# commonplace-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 commonplace-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

- `npm i -g commonplace-mcp` 证据：`README.md` Claim：`clm_0003` supported 0.86
- `claude mcp add commonplace commonplace-mcp` 证据：`README.md` Claim：`clm_0004` supported 0.86

## 继续前判断卡

- **当前建议**：先做权限沙盒试用
- **为什么**：项目存在安装命令、宿主配置或本地写入线索，不建议直接进入主力环境，应先在隔离环境试装。

### 30 秒判断

- **现在怎么做**：先做权限沙盒试用
- **最小安全下一步**：先跑 Prompt Preview；若仍要安装，只在隔离环境试装
- **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、宿主 AI 配置、本地环境或项目文件

### 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

### 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。 证据：`CLAUDE.md`
- **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`
- **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：`CLAUDE.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`
- **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。）
- **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_0005` 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_0006` 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：113
- 重要文件覆盖：40/113
- 证据索引条目：80
- 角色 / Skill 条目：5

### 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 commonplace-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 commonplace-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 commonplace-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 5 个角色 / Skill / 项目文档条目。

- **Project Rules**（project_doc）：Project-specific rules for this codebase. Universal coding rules git practices, quality standards, task sizing, language rules, etc. are in ~/.claude/CLAUDE.md and apply automatically. 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：`CLAUDE.md`
- **commonplace**（project_doc）：Commonplace is a commonplace book for your agent: a local-first store of markdown notes, each paired with a sidecar embeddings file, served to Claude Code or any MCP client over stdio. There is no database -- notes live as .md files on disk, embeddings live next to them as .embedding sidecars, and search is in-memory cosine similarity over those sidecars. 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`
- **Contributing to commonplace**（project_doc）：Thanks for your interest in contributing! This project is small and opinionated; the rules below keep the history clean and the green lights honest. 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：`CONTRIBUTING.md`
- **Sidecar binary format .embedding**（project_doc）：This document specifies the on-disk wire format for the binary .embedding sidecars that pair with each memory .md file. The format is owned by DAR-910 https://linear.app/darkdragonsastro/issue/DAR-910 ; the encode and decode implementations live in src/store/sidecar.ts ../src/store/sidecar.ts . 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：`docs/sidecar-format.md`
- **Changelog**（project_doc）：All notable changes to commonplace-mcp are documented in this file. 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`

## 证据索引

- 共索引 80 条证据。

- **Project Rules**（documentation）：Project-specific rules for this codebase. Universal coding rules git practices, quality standards, task sizing, language rules, etc. are in ~/.claude/CLAUDE.md and apply automatically. 证据：`CLAUDE.md`
- **commonplace**（documentation）：Commonplace is a commonplace book for your agent: a local-first store of markdown notes, each paired with a sidecar embeddings file, served to Claude Code or any MCP client over stdio. There is no database -- notes live as .md files on disk, embeddings live next to them as .embedding sidecars, and search is in-memory cosine similarity over those sidecars. 证据：`README.md`
- **Package**（package_manifest）：{ "name": "commonplace-mcp", "version": "0.2.1", "description": "Local-first commonplace book with embedding-backed semantic search via MCP", "license": "MIT", "type": "module", "main": "dist/index.js", "bin": { "commonplace": "dist/index.js", "commonplace-mcp": "dist/bin/commonplace-mcp.js" }, "files": "dist/", "LICENSE", "README.md", "CHANGELOG.md" , "repository": { "type": "git", "url": "git+https://github.com/rickbassham/commonplace.git" }, "homepage": "https://github.com/rickbassham/commonplace readme", "bugs": { "url": "https://github.com/rickbassham/commonplace/issues" }, "keywords": "mcp", "model-context-protocol", "memory", "embeddings", "semantic-search", "claude", "ai-agent" , "e… 证据：`package.json`
- **Contributing to commonplace**（documentation）：Thanks for your interest in contributing! This project is small and opinionated; the rules below keep the history clean and the green lights honest. 证据：`CONTRIBUTING.md`
- **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`
- **Sidecar binary format .embedding**（documentation）：This document specifies the on-disk wire format for the binary .embedding sidecars that pair with each memory .md file. The format is owned by DAR-910 https://linear.app/darkdragonsastro/issue/DAR-910 ; the encode and decode implementations live in src/store/sidecar.ts ../src/store/sidecar.ts . 证据：`docs/sidecar-format.md`
- **Changelog**（documentation）：All notable changes to commonplace-mcp are documented in this file. 证据：`CHANGELOG.md`
- **.Versionrc**（structured_config）：{ "bumpFiles": { "filename": "package.json", "type": "json" }, { "filename": "src/server/server.ts", "updater": "scripts/server-version-updater.cjs" } , "header": " Changelog\n\nAll notable changes to commonplace-mcp are documented in this file.\n\nThe format follows Keep a Changelog https://keepachangelog.com/en/1.1.0/ ,\nand this project adheres to Semantic Versioning https://semver.org/ .\n", "types": { "type": "feat", "section": "Added" }, { "type": "fix", "section": "Fixed" }, { "type": "perf", "section": "Performance" }, { "type": "refactor", "section": "Changed" }, { "type": "revert", "section": "Reverted" }, { "type": "chore", "hidden": true }, { "type": "test", "hidden": true }, {… 证据：`.versionrc.json`
- **Tsconfig.Build**（structured_config）：{ "extends": "./tsconfig.json", "compilerOptions": { "outDir": "dist", "rootDir": "src" }, "include": "src/ / " , "exclude": "node modules", "dist", "tests", " / .test.ts", "vitest.config.ts" } 证据：`tsconfig.build.json`
- **Tsconfig**（structured_config）：{ "compilerOptions": { "target": "ES2022", "module": "NodeNext", "moduleResolution": "NodeNext", "strict": true, "esModuleInterop": true, "skipLibCheck": true, "forceConsistentCasingInFileNames": true, "resolveJsonModule": true, "isolatedModules": true, "noUncheckedIndexedAccess": true, "noImplicitOverride": true, "outDir": "dist", "declaration": true, "declarationMap": true, "sourceMap": true, "types": "node", "vitest/globals" }, "include": "src/ / ", "tests/ / ", "vitest.config.ts" , "exclude": "node modules", "dist" } 证据：`tsconfig.json`
- **Cycles 2**（structured_config）：{ "root": "a", "depth": 2, "memories": { "name": "a", "relations": { "to": "b", "type": "related-to" } }, { "name": "b", "relations": { "to": "a", "type": "related-to" } } } 证据：`tests/fixtures/graph/cycles-2.json`
- **Cycles 3**（structured_config）：{ "root": "a", "depth": 3, "memories": { "name": "a", "relations": { "to": "b", "type": "related-to" } }, { "name": "b", "relations": { "to": "c", "type": "related-to" } }, { "name": "c", "relations": { "to": "a", "type": "related-to" } } } 证据：`tests/fixtures/graph/cycles-3.json`
- **Depth 0**（structured_config）：{ "root": "root", "depth": 0, "memories": { "name": "root", "relations": { "to": "alpha", "type": "related-to" } }, { "name": "alpha" } } 证据：`tests/fixtures/graph/depth-0.json`
- **Depth 1**（structured_config）：{ "root": "root", "depth": 1, "memories": { "name": "root", "relations": { "to": "alpha", "type": "related-to" }, { "to": "beta", "type": "builds-on" } }, { "name": "alpha", "relations": { "to": "gamma", "type": "related-to" } }, { "name": "beta" }, { "name": "gamma" } } 证据：`tests/fixtures/graph/depth-1.json`
- **Depth 2**（structured_config）：{ "root": "root", "depth": 2, "memories": { "name": "root", "relations": { "to": "alpha", "type": "related-to" }, { "to": "beta", "type": "builds-on" } }, { "name": "alpha", "relations": { "to": "gamma", "type": "related-to" } }, { "name": "beta" }, { "name": "gamma" } } 证据：`tests/fixtures/graph/depth-2.json`
- **.editorconfig**（source_file）：charset = utf-8 end of line = lf insert final newline = true trim trailing whitespace = true indent style = space indent size = 2 证据：`.editorconfig`
- **.gitignore**（source_file）：node modules/ dist/ .embedding .DS Store .env.local 证据：`.gitignore`
- **.nvmrc**（source_file）：24 证据：`.nvmrc`
- **CHANGELOG.md is hand-authored Keep-a-Changelog markdown. Prettier rewraps**（source_file）：node modules/ dist/ pnpm-lock.yaml .embedding 证据：`.prettierignore`
- **.prettierrc**（source_file）：{ "semi": true, "singleQuote": true, "trailingComma": "all", "printWidth": 100, "tabWidth": 2, "useTabs": false, "endOfLine": "lf" } 证据：`.prettierrc`
- **Makefile**（source_file）：.PHONY: help install build test typecheck lint format format-check audit release release-dry 证据：`Makefile`
- **Eslint.Config**（source_file）：// @ts-check import eslint from '@eslint/js'; import tseslint from 'typescript-eslint'; 证据：`eslint.config.js`
- **!/usr/bin/env node**（source_file）：!/usr/bin/env node / Regenerate the DAR-933 fixture corpus snapshot files tests/fixtures/graph/ .mermaid and .dot from the spec JSONs in the same directory. Run once at fixture-add time; in CI the snapshot tests compare against these committed files. Imports the live TypeScript sources via tsx -- no make build is required before running. Earlier revisions imported from dist/ , which silently broke when the build was stale; see DAR-933 review f-3. Usage: pnpm exec tsx scripts/dar933-generate-fixtures.mjs / 证据：`scripts/dar933-generate-fixtures.mjs`
- **!/usr/bin/env bash**（source_file）：!/usr/bin/env bash DAR-960: derive an npm dist-tag from a semver version string. Usage: scripts/derive-dist-tag.sh Example: scripts/derive-dist-tag.sh 0.1.0 - latest scripts/derive-dist-tag.sh 0.1.0-beta.1 - beta scripts/derive-dist-tag.sh 1.0.0-alpha - alpha Rule: - Input must be a MAJOR.MINOR.PATCH semver core, optionally followed by - and/or + . A leading v is rejected: the workflow strips the leading v from the git tag before invoking this script, so by the time we see the version it must be bare. - If there is no pre-release identifier, print latest . - Otherwise, print the alphabetic prefix of the first pre-release segment the part before the first . . Empty alphabetic prefix is a mal… 证据：`scripts/derive-dist-tag.sh`
- **!/usr/bin/env bash**（source_file）：!/usr/bin/env bash DAR-955 cold-start race reproduction maintainer-only diagnostic . Triggers the race that motivated tests/global-setup.ts: when two vitest workers concurrently load transformers.js with a cold model cache, they corrupt the in-flight model.onnx and downstream parsing fails with one of: - MCP error -32000: Connection closed child process died - Load model from .../onnx/model.onnx failed: Protobuf parsing failed. - mutex lock failed libc++ abi in ORT init when files mid-write The fix is tests/global-setup.ts : serialise the download into the vitest main process before any workers fork. The boot-ordering invariants in tests/server-bin-cold-start.integration.test.ts ac-2 lock t… 证据：`scripts/reproduce-cold-start-race.sh`
- **Server Version Updater**（source_file）：// DAR-963: commit-and-tag-version updater for src/server/server.ts. // // commit-and-tag-version walks every entry in bumpFiles and either uses a // built-in type "json", "plain-text" or a custom updater module that // exports readVersion contents and writeVersion contents, newVersion . // // For src/server/server.ts we need a scoped find/replace: ONLY the // SERVER VERSION constant should change, not any other string that // happens to look like a version. The built-in plain-text updater would // rewrite every occurrence of the current version anywhere in the file -- // brittle. A 6-line custom updater is the right shape. 证据：`scripts/server-version-updater.cjs`
- **!/usr/bin/env bash**（source_file）：!/usr/bin/env bash DAR-914: Apply branch protection to main for the commonplace repo. This script is the reproducible source of truth for the branch protection settings -- review changes here, run the script out-of-band to apply them. CI does not run this script; a maintainer with admin rights does. Required: - gh CLI authenticated as a user with admin on the repo. - The CI workflow .github/workflows/ci.yml defines a matrix on node-version: 22, 24 and a single job named ci . The status-check contexts below "ci 22 " / "ci 24 " must match the rendered matrix-leg job names. Drift between this script and ci.yml is checked by the DAR-914 contract tests. Usage: ./scripts/setup-branch-protection.s… 证据：`scripts/setup-branch-protection.sh`
- **!/usr/bin/env node**（source_file）：!/usr/bin/env node / Bin entry: the commonplace CLI dispatcher DAR-918, extended by DAR-961 . Subcommand surface single source of truth: USAGE in src/cli/migrate.ts : commonplace migrate detect known external memory sources commonplace migrate --from import from a known source; --dry-run / --auto supported commonplace migrate rebuild sidecars for an existing memory dir; --dry-run / --prune-dangling supported The legacy DAR-918 path rebuild sidecars for an existing dir and the DAR-961 detection / import paths share this dispatcher; the bare-bin usage message and the parser usage error message are rendered from the same exported USAGE constant so the two cannot drift. Bin convention package.j… 证据：`src/index.ts`
- **Atomic Write.Test**（source_file）：/ DAR-923 contract tests for the atomicWrite helper. Test names mirror the contract envelope on DAR-923 round 1, approved : - ac-1: write-temp-then-rename + same-filesystem guard - ac-2: fsync semantics + descriptor lifecycle The helper writes bytes to a tmp file colocated with the target, fsyncs the tmpfile, renames over the target, then fsyncs the directory. It throws when tmpdir and target dir straddle filesystems and surfaces fsync errors. The helper exposes an internal atomicWriteHooks test seam to inject a fake fs.promises -shaped dependency. This avoids the non-configurable descriptor problem on the fs/promises namespace and lets us observe the exact fsync / rename / close call order… 证据：`tests/atomic-write.test.ts`
- **Ci Workflow.Test**（source_file）：/ DAR-914 contract tests. Verifies the CI workflow at .github/workflows/ci.yml , the reproducible branch-protection script at scripts/setup-branch-protection.sh , and contributor documentation. The contract tests assert structural properties of the committed artifacts -- they do not invoke gh api against the live GitHub repo live state requires admin token + network access; verifying the artifact is the contract . / 证据：`tests/ci-workflow.test.ts`
- **Derive Dist Tag.Test**（source_file）：/ DAR-960 contract tests: dist-tag derivation script. The script lives at scripts/derive-dist-tag.sh . It reads a version string the package version stripped of the leading v from the git tag and prints a single dist-tag value on stdout. The release workflow uses its stdout as the value passed to pnpm publish --tag . Living in a script rather than inline shell in the workflow makes it unit-testable -- which is exactly what this file exercises. Rule: strip the leading v the workflow does this before invoking the script -- the script itself rejects a leading v as malformed , then look at any pre-release identifier after the first - . If absent, dist-tag is latest . If present, dist-tag is the… 证据：`tests/derive-dist-tag.test.ts`
- **Embedder.Integration.Test**（source_file）：/ DAR-912 contract integration tests. These tests load the real Xenova/bge-base-en-v1.5 model via @huggingface/transformers and run an actual embed round trip. The first run pulls weights from the HF hub ~6s cold start, see AC-2 ; subsequent runs are warm thanks to transformers.js's local cache. The wall-clock latency is intentionally NOT asserted -- AC-2's "~6s" is descriptive guidance, not a measurable target see contract envelope's explicit non goals . What IS asserted: - the returned vector is a Float32Array of length 768 - the vector's L2 norm is within 1e-3 of 1.0 so cosine == dot product - dim reads back as 768 and modelId round-trips - dim equals the returned vector's length after a… 证据：`tests/embedder.integration.test.ts`
- **Embedder.Readonly Types**（source_file）：/ DAR-912: type-level guard that Embedder.modelId and Embedder.dim are declared readonly . This file is not a runtime test -- it participates in tsc --noEmit and therefore make typecheck . The // @ts-expect-error markers below FAIL the build if either property becomes assignable, which is exactly the readonly contract required by AC-1. The test runner does not execute this file; it is only here so the compiler sees the intentionally-erroneous assignments. / 证据：`tests/embedder.readonly-types.ts`
- **Embedder.Test**（source_file）：/ DAR-912 contract tests unit . Behavioral tests for the Embedder wrapper around @huggingface/transformers . The pipeline factory is mocked so these tests run hermetically without pulling real model weights from the HF hub. The integration counterparts real model load + embed live in tests/embedder.integration.test.ts . Test names mirror the contract envelope on DAR-912 round 1, approved . / 证据：`tests/embedder.test.ts`
- **Global Setup**（source_file）：/ Vitest globalSetup: serially warm the transformers.js model cache for the embedder used by the spawned-bin and embedder integration tests. Why this exists DAR-955 root cause The DAR-919 spawned-bin integration test was flaky on cold-cache make test runs. Surface symptoms varied: - MCP error -32000: Connection closed the bot review report on PR 15 / DAR-929 , and - a non-error CallToolResult with isError: true whose payload read Load model from .../onnx/model.onnx failed:Protobuf parsing failed. Both symptoms have the same cause: vitest defaults to running test files in parallel forked workers. Two integration tests embedder.integration.test.ts and server-bin.integration.test.ts each load… 证据：`tests/global-setup.ts`
- **Graph Bin.Integration.Test**（source_file）：/ DAR-933 ac-1 / ac-3 / ac-6 spawned-bin coverage: prove the built commonplace bin dispatches the graph subcommand and produces a mermaid / json / dot rendering for a fixture memory dir. The in-process suite graph-cli.test.ts drives graphMain directly with a stub embedder so it can run fast. This file exists so the spawned-bin wiring -- the dispatcher's argv 0 === 'graph' short- circuit, the embedder factory, the real MemoryStore.scan cold-start -- is covered end-to-end. A missing wire would surface here as a non-zero exit or an empty stdout despite an in-process test passing. Slow on purpose: pays the real Embedder cold-start price. / 证据：`tests/graph-bin.integration.test.ts`
- **Graph Cli.Test**（source_file）：/ DAR-933 contract tests for the commonplace graph subcommand. Test names mirror the approved contract envelope on DAR-933. The CLI surface is broken across: - argv parsing: parseGraphArgs - mermaid / json / dot renderers - dispatcher integration bare bin + unknown-subcommand error paths - end-to-end via the in-process graphMain the spawned-bin form lives in tests/graph-bin.integration.test.ts Snapshot fixtures live under tests/fixtures/graph/ ; each case has one .json describing the memories and one expected-output file per format .mermaid , .json , .dot . / 证据：`tests/graph-cli.test.ts`
- **Graph.Test**（source_file）：/ DAR-926 contract tests. Behavioral tests for the MemoryGraph class -- an in-memory adjacency structure built from the relations and supersedes frontmatter graph fields DAR-925 layered over the MemoryStore entries DAR-916 . Test names mirror the contract envelope on DAR-926 round 1, approved . / 证据：`tests/graph.test.ts`
- **Memory Graph.Test**（source_file）：/ DAR-925 contract tests. Behavioral tests for the relations and supersedes frontmatter graph fields layered on top of the DAR-911 memory I/O primitives: - readMemory path - { ..., relations, supersedes } - writeMemory path, memory accepts relations / supersedes - contentSha memory MUST NOT depend on relations / supersedes Test names mirror the contract envelope on DAR-925 round 1 . / 证据：`tests/memory-graph.test.ts`
- **Memory Store Multiprocess.Test**（source_file）：/ DAR-923 contract tests for MemoryStore integration with the multi-process safety primitives: - ac-1: every .md and .embedding write goes through the atomic helper - ac-3: per-name advisory locks around save and delete - ac-4: mtime-based external-writer rescan in search and list/all Test names mirror the contract envelope on DAR-923 round 1, approved . / 证据：`tests/memory-store-multiprocess.test.ts`
- **Memory Store Search.Test**（source_file）：/ DAR-917 contract tests. Behavioral tests for MemoryStore.search query, opts? -- brute-force top-k cosine search over the in-memory index. Vectors in the store are normalized at write time DAR-916 , so cosine reduces to a dot product. Test names mirror the contract envelope on DAR-917 round 1, approved . The Embedder dependency is stubbed no real model load so these tests run hermetically and quickly. Stubs return small deterministic vectors so dot products and ordering can be hand-computed. / 证据：`tests/memory-store-search.test.ts`
- **Memory Store Stress.Test**（source_file）：/ DAR-923 contract tests for multi-process stress + crash safety: - ac-5: 4 child processes each saving 50 distinct memories produce 200 valid .md, .embedding pairs with no errors and no corruption. - ac-6: 2 child processes racing on the same name resolve to exactly one winner and one clear "already exists" / lock-busy loser, with no corruption of the winner's files. - ac-7: SIGKILL'ing a process mid-write to an existing memory leaves the prior files byte-equal, leaves no orphan tmpfile, and a fresh scan in a new process reads the prior memory without throwing. Each test orchestrates one or more child processes via node:child process running tests/fixtures/save-worker.ts under tsx. The wor… 证据：`tests/memory-store-stress.test.ts`
- **Memory Store.Test**（source_file）：/ DAR-916 contract tests. Behavioral tests for the MemoryStore class -- the in-memory vector index backed by / .md + .embedding sidecar files. Test names mirror the contract envelope on DAR-916 round 1, approved . The Embedder dependency is stubbed no real model load so these tests run hermetically and quickly. The stub satisfies the structural contract the MemoryStore depends on: modelId , dim , and embed text . / 证据：`tests/memory-store.test.ts`
- **Memory.Test**（source_file）：/ DAR-911 contract tests. Behavioral tests for the memory .md file I/O primitives: - readMemory path - { name, description, type, body, raw } - writeMemory path, memory - contentSha memory - 64-char lowercase sha256 hex Test names mirror the contract envelope on DAR-911 round 1 . / 证据：`tests/memory.test.ts`
- **Mentions Store.Test**（source_file）：/ DAR-927 contract tests -- MemoryStore wiring. Behavioral tests that exercise the integration of extractMentions into MemoryStore.scan and MemoryStore.save : each extracted name becomes one mentions edge in the configured MemoryGraph , gated by the env var COMMONPLACE EXTRACT MENTIONS default on . Test names mirror the contract envelope on DAR-927 round 1, approved . / 证据：`tests/mentions-store.test.ts`
- **Mentions.Test**（source_file）：/ DAR-927 contract tests -- pure tokenizer extractMentions . Behavioral tests for the name body-mention extractor. Test names mirror the contract envelope on DAR-927 round 1, approved . The tokenizer is a pure function over a markdown body string. It returns unique mention names in first-occurrence order, restricting matches to the same ^ a-z0-9 +$ rule used by memory file names DAR-911 / DAR-925 . / 证据：`tests/mentions.test.ts`
- **Migrate Bin.Integration.Test**（source_file）：/ DAR-918 ac-7: spawn the built commonplace bin with the migrate subcommand and assert that: - migrate against a fixture directory produces the expected sidecars, prints a non-empty summary on stdout containing the labels 'loaded', 'embedded', 're-embedded', and 'orphaned' along with their counts, and exits 0. - migrate with no positional argument prints a usage message to stderr and exits non-zero. The build is performed by tests/scaffolding.integration.test.ts via the same make build target; we re-run it here defensively because vitest does not order test files. The bin's first invocation pays the transformers.js model load cost ~6-12s on a warm cache, longer cold . The vitest globalSetup… 证据：`tests/migrate-bin.integration.test.ts`
- **Migrate Import.Test**（source_file）：/ DAR-961 contract tests for the auto-migration importer: - commonplace migrate no args detects Claude Code project-memory directories at ~/.claude/projects/ \/memory/ and reports counts without writing anything. - commonplace migrate --from claude-code copies each compatible .md to / .md , then runs the existing scan/embed pass so each imported file gets its .embedding sidecar. - Conflict policy is skip-and-report default , preserving existing target files byte-identical. - --dry-run reports what would be imported without writing. Test names mirror the approved contract envelope on DAR-961. / 证据：`tests/migrate-import.test.ts`
- **Migrate Readme.Test**（source_file）：/ DAR-961 ac-7: README sanity tests for the migration section. Asserts the migration documentation covers: - the commonplace migrate entry-point with --from claude-code , --auto , and --dry-run - the skip-and-report conflict policy - the "Migrating from mem0 / Letta" dual-MCP-server pattern, with an explicit note that no commonplace-side integration code is required. These are doc-drift guards rather than prose-quality assertions: each test pins one concrete claim that the issue body says must be present, so a future README rewrite that drops the claim fails the test rather than silently regressing. / 证据：`tests/migrate-readme.test.ts`
- **Migrate Resilience.Test**（source_file）：/ DAR-966 contract tests: migrate-path resilience to permissively-formatted harness frontmatter, MEMORY.md exclusion, structured skip reasons for unrecoverable source files, and skip-and-warn semantics on MemoryStore.scan . Test names mirror the approved contract envelope on DAR-966. / 证据：`tests/migrate-resilience.test.ts`
- **Migrate.Test**（source_file）：/ DAR-918 contract tests for runMigrate programmatic API and the argv-parsing wrapper. Test names mirror the approved contract envelope on DAR-918. / 证据：`tests/migrate.test.ts`
- **Readme.Test**（source_file）：/ README sanity tests. Two narrow goals: 1. Catch drift between README and code: every inputSchema property of every registered MCP tool, and every COMMONPLACE env var the bin reads, must appear somewhere in the README. If a property is renamed or removed, the test fails until the README catches up. 2. Verify the verbatim install commands the publish smoke test DAR-921 depends on are present. The publish smoke test executes these strings exactly; if either drifts in the README, users will follow stale instructions. Everything else about README quality concept blurb, prose, section ordering, emoji-free, License/Contributing presence is a review-time concern, not a unit-test concern. Assertin… 证据：`tests/readme.test.ts`
- **Release Workflow.Test**（source_file）：/ DAR-960 contract tests: release workflow YAML structure. Verifies the release workflow at .github/workflows/release.yml and the release-process docs in CONTRIBUTING.md . Mirrors the structural-assertion pattern used in tests/ci-workflow.test.ts DAR-914 : we parse the YAML with the yaml package and assert structural properties of the committed artifact rather than executing it against the live GitHub repo. / 证据：`tests/release-workflow.test.ts`
- **Scaffolding.Integration.Test**（source_file）：/ DAR-908 contract integration tests. These shell out to make and assert exit codes / stdout for the required targets. A few targets are intentionally NOT spawned from inside vitest: - make test would recurse vitest invoking vitest . - make install mutates node modules under us mid-run. - make audit depends on the live registry and is flaky as a unit test; the recipe correctness is asserted statically. Their recipes are validated structurally in scaffolding.test.ts and operationally by CI invoking the targets directly. They are declared in the implementation envelope's untested with reasons. / 证据：`tests/scaffolding.integration.test.ts`
- **Scaffolding.Test**（source_file）：/ DAR-908 contract tests. These tests verify the scaffolding contract: package metadata, TS/ESM config, src/ skeleton, Makefile-driven dev loop, lint/format/editor configs, license, gitignore, and the populated CLAUDE.md. They are "unit" in the sense of assertions over file contents on disk; a separate suite scaffolding.integration.test.ts covers the make -target integration checks. / 证据：`tests/scaffolding.test.ts`
- **Scope.Test**（source_file）：/ DAR-924 ac-2 / ac-7: unit tests for the scope-detection function. Each detection branch env / roots / cwd / none is exercised independently of the spawned bin so the priority order can be asserted without booting a real MCP server. / 证据：`tests/scope.test.ts`
- **Server Bin Cold Start.Integration.Test**（source_file）：/ DAR-955 cold-start invariants for the spawned-bin integration harness. Background tests/server-bin.integration.test.ts DAR-919 was flaky on make test runs from a cold transformers.js model cache: the first invocation occasionally failed with one of two surface symptoms, - MCP error -32000: Connection closed child process died , or - a CallToolResult with isError: true whose payload reads Load model from .../onnx/model.onnx failed:Protobuf parsing failed. Both symptoms have the same underlying cause: Concurrent transformers.js downloads to a shared on-disk model cache corrupt the in-flight files. Two vitest workers this file, the bin integration test, and the in-process embedder integratio… 证据：`tests/server-bin-cold-start.integration.test.ts`
- **Server Bin Graph Wiring.Test**（source_file）：/ DAR-928 ac-5 unit-level assertion: the boot module constructs a MemoryGraph , passes it to MemoryStore { dir, embedder, graph } , and passes it to createDefaultHandlers { ..., graph } . We assert structurally on the source text. The end-to-end behavioural proof that the graph is actually wired i.e. populated by save/scan and available to link/unlink lives in server-bin-link.integration.test.ts . Note: as of DAR-924 the wiring lives in src/bin/boot.ts rather than src/bin/commonplace-mcp.ts the bin reduces to a thin shell that delegates to bootServer . The contract is unchanged -- a graph is constructed and threaded through MemoryStore + createDefaultHandlers -- just relocated. / 证据：`tests/server-bin-graph-wiring.test.ts`
- **Server Bin Graph.Integration.Test**（source_file）：/ DAR-932 ac-7 spawned-bin integration: end-to-end coverage for memory graph and memory path against the real bin and its real MemoryGraph instance. This proves the boot path wires both per-scope graphs into the new handler factories so the running server can serve traversal queries -- a missing wire-up would surface as UnknownToolError, 'not implemented', or empty traversal results despite memory link having succeeded. Slow on purpose: pays the real Embedder cold-start price. / 证据：`tests/server-bin-graph.integration.test.ts`
- **Server Bin Link.Integration.Test**（source_file）：/ DAR-928 bin integration test: spawn the built commonplace-mcp bin and exercise memory link / memory unlink end-to-end via real MCP stdio framing. This proves: - ac-5: the bin instantiates a MemoryGraph , wires it into both the store and the handler map, and the running server's graph is populated by scan / save verified by linking between two memory save'd entries -- a missing graph or a graph not populated by save would surface as "target memory does not exist" . - ac-6: round trip save A, save B, link A- B, hand-inspect A.md to verify the edge appears with valid YAML, then unlink and confirm the edge is gone and frontmatter is still valid YAML. Slow on purpose -- pays the real Embedder… 证据：`tests/server-bin-link.integration.test.ts`
- 其余 20 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：`CLAUDE.md`, `README.md`, `package.json`
- **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：`CLAUDE.md`, `README.md`, `package.json`

## 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**：importance `high`
  - source_paths: README.md, package.json, CLAUDE.md
- **快速入门与安装**：importance `high`
  - source_paths: README.md, package.json, .nvmrc
- **系统架构**：importance `high`
  - source_paths: CLAUDE.md, src/index.ts, src/embedder/index.ts, src/store/memory-store.ts, src/server/server.ts
- **技术栈详解**：importance `medium`
  - source_paths: package.json, tsconfig.json, tsconfig.build.json, vitest.config.ts, eslint.config.js
- **嵌入模型与语义搜索**：importance `high`
  - source_paths: src/embedder/index.ts, vitest.config.ts, tests/global-setup.ts
- **内存格式与侧载文件**：importance `high`
  - source_paths: docs/sidecar-format.md, src/store/memory.ts, src/store/sidecar.ts, src/store/memory-store.ts
- **搜索功能详解**：importance `high`
  - source_paths: README.md, src/server/defaults.ts, src/server/tools.ts, src/server/handlers.ts
- **图功能与关系管理**：importance `medium`
  - source_paths: src/store/graph.ts, src/store/mentions.ts, src/cli/graph.ts

## Repo Inspection Evidence / 源码检查证据

- repo_clone_verified: true
- repo_inspection_verified: true
- repo_commit: `bda917d02f7cf739fc731a3f895316b118ec2ba5`
- inspected_files: `pnpm-lock.yaml`, `package.json`, `README.md`, `docs/sidecar-format.md`, `src/index.ts`, `src/cli/migrate.ts`, `src/cli/graph.ts`, `src/bin/env.ts`, `src/bin/commonplace-mcp.ts`, `src/bin/scope.ts`, `src/bin/boot.ts`, `src/embedder/index.ts`, `src/server/defaults.ts`, `src/server/tools.ts`, `src/server/index.ts`, `src/server/server.ts`, `src/server/handlers.ts`, `src/store/mentions.ts`, `src/store/sidecar.ts`, `src/store/memory-store.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: 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 | github_repo:1232879661 | https://github.com/rickbassham/commonplace | 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 2: 运行可能依赖外部服务

- Trigger: 项目说明出现 external service/cloud/webhook/database 等运行依赖关键词。
- Host AI rule: 确认是否有离线 demo、mock 数据或可替代服务。
- Why it matters: 本地安装成功不等于能力可用，外部服务不可用会阻断体验。
- Evidence: packet_text.keyword_scan | github_repo:1232879661 | https://github.com/rickbassham/commonplace | matched external service / cloud / webhook / database keyword
- 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 | github_repo:1232879661 | https://github.com/rickbassham/commonplace | 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。

### Constraint 4: 下游验证发现风险项

- Trigger: no_demo
- Host AI rule: 进入安全/权限治理复核队列。
- Why it matters: 下游已经要求复核，不能在页面中弱化。
- Evidence: downstream_validation.risk_items | github_repo:1232879661 | https://github.com/rickbassham/commonplace | 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: 存在安全注意事项

- Trigger: No sandbox install has been executed yet; downstream must verify before user use.
- Host AI rule: 转成明确权限清单和安全审查提示。
- Why it matters: 用户安装前需要知道权限边界和敏感操作。
- Evidence: risks.safety_notes | github_repo:1232879661 | https://github.com/rickbassham/commonplace | No sandbox install has been executed yet; downstream must verify before user use.
- 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: 存在评分风险

- Trigger: no_demo
- Host AI rule: 把风险写入边界卡，并确认是否需要人工复核。
- Why it matters: 风险会影响是否适合普通用户安装。
- Evidence: risks.scoring_risks | github_repo:1232879661 | https://github.com/rickbassham/commonplace | 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 7: 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 | github_repo:1232879661 | https://github.com/rickbassham/commonplace | 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 8: 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 | github_repo:1232879661 | https://github.com/rickbassham/commonplace | 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。