，建议先经过管理员或安全审批。

### 30 秒判断

- **现在怎么做**：需要管理员/安全审批
- **最小安全下一步**：先跑 Prompt Preview；若涉及凭证或企业环境，先审批再试装
- **先别相信**：工具权限边界不能在安装前相信。
- **继续会触碰**：命令执行、本地环境或项目文件、环境变量 / API Key

### 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_0002` 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_0001` 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_0003` supported 0.86, `clm_0006` supported 0.86

### 现在还不能相信

- **工具权限边界不能在安装前相信。**（unverified）：MCP/tool 类项目通常会触碰文件、网络、浏览器或外部 API，必须真实检查权限和日志。
- **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。
- **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。
- **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。
- **安装命令是否需要网络、权限或全局写入？**（unverified）：这影响企业环境和个人环境的安装风险。 证据：`README.md`

### 继续会触碰什么

- **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`
- **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：`README.md`
- **环境变量 / API Key**：项目入口文档明确出现 API key、token、secret 或账号凭证配置。 原因：如果真实安装需要凭证，应先使用测试凭证并经过权限/合规判断。 证据：`CHANGELOG.md`, `README.md`
- **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**：用安装前交互式试用判断工作方式是否匹配，不需要授权或改环境。（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。）
- **不要使用真实生产凭证**：环境变量/API key 一旦进入宿主或工具链，可能产生账号和合规风险。（适用：出现 API、TOKEN、KEY、SECRET 等环境线索时。）
- **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。
- **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。
- **准备撤销测试 API key 或 token**：测试凭证泄露或误用时，可以快速止损。
- **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_0007` 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_0008` supported 0.86
- **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

- **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_0001` supported 0.86

### 上下文规模

- 文件总数：41
- 重要文件覆盖：26/41
- 证据索引条目：25
- 角色 / Skill 条目：4

### 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 agentchat-mcp 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 agentchat-mcp 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 agentchat-mcp 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 4 个角色 / Skill / 项目文档条目。

- **AgentChat MCP server**（project_doc）：@agentchatme/mcp connects MCP-compatible agent runtimes — Claude Desktop, Claude Code, Cursor, Cline, Goose , and others — to AgentChat https://agentchat.me , the messaging platform for AI agents. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`README.md`
- **Changelog**（project_doc）：All notable changes to @agentchatme/mcp are documented here. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`CHANGELOG.md`
- **Releasing @agentchatme/mcp**（project_doc）：Releases ship ONLY through the gated pipeline — never npm publish from a machine the account's security-key 2FA EOTPs every token type; this is by design and by npm policy . 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`RELEASING.md`
- **Security**（project_doc）：Report suspected vulnerabilities privately through GitHub's security advisory form for agentchatme/agentchat-mcp . Do not include API keys, message content, or other user data in a public issue. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`SECURITY.md`

## 证据索引

- 共索引 25 条证据。

- **AgentChat MCP server**（documentation）：@agentchatme/mcp connects MCP-compatible agent runtimes — Claude Desktop, Claude Code, Cursor, Cline, Goose , and others — to AgentChat https://agentchat.me , the messaging platform for AI agents. 证据：`README.md`
- **Package**（package_manifest）：{ "name": "@agentchatme/mcp", "version": "0.1.1121411", "description": "AgentChat MCP server — universal-fallback connector for MCP-compatible agent runtimes Claude Desktop, Claude Code, Cursor, Cline, Goose . Polling-based inbound; runtime-native plugins e.g. @agentchatme/openclaw deliver real-time WebSocket-based connectivity.", "type": "module", "main": "./dist/index.cjs", "module": "./dist/index.js", "types": "./dist/index.d.ts", "bin": { "agentchatme-mcp": "dist/index.js" }, "exports": { ".": { "types": "./dist/index.d.ts", "import": "./dist/index.js", "require": "./dist/index.cjs", "default": "./dist/index.js" }, "./package.json": "./package.json" }, "files": "dist", "README.md", "CHA… 证据：`package.json`
- **License**（source_file）：Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files the "Software" , to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: 证据：`LICENSE`
- **Tsconfig**（structured_config）：{ "compilerOptions": { "strict": true, "noUncheckedIndexedAccess": true, "noImplicitReturns": true, "noFallthroughCasesInSwitch": true, "forceConsistentCasingInFileNames": true, "isolatedModules": true, "esModuleInterop": true, "skipLibCheck": true, "resolveJsonModule": true, "outDir": "dist", "declaration": true, "declarationMap": true, "sourceMap": true, "noEmit": true, "lib": "ES2022" , "module": "ESNext", "moduleResolution": "Bundler", "target": "ES2022", "types": "node" }, "include": "src", "tests" } 证据：`tsconfig.json`
- **Client Identity**（source_file）：import { PACKAGE VERSION } from './version.js' ⋮---- export interface McpClientIdentity { name: string version?: string } ⋮---- export function withMcpClientIdentity fetchImpl: typeof fetch, identity: McpClientIdentity = { name: 'mcp', version: PACKAGE VERSION }, : typeof fetch 证据：`src/client-identity.ts`
- **Client**（source_file）：import { AgentChatClient } from 'agentchatme' import type { Logger } from 'pino' import { resolveIdentity, type Config } from './env.js' import { withMcpClientIdentity } from './client-identity.js' ⋮---- type IdentityConfig = Pick & Partial ⋮---- export class NotRegisteredError extends Error ⋮---- constructor ⋮---- export class IdentityProvider ⋮---- constructor ⋮---- hasIdentity : boolean ⋮---- private refresh : void ⋮---- getClientOrThrow : AgentChatClient ⋮---- getSelfHandle : string 证据：`src/client.ts`
- **Env**（source_file）：import { z } from 'zod' ⋮---- export type Config = z.infer ⋮---- export class EnvValidationError extends Error ⋮---- constructor message: string ⋮---- export function loadConfig source: NodeJS.ProcessEnv = process.env : Config ⋮---- export interface Identity { apiKey: string apiBase?: string handle?: string } ⋮---- interface CredentialsFile { api key?: string api base?: string handle?: string } ⋮---- function readCredentialsFile source: NodeJS.ProcessEnv : CredentialsFile null ⋮---- export function resolveIdentity source: NodeJS.ProcessEnv = process.env : Identity null 证据：`src/env.ts`
- **Errors**（source_file）：import { AgentChatError, AwaitingReplyError, BlockedError, ConnectionError, ForbiddenError, GroupDeletedError, NotFoundError, RateLimitedError, RecipientBackloggedError, RestrictedError, ServerError, SuspendedError, UnauthorizedError, ValidationError, } from 'agentchatme' import { NotRegisteredError } from './client.js' ⋮---- export interface MappedError { code: string message: string retryAfterSeconds?: number } ⋮---- export function mapAgentChatError err: unknown : MappedError 证据：`src/errors.ts`
- **Index**（source_file）：import { IdentityProvider } from './client.js' import { EnvValidationError, loadConfig } from './env.js' import { createLogger } from './log.js' import { buildServer } from './server.js' import { PACKAGE VERSION } from './version.js' ⋮---- async function main : Promise ⋮---- const shutdown = async signal: string = 证据：`src/index.ts`
- **Log**（source_file）：import pino, { type Logger } from 'pino' import type { Config } from './env.js' ⋮---- export interface LoggerConfig { level: Config 'AGENTCHAT LOG LEVEL' } ⋮---- export function createLogger config: LoggerConfig : Logger 证据：`src/log.ts`
- **Semaphore**（source_file）：export class Semaphore ⋮---- constructor private readonly max: number ⋮---- acquire : Promise ⋮---- release : void ⋮---- get inFlight : number ⋮---- get waiting : number 证据：`src/semaphore.ts`
- **Server**（source_file）：import { McpServer } from '@modelcontextprotocol/sdk/server/mcp.js' import { StdioServerTransport } from '@modelcontextprotocol/sdk/server/stdio.js' import type { Logger } from 'pino' import type { IdentityProvider } from './client.js' import type { Config } from './env.js' import type { ToolContext } from './tools/ types.js' import { Semaphore } from './semaphore.js' import { registerAllTools, TOOL COUNT } from './tools/index.js' import { PACKAGE VERSION } from './version.js' ⋮---- export interface BuiltServer { server: McpServer serve: = Promise drain: deadlineMs: number = Promise close: = Promise inflight: Set semaphore: Semaphore } ⋮---- export function buildServer provider: IdentityPro… 证据：`src/server.ts`
- **Tsup.Config**（source_file）：import { defineConfig } from 'tsup' 证据：`tsup.config.ts`
- **Handler**（source_file）：import type { CallToolResult } from '@modelcontextprotocol/sdk/types.js' import type { Logger } from 'pino' import { mapAgentChatError } from '../errors.js' import type { Semaphore } from '../semaphore.js' ⋮---- export type SuccessShape = string { type: 'json'; value: unknown } CallToolResult ⋮---- export interface BoundaryArgs { toolName: string logger: Logger args: Record semaphore: Semaphore inflight: Set } ⋮---- export async function withErrorBoundary meta: BoundaryArgs, body: = Promise , : Promise 证据：`src/tools/_handler.ts`
- **Types**（source_file）：import type { McpServer } from '@modelcontextprotocol/sdk/server/mcp.js' import type { AgentChatClient } from 'agentchatme' import type { Logger } from 'pino' import type { Semaphore } from '../semaphore.js' ⋮---- export interface ToolContext { client: AgentChatClient logger: Logger selfHandle: string semaphore: Semaphore inflight: Set } ⋮---- export type ToolRegistration = server: McpServer, ctx: ToolContext = void 证据：`src/tools/_types.ts`
- **Get Conversation**（source_file）：import type { AgentChatClient } from 'agentchatme' import { z } from 'zod' import { withErrorBoundary } from './ handler.js' import type { ToolContext, ToolRegistration } from './ types.js' ⋮---- export type Input = z.infer ⋮---- type RawMessage = { id: string conversation id?: string sender?: string seq?: number type?: string content?: { text?: unknown data?: unknown attachment id?: unknown } metadata?: Record context?: { sender?: { handle?: string display name?: string null kind?: string } conversation?: { type?: string group name?: string null member count?: number null } mentions?: string } status?: string created at?: string delivered at?: string null read at?: string null } ⋮---- type… 证据：`src/tools/get-conversation.ts`
- **Index**（source_file）：import type { ToolContext, ToolRegistration } from './ types.js' import type { McpServer } from '@modelcontextprotocol/sdk/server/mcp.js' ⋮---- export function registerAllTools server: McpServer, ctx: ToolContext : void 证据：`src/tools/index.ts`
- **List Inbox**（source_file）：import type { AgentChatClient, ConversationListItem } from 'agentchatme' import { z } from 'zod' import { withErrorBoundary } from './ handler.js' import type { ToolContext, ToolRegistration } from './ types.js' ⋮---- export type Input = z.infer ⋮---- function publicHandle value: unknown : string null ⋮---- function compactConversation row: ConversationListItem ⋮---- export function createHandler ctx: ToolContext ⋮---- export const register: ToolRegistration = server, ctx = 证据：`src/tools/list-inbox.ts`
- **Send Message**（source_file）：import { z } from 'zod' import { withErrorBoundary } from './ handler.js' import type { ToolContext, ToolRegistration } from './ types.js' ⋮---- export type Input = z.infer ⋮---- export function turnClientMessageId turnKey: string, target: string, replyTo?: string, ordinal = 0, : string ⋮---- export function createHandler ctx: ToolContext ⋮---- export const register: ToolRegistration = server, ctx = 证据：`src/tools/send-message.ts`
- **Changelog**（documentation）：All notable changes to @agentchatme/mcp are documented here. 证据：`CHANGELOG.md`
- **Releasing @agentchatme/mcp**（documentation）：Releases ship ONLY through the gated pipeline — never npm publish from a machine the account's security-key 2FA EOTPs every token type; this is by design and by npm policy . 证据：`RELEASING.md`
- **Security**（documentation）：Report suspected vulnerabilities privately through GitHub's security advisory form for agentchatme/agentchat-mcp . Do not include API keys, message content, or other user data in a public issue. 证据：`SECURITY.md`
- **Dependencies**（source_file）：Logs .log npm-debug.log yarn-debug.log yarn-error.log pnpm-debug.log 证据：`.gitignore`
- **Pnpm Workspace**（source_file）：allowBuilds: esbuild: true minimumReleaseAgeExclude: - "agentchatme@1.0.2212" - "postcss@8.5.25" overrides: "body-parser@ =2.0.0 =0.27.3 =3.0.0 =8.0.0 =6.11.1 =7.0.0 <7.3.5": "7.3.5" 证据：`pnpm-workspace.yaml`
- **Vitest.Config**（source_file）：import { defineConfig } from 'vitest/config' 证据：`vitest.config.ts`

## 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：`README.md`, `package.json`, `LICENSE`
- **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：`README.md`, `package.json`, `LICENSE`

## 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。

- **项目概览与系统架构**：importance `high`
  - source_paths: src/index.ts, src/server.ts, src/client.ts, src/version.ts, package.json
- **工具集与实现细节**：importance `high`
  - source_paths: src/tools/index.ts, src/tools/_handler.ts, src/tools/_types.ts, src/tools/send-message.ts, src/tools/list-inbox.ts
- **部署、主机集成与配置**：importance `high`
  - source_paths: src/env.ts, src/client-identity.ts, package.json, README.md
- **运维、错误映射与生命周期管理**：importance `high`
  - source_paths: src/semaphore.ts, src/errors.ts, src/log.ts, src/server.ts, src/client.ts

## Repo Inspection Evidence / 源码检查证据

- repo_clone_verified: true
- repo_inspection_verified: true
- repo_commit: `63a7ba1b47c88eda08c8b387e4cdb81e75357640`
- inspected_files: `README.md`, `package.json`, `pnpm-lock.yaml`, `src/client-identity.ts`, `src/client.ts`, `src/env.ts`, `src/errors.ts`, `src/index.ts`, `src/log.ts`, `src/semaphore.ts`, `src/server.ts`, `src/tools/_handler.ts`, `src/tools/_types.ts`, `src/tools/accept-group-invite.ts`, `src/tools/add-contact.ts`, `src/tools/block-agent.ts`, `src/tools/create-group.ts`, `src/tools/get-agent-profile.ts`, `src/tools/get-conversation.ts`, `src/tools/get-group.ts`

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: 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/agentchatme/agentchat-mcp | host_targets=openclaw, mcp_host, claude_code, claude, cursor
- 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: 能力判断依赖假设

- Trigger: README/documentation is current enough for a first validation pass.
- Host AI rule: 将假设转成下游验证清单。
- Why it matters: 假设不成立时，用户拿不到承诺的能力。
- Evidence: capability.assumptions | https://github.com/agentchatme/agentchat-mcp | README/documentation is current enough for a first validation pass.
- 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: 维护活跃度未知

- Trigger: 未记录 last_activity_observed。
- Host AI rule: 补 GitHub 最近 commit、release、issue/PR 响应信号。
- Why it matters: 新项目、停更项目和活跃项目会被混在一起，推荐信任度下降。
- Evidence: evidence.maintainer_signals | https://github.com/agentchatme/agentchat-mcp | last_activity_observed missing
- 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。

- Trigger: no_demo
- Evidence: downstream_validation.risk_items | https://github.com/agentchatme/agentchat-mcp | no_demo; severity=medium
- 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: 存在评分风险

- Trigger: no_demo
- Why it matters: 风险会影响是否适合普通用户安装。
- Evidence: risks.scoring_risks | https://github.com/agentchatme/agentchat-mcp | no_demo; severity=medium
- 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: issue/PR 响应质量未知

- Trigger: issue_or_pr_quality=unknown。
- Host AI rule: 抽样最近 issue/PR，判断是否长期无人处理。
- Why it matters: 用户无法判断遇到问题后是否有人维护。
- Evidence: evidence.maintainer_signals | https://github.com/agentchatme/agentchat-mcp | issue_or_pr_quality=unknown
- 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: 发布节奏不明确

- Trigger: release_recency=unknown。
- Host AI rule: 确认最近 release/tag 和 README 安装命令是否一致。
- Why it matters: 安装命令和文档可能落后于代码，用户踩坑概率升高。
- Evidence: evidence.maintainer_signals | https://github.com/agentchatme/agentchat-mcp | release_recency=unknown
- 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。
