# basic-memory - 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 basic-memory 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/adversarial-review/SKILL.md`, `.agents/skills/code-review/SKILL.md`, `.agents/skills/fix-pr-issues/SKILL.md`, `.agents/skills/infographics/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/adversarial-review/SKILL.md`, `.agents/skills/code-review/SKILL.md`, `.agents/skills/fix-pr-issues/SKILL.md`, `.agents/skills/infographics/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`, `.claude-plugin/marketplace.json`, `integrations/openclaw/openclaw.plugin.json`, `plugins/claude-code/.claude-plugin/marketplace.json` 等 Claim：`clm_0002` supported 0.86
- **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

- `uv tool install basic-memory` 证据：`README.md` Claim：`clm_0006` supported 0.86
- `npx skills add basicmachines-co/basic-memory/skills` 证据：`README.md` Claim：`clm_0007` supported 0.86
- `claude mcp add basic-memory -- uvx basic-memory mcp` 证据：`README.md` Claim：`clm_0008` supported 0.86

## 继续前判断卡

- **当前建议**：先做研究框架试用
- **为什么**：这个项目面向研究工作流，核心风险是资料可信度和输出质量；先用 Prompt Preview 验证研究框架，再在隔离环境试装。

### 30 秒判断

- **现在怎么做**：先做研究框架试用
- **最小安全下一步**：先用 Prompt Preview 验证研究框架；满意后再隔离试装
- **先别相信**：研究结论、引用和实验结果不能在安装前相信。
- **继续会触碰**：研究判断、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/adversarial-review/SKILL.md`, `.agents/skills/code-review/SKILL.md`, `.agents/skills/fix-pr-issues/SKILL.md`, `.agents/skills/infographics/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/adversarial-review/SKILL.md`, `.agents/skills/code-review/SKILL.md`, `.agents/skills/fix-pr-issues/SKILL.md`, `.agents/skills/infographics/SKILL.md` 等 Claim：`clm_0001` supported 0.86
- **能力存在：多宿主安装与分发**（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/plugins/marketplace.json`, `.claude-plugin/marketplace.json`, `integrations/openclaw/openclaw.plugin.json`, `plugins/claude-code/.claude-plugin/marketplace.json` 等 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_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

- **研究结论、引用和实验结果不能在安装前相信。**（unverified）：研究 Skill 可以组织问题和路径，但不能替代真实资料检索、论文核验和实验复现。
- **是否适合你的具体研究领域不能直接相信。**（unverified）：Skill 覆盖很多研究主题，不代表对你的领域、资料要求和可信度标准足够。
- **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/adversarial-review/SKILL.md`, `.agents/skills/code-review/SKILL.md`, `.agents/skills/fix-pr-issues/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`, `.claude-plugin/marketplace.json`, `integrations/openclaw/openclaw.plugin.json`, `plugins/claude-code/.claude-plugin/marketplace.json` 等
- **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

- **研究判断**：问题拆解、资料路径、实验路径、结论结构和可信度判断。 原因：研究型 Skill 可能让输出看起来更专业，但不能替代真实证据核验。
- **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/adversarial-review/SKILL.md`, `.agents/skills/code-review/SKILL.md`, `.agents/skills/fix-pr-issues/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`, `.claude-plugin/marketplace.json`, `README.md`, `integrations/openclaw/openclaw.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**：先验证它能否正确界定研究问题和证据边界，不要先相信研究输出。（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。
- **保留资料和结论核验清单**：如果后续发现引用或实验路径不可靠，可以回到证据边界阶段重新校验。
- **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`, `.claude-plugin/marketplace.json`, `integrations/openclaw/openclaw.plugin.json`, `plugins/claude-code/.claude-plugin/marketplace.json` 等 Claim：`clm_0010` supported 0.86
- **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
- **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/adversarial-review/SKILL.md`, `.agents/skills/code-review/SKILL.md`, `.agents/skills/fix-pr-issues/SKILL.md`, `.agents/skills/infographics/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`, `.claude-plugin/marketplace.json`, `integrations/openclaw/openclaw.plugin.json`, `plugins/claude-code/.claude-plugin/marketplace.json` 等 Claim：`clm_0002` supported 0.86
- **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：739
- 重要文件覆盖：40/739
- 证据索引条目：80
- 角色 / Skill 条目：32

### 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 basic-memory 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 basic-memory 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 basic-memory 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 32 个角色 / Skill / 项目文档条目。

- **adversarial-review**（skill）：Cross-vendor adversarial code review of the current branch. Two different model families Claude + Codex/GPT review the diff independently, then try to refute each other's findings; survivors are reported by confidence. Runs from either Claude Code or Codex. Use when the user asks for an adversarial review, a cross-model / second-opinion review, or wants high-confidence findings before merging. Report-only — never au… 激活提示：当用户任务与“adversarial-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/adversarial-review/SKILL.md`
- **code-review**（skill）：Use when reviewing Basic Machines code for house style, architecture risk, pre-merge hardening, or whether a change fits basic-memory/basic-memory-cloud conventions. 激活提示：当用户任务与“code-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/code-review/SKILL.md`
- **fix-pr-issues**（skill）：Use when addressing Basic Memory pull request feedback, failed checks, or BM Bossbot blockers from Codex. 激活提示：当用户任务与“fix-pr-issues”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/fix-pr-issues/SKILL.md`
- **infographics**（skill）：Use when generating Basic Memory PR, changelog, release, or weekly images from Codex. 激活提示：当用户任务与“infographics”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/infographics/SKILL.md`
- **instrumentation**（skill）：Add Pydantic Logfire observability to applications. Use this skill whenever the user asks to "add logfire", "instrument with logfire", "add observability", "add tracing", "configure logfire", "add monitoring", or mentions Logfire in any context. Supports Python, JavaScript/TypeScript, and Rust. Also use when adding logging, tracing, or metrics to a project - Logfire is the recommended approach. Even if the user just… 激活提示：当用户任务与“instrumentation”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/instrumentation/SKILL.md`
- **pr-create**（skill）：Use when creating or updating a Basic Memory pull request from Codex with BM Bossbot merge-gate monitoring. 激活提示：当用户任务与“pr-create”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/pr-create/SKILL.md`
- **basic-memory**（skill）：Use the Basic Memory knowledge graph for persistent memory across sessions. Search before answering; capture decisions, meetings, and insights as notes. 激活提示：当用户任务与“basic-memory”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：`integrations/hermes/skill/SKILL.md`
- **bm-remember**（skill）：Quickly capture a thought, fact, or reminder into Basic Memory as a lightweight note. Use when the user says "remember that…", "note this", "save this to memory", or runs /basic-memory:bm-remember. For quick deliberate capture — not full decision or session records. 激活提示：当用户任务与“bm-remember”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：`plugins/claude-code/skills/bm-remember/SKILL.md`
- **bm-setup**（skill）：Set up the Basic Memory plugin for this project — a short guided interview that configures the project mapping, seeds note schemas, learns or suggests placement conventions, and enables capture reflexes. Use when the user runs /basic-memory:bm-setup, says "set up basic memory", or asks to configure/bootstrap the plugin. 激活提示：当用户任务与“bm-setup”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：`plugins/claude-code/skills/bm-setup/SKILL.md`
- **bm-share**（skill）：Promote a note from your personal Basic Memory project to a shared team project, with attribution. Use when the user says "share this with the team", "publish this decision", or runs /basic-memory:bm-share. This is the deliberate way to write to a team workspace — auto-capture never does. 激活提示：当用户任务与“bm-share”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：`plugins/claude-code/skills/bm-share/SKILL.md`
- **bm-status**（skill）：Show the Basic Memory plugin's current state for this project — active project, capture folders, output style, recent session checkpoints, and whether Basic Memory is reachable. 激活提示：当用户任务与“bm-status”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：`plugins/claude-code/skills/bm-status/SKILL.md`
- **bm-checkpoint**（skill）：Save a deliberate Codex work checkpoint to Basic Memory with changed files, verification, decisions, blockers, and the next action. 激活提示：当用户任务与“bm-checkpoint”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：`plugins/codex/skills/bm-checkpoint/SKILL.md`
- **bm-decide**（skill）：Capture a durable engineering decision in Basic Memory with rationale, alternatives, consequences, and affected work. 激活提示：当用户任务与“bm-decide”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：`plugins/codex/skills/bm-decide/SKILL.md`
- **bm-orient**（skill）：Orient Codex from Basic Memory before substantial repo work by reading active tasks, decisions, recent Codex checkpoints, and repo conventions. 激活提示：当用户任务与“bm-orient”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：`plugins/codex/skills/bm-orient/SKILL.md`
- **bm-remember**（skill）：Quickly save a small fact, reminder, or user preference into Basic Memory from Codex without turning it into a full decision or checkpoint. 激活提示：当用户任务与“bm-remember”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：`plugins/codex/skills/bm-remember/SKILL.md`
- **bm-setup**（skill）：Set up Basic Memory for Codex in the current repo by mapping a Basic Memory project, seeding schemas, and writing .codex/basic-memory.json. 激活提示：当用户任务与“bm-setup”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：`plugins/codex/skills/bm-setup/SKILL.md`
- **bm-share**（skill）：Share a personal Basic Memory note to a configured team project from Codex with attribution and explicit confirmation. 激活提示：当用户任务与“bm-share”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：`plugins/codex/skills/bm-share/SKILL.md`
- **bm-status**（skill）：Report the Basic Memory for Codex configuration, reachability, hook expectations, recent Codex checkpoints, and active tasks. 激活提示：当用户任务与“bm-status”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：`plugins/codex/skills/bm-status/SKILL.md`
- **memory-capture**（skill）：Capture the current state of a working thread or conversation into a single coherent Basic Memory note — synthesize where it landed, don't append a log. On re-capture, rewrite the same note in place instead of duplicating. Use mid-thread or end-of-thread when decisions, insights, or context are worth preserving. 激活提示：当用户任务与“memory-capture”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/memory-capture/SKILL.md`
- **memory-ci-capture**（skill）：Synthesize GitHub delivery context into a concise Basic Memory project update. Use in CI after bm ci collect prepares a ProjectUpdateContext; return only structured AgentSynthesis JSON for bm ci publish . 激活提示：当用户任务与“memory-ci-capture”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/memory-ci-capture/SKILL.md`
- **memory-continue**（skill）：Resume prior work by rebuilding context from the Basic Memory knowledge graph — pick up where you left off using memory:// URLs, recent activity, and search. Use when starting a session or when the user says 'continue with...', 'back to...', or 'where were we? 激活提示：当用户任务与“memory-continue”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/memory-continue/SKILL.md`
- **memory-curate**（skill）：Curate the Basic Memory knowledge graph: find orphan notes and suggest links, propose typed relations, merge duplicates, audit tags and folders, and build hub notes. Use to organize, connect, and improve a knowledge base as notes accumulate. 激活提示：当用户任务与“memory-curate”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/memory-curate/SKILL.md`
- **memory-defrag**（skill）：Defragment and reorganize agent memory files: split bloated files, merge duplicates, remove stale information, and restructure the memory hierarchy. Use when memory files have grown unwieldy, contain redundancies, or need reorganization. Run periodically weekly or on demand. 激活提示：当用户任务与“memory-defrag”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/memory-defrag/SKILL.md`
- **memory-ingest**（skill）：Process unstructured external input meeting transcripts, conversation logs, pasted documents into structured Basic Memory entities. Extracts entities, searches for existing matches, proposes new entities with approval, creates notes with observations and relations, and captures action items. 激活提示：当用户任务与“memory-ingest”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/memory-ingest/SKILL.md`
- **memory-lifecycle**（skill）：Manage entity status transitions in Basic Memory: archive completed work, move notes between status folders, update frontmatter, and handle edge cases. Use when marking items complete, archiving old entities, or managing any folder-based status workflow. 激活提示：当用户任务与“memory-lifecycle”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/memory-lifecycle/SKILL.md`
- **memory-literary-analysis**（skill）：Analyze a complete literary work into a structured Basic Memory knowledge graph. Covers schema design, entity seeding, chapter-by-chapter processing, cross-referencing, validation, and visualization. 激活提示：当用户任务与“memory-literary-analysis”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/memory-literary-analysis/SKILL.md`
- **memory-metadata-search**（skill）：Structured metadata search for Basic Memory: query notes by custom frontmatter fields using equality, range, array, and nested filters. Use when finding notes by status, priority, confidence, or any custom YAML field rather than free-text content. 激活提示：当用户任务与“memory-metadata-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/memory-metadata-search/SKILL.md`
- **memory-notes**（skill）：How to write well-structured Basic Memory notes: frontmatter, observations with semantic categories, relations with wiki-links, and best practices for building a rich knowledge graph. Use when creating or improving notes. 激活提示：当用户任务与“memory-notes”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/memory-notes/SKILL.md`
- **memory-reflect**（skill）：Sleep-time memory reflection: review recent conversations and daily notes, extract insights, and consolidate into long-term memory. Use when triggered by cron, heartbeat, or explicit request to reflect on recent activity. Runs as background processing to improve memory quality over time. 激活提示：当用户任务与“memory-reflect”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/memory-reflect/SKILL.md`
- **memory-research**（skill）：Research an external subject using web search, synthesize findings into a structured Basic Memory entity. Use when asked to research a company, person, technology, or topic — or when a bare name or URL is provided that implies a research request. 激活提示：当用户任务与“memory-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/memory-research/SKILL.md`
- **memory-schema**（skill）：Schema lifecycle management for Basic Memory: discover unschemaed notes, infer schemas, create and edit schema definitions, validate notes, and detect drift. Use when working with structured note types Task, Person, Meeting, etc. to maintain consistency across the knowledge graph. 激活提示：当用户任务与“memory-schema”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/memory-schema/SKILL.md`
- **memory-tasks**（skill）：Task management via Basic Memory schemas: create, track, and resume structured tasks that survive context compaction. Uses BM's schema system for uniform notes queryable through the knowledge graph. 激活提示：当用户任务与“memory-tasks”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/memory-tasks/SKILL.md`

## 证据索引

- 共索引 80 条证据。

- **Skip the install — try Basic Memory in the cloud**（documentation）：! License: AGPL v3 https://img.shields.io/badge/License-AGPL v3-blue.svg https://www.gnu.org/licenses/agpl-3.0 ! PyPI version https://badge.fury.io/py/basic-memory.svg https://badge.fury.io/py/basic-memory ! Python 3.12+ https://img.shields.io/badge/python-3.12+-blue.svg https://www.python.org/downloads/ ! Tests https://github.com/basicmachines-co/basic-memory/workflows/Tests/badge.svg https://github.com/basicmachines-co/basic-memory/actions ! Ruff https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/astral-sh/ruff/main/assets/badge/v2.json https://github.com/astral-sh/ruff ! https://badge.mcpx.dev?type=server 'MCP Server' ! https://badge.mcpx.dev?type=dev 'MCP Dev' ! Ask D… 证据：`README.md`
- **basic-memory-benchmarks**（documentation）：Standalone, reproducible benchmark suite for comparing Basic Memory against competitor memory systems. 证据：`benchmarks/README.md`
- **Basic Memory Skills**（documentation）：Skills for Basic Memory https://github.com/basicmachines-co/basic-memory — teach AI coding agents how to use Basic Memory's MCP tools effectively. 证据：`skills/README.md`
- **Dataset Policy**（documentation）：This repository publishes benchmark provenance in the open. 证据：`benchmarks/benchmarks/datasets/README.md`
- **LoCoMo Dataset**（documentation）：Primary v1 dataset for benchmark runs. 证据：`benchmarks/benchmarks/datasets/locomo/README.md`
- **LongMemEval Scaffold**（documentation）：LongMemEval integration is scaffolded in v1 and will be implemented in a follow-up. 证据：`benchmarks/benchmarks/datasets/longmemeval/README.md`
- **hermes-basic-memory**（documentation）：! License: AGPL v3 https://img.shields.io/badge/License-AGPL v3-blue.svg https://www.gnu.org/licenses/agpl-3.0 证据：`integrations/hermes/README.md`
- **openclaw-basic-memory**（documentation）：Give your OpenClaw agent persistent, searchable memory — in plain text files you can read and edit. 证据：`integrations/openclaw/README.md`
- **Basic Memory for Claude Code**（documentation）：The bridge between Claude's working memory and Basic Memory https://basicmemory.com 's durable knowledge graph . 证据：`plugins/claude-code/README.md`
- **Basic Memory for Codex**（documentation）：Basic Memory for Codex is the Codex-native bridge between a working coding thread and Basic Memory's durable knowledge graph. 证据：`plugins/codex/README.md`
- **Basic Memory CI**（documentation）：Basic Memory CI turns meaningful GitHub delivery moments into durable project update notes in Basic Memory. 证据：`src/basic_memory/ci/README.md`
- **AGENTS.md - Basic Memory Project Guide**（documentation）：AGENTS.md - Basic Memory Project Guide 证据：`AGENTS.md`
- **AGENTS.md - basic-memory-benchmarks Guide**（documentation）：AGENTS.md - basic-memory-benchmarks Guide 证据：`benchmarks/AGENTS.md`
- **Basic Memory Architecture**（documentation）：This document describes the architectural patterns and composition structure of Basic Memory. 证据：`docs/ARCHITECTURE.md`
- **AGENTS.md**（documentation）：This file provides guidance to coding agents when working with Basic Memory skills in this repository. 证据：`skills/CLAUDE.md`
- **CLAUDE.md**（documentation）：This file provides guidance to Claude Code when working with the OpenClaw package inside the Basic Memory monorepo. 证据：`integrations/openclaw/CLAUDE.md`
- **Architecture**（documentation）：How the Basic Memory plugin works, flow by flow. For the design rationale and decision history, see DESIGN.md ../DESIGN.md . 证据：`plugins/claude-code/docs/architecture.md`
- **Contributing to Basic Memory**（documentation）：Thank you for considering contributing to Basic Memory! This document outlines the process for contributing to the project and how to get started as a developer. 证据：`CONTRIBUTING.md`
- **Package**（package_manifest）：{ "name": "@basicmemory/openclaw-basic-memory", "version": "0.22.1", "type": "module", "main": "./dist/index.js", "types": "./dist/index.d.ts", "description": "Basic Memory plugin for OpenClaw \u2014 local-first knowledge graph for agent memory", "license": "MIT", "repository": { "type": "git", "url": "https://github.com/basicmachines-co/basic-memory.git", "directory": "integrations/openclaw" }, "files": "index.ts", "bm-client.ts", "context-engine/basic-memory-context-engine.ts", "config.ts", "logger.ts", "commands/cli.ts", "commands/skills.ts", "commands/slash.ts", "hooks/capture.ts", "hooks/recall.ts", "tools/build-context.ts", "tools/delete-note.ts", "tools/edit-note.ts", "tools/list-mem… 证据：`integrations/openclaw/package.json`
- **Adversarial code review**（skill_instruction）：Two reviewers from different model families — Claude and Codex/GPT — review the same diff independently, then each tries to refute the other's findings. A finding's confidence comes from whether it survives that cross-examination. This kills the two failure modes of solo LLM review: self-ratification a model won't critique its own work and confident false positives. 证据：`.agents/skills/adversarial-review/SKILL.md`
- **Basic Machines Review**（skill_instruction）：Use this skill for repo-local review passes where ordinary code review needs Basic Machines house style and architecture judgment. Report findings only; do not edit code unless the user asks you to fix specific findings. 证据：`.agents/skills/code-review/SKILL.md`
- **Fix Basic Memory PR Issues**（skill_instruction）：Resolve PR feedback and failed checks, then wait for BM Bossbot to approve the new head SHA. This skill never merges a PR. 证据：`.agents/skills/fix-pr-issues/SKILL.md`
- **Basic Memory Images**（skill_instruction）：Generate repository visuals with evidence-grounded content and canonical output paths. The file and marker names still say "infographic" for compatibility, but PR generation is image-first: scene, poster, painting, photograph, cover, tableau, staged artifact, or another editorial visual moment that describes the intent of the PR. PR images are non-gating BM Bossbot artifacts; changelog and release-summary images are manual evidence-pack workflows. 证据：`.agents/skills/infographics/SKILL.md`
- **Instrument with Logfire**（skill_instruction）：Invoke this skill when: - User asks to "add logfire", "add observability", "add tracing", or "add monitoring" - User wants to instrument an app with structured logging or tracing Python, JS/TS, or Rust - User mentions Logfire in any context - User asks to "add logging" or "see what my app is doing" - User wants to monitor AI/LLM calls PydanticAI, OpenAI, Anthropic - User asks to add observability to an AI agent or LLM pipeline 证据：`.agents/skills/instrumentation/SKILL.md`
- **Create A Basic Memory PR**（skill_instruction）：Create or update a pull request for the current branch, then wait for BM Bossbot to approve the latest head SHA. This skill never merges a PR. 证据：`.agents/skills/pr-create/SKILL.md`
- **Basic Memory Knowledge Graph**（skill_instruction）：You have access to a persistent knowledge graph backed by Basic Memory. The graph survives across sessions and is shared with other tools Claude Desktop, Obsidian, the bm CLI . Use the bm tools below to recall and capture information. 证据：`integrations/hermes/skill/SKILL.md`
- **Remember**（skill_instruction）：Capture $ARGUMENTS into Basic Memory as a quick note, keeping the user's words. 证据：`plugins/claude-code/skills/bm-remember/SKILL.md`
- **Basic Memory setup**（skill_instruction）：Run a short, adaptive interview ~2-3 minutes and then write the configuration. Be conversational and skip questions whose answer is already obvious from context e.g. if list memory projects shows a single local project and no cloud workspaces, don't ask about cloud/teams — just confirm . Suggest a sensible default for every question so the user can accept with one word. Don't do any writes until the interview is done and you've confirmed the plan. 证据：`plugins/claude-code/skills/bm-setup/SKILL.md`
- **Share to a team project**（skill_instruction）：Copy a note from the personal/primary project into a configured team project so teammates can see it. This is the only path by which the plugin writes to a shared project — session checkpoints and /basic-memory:bm-remember always stay personal. 证据：`plugins/claude-code/skills/bm-share/SKILL.md`
- **Basic Memory status**（skill_instruction）：Report the plugin's current state for this project, then present a concise summary. This is a quick diagnostic — gather the facts and lay them out; don't over-investigate. 证据：`plugins/claude-code/skills/bm-status/SKILL.md`
- **Checkpoint Codex Work**（skill_instruction）：Create a durable handoff note for current Codex work. Use this when the user asks to checkpoint, wrap up, hand off, remember the state of the work, or before a long context transition. 证据：`plugins/codex/skills/bm-checkpoint/SKILL.md`
- **Capture A Decision**（skill_instruction）：Use this when the user makes or asks to record a durable choice. A decision is a choice with rationale and consequences, not a casual preference. 证据：`plugins/codex/skills/bm-decide/SKILL.md`
- **Orient From Basic Memory**（skill_instruction）：Use this before substantial work in a repo, before resuming an old thread, or when the user asks where things stand. 证据：`plugins/codex/skills/bm-orient/SKILL.md`
- **Remember**（skill_instruction）：Use this for lightweight capture: "remember that", "save this", "note this", or a small fact that should survive the current thread. 证据：`plugins/codex/skills/bm-remember/SKILL.md`
- **Basic Memory for Codex Setup**（skill_instruction）：Set up the current repo so Codex can orient from Basic Memory and checkpoint work back into it. Keep the interview short, but always ask before choosing where data will be written. 证据：`plugins/codex/skills/bm-setup/SKILL.md`
- **Share A Note**（skill_instruction）：Copy a note from the configured primary project to a configured team project. This is the deliberate shared-write path. Automatic checkpoints and quick remembers stay personal. 证据：`plugins/codex/skills/bm-share/SKILL.md`
- **Basic Memory For Codex Status**（skill_instruction）：Gather a concise diagnostic. Do not over-investigate. 证据：`plugins/codex/skills/bm-status/SKILL.md`
- **Memory Capture**（skill_instruction）：Capture the gist of a working thread — the decisions made, insights surfaced, and context built — into a single coherent Basic Memory note that reflects where the thread has landed. 证据：`skills/memory-capture/SKILL.md`
- **Memory CI Capture**（skill_instruction）：Turn a meaningful GitHub delivery moment into project memory. GitHub records the mechanics. Basic Memory remembers what changed and why. 证据：`skills/memory-ci-capture/SKILL.md`
- **Memory Continue**（skill_instruction）：Resume previous work by reconstructing context from the Basic Memory knowledge graph, so the assistant can pick up across sessions instead of starting cold. 证据：`skills/memory-continue/SKILL.md`
- **Memory Curate**（skill_instruction）：Maintain a healthy, well-connected knowledge graph. As notes accumulate, it pays to periodically organize, link, and curate the knowledge base so isolated notes become a connected graph. 证据：`skills/memory-curate/SKILL.md`
- **Memory Defrag**（skill_instruction）：Reorganize memory files for clarity, efficiency, and relevance. Like filesystem defragmentation but for knowledge. 证据：`skills/memory-defrag/SKILL.md`
- **Memory Ingest**（skill_instruction）：Turn raw, unstructured input into structured Basic Memory entities. Meeting transcripts, conversation logs, pasted documents, email threads — anything with information worth preserving gets parsed, cross-referenced against existing knowledge, and written as proper notes. 证据：`skills/memory-ingest/SKILL.md`
- **Memory Lifecycle**（skill_instruction）：Manage how entities move through status stages in Basic Memory. The core principle: archive, never delete. Completed work is valuable context — move it out of the active view, but keep it in the knowledge graph. 证据：`skills/memory-lifecycle/SKILL.md`
- **Memory Literary Analysis**（skill_instruction）：Transform a complete literary work into a structured knowledge graph. Characters, themes, chapters, locations, symbols, and literary devices become interconnected notes — searchable, validatable, and visualizable. 证据：`skills/memory-literary-analysis/SKILL.md`
- **Memory Metadata Search**（skill_instruction）：Find notes by their structured frontmatter fields instead of or in addition to free-text content. Any custom YAML key in a note's frontmatter beyond the standard set title , type , tags , permalink , schema is automatically indexed as entity metadata and becomes queryable. 证据：`skills/memory-metadata-search/SKILL.md`
- **Memory Notes**（skill_instruction）：Write well-structured notes that Basic Memory can parse into a searchable knowledge graph. Every note is a markdown file with three key sections: frontmatter, observations, and relations. 证据：`skills/memory-notes/SKILL.md`
- **Memory Reflect**（skill_instruction）：Review recent activity and consolidate valuable insights into long-term memory. 证据：`skills/memory-reflect/SKILL.md`
- **Memory Research**（skill_instruction）：Research an external subject, synthesize what you find, and create a structured Basic Memory entity — with the user's approval. 证据：`skills/memory-research/SKILL.md`
- **Memory Schema**（skill_instruction）：Manage structured note types using Basic Memory's Picoschema system. Schemas define what fields a note type should have, making notes uniform, queryable, and validatable. 证据：`skills/memory-schema/SKILL.md`
- **Memory Tasks**（skill_instruction）：Manage work-in-progress using Basic Memory's schema system. Tasks are just notes with type: Task — they live in the knowledge graph, validate against a schema, and survive context compaction. 证据：`skills/memory-tasks/SKILL.md`
- **Marketplace**（structured_config）：{ "name": "basicmachines-co", "owner": { "name": "Basic Machines", "email": "hello@basicmachines.co" }, "metadata": { "description": "Official Basic Memory plugins from the canonical basic-memory repository", "version": "0.22.1" }, "plugins": { "name": "basic-memory", "source": "./plugins/claude-code", "description": "The bridge between Claude's working memory and Basic Memory's durable knowledge graph \u2014 session briefings, pre-compaction checkpoints, and capture reflexes", "version": "0.22.1", "author": { "name": "Basic Machines" }, "keywords": "memory", "knowledge", "mcp", "specs", "context" } } 证据：`.claude-plugin/marketplace.json`
- **Marketplace**（structured_config）：{ "name": "basic-memory-local", "interface": { "displayName": "Basic Memory Local" }, "plugins": { "name": "codex", "source": { "source": "local", "path": "./plugins/codex" }, "policy": { "installation": "AVAILABLE", "authentication": "ON INSTALL" }, "category": "Developer Tools" } } 证据：`.agents/plugins/marketplace.json`
- **Marketplace**（structured_config）：{ "name": "basicmachines-co", "owner": { "name": "Basic Machines", "email": "hello@basicmachines.co" }, "metadata": { "description": "Official plugins from Basic Machines for knowledge management and AI-assisted development", "version": "0.22.1" }, "plugins": { "name": "basic-memory", "source": "./", "description": "The bridge between Claude's working memory and Basic Memory's durable knowledge graph \u2014 session briefings, pre-compaction checkpoints, and capture reflexes", "version": "0.22.1", "author": { "name": "Basic Machines" }, "keywords": "memory", "knowledge", "mcp", "specs", "context" } } 证据：`plugins/claude-code/.claude-plugin/marketplace.json`
- **Plugin**（structured_config）：{ "name": "basic-memory", "description": "The bridge between Claude's working memory and Basic Memory's durable knowledge graph \u2014 session briefings, pre-compaction checkpoints, and capture reflexes", "version": "0.22.1", "author": { "name": "Basic Machines" }, "repository": "https://github.com/basicmachines-co/basic-memory/tree/main/plugins/claude-code" } 证据：`plugins/claude-code/.claude-plugin/plugin.json`
- **Plugin**（structured_config）：{ "name": "codex", "version": "0.22.1", "description": "A Codex-native bridge to Basic Memory for durable engineering context, decisions, and resumable checkpoints.", "author": { "name": "Basic Machines", "email": "hello@basicmachines.co", "url": "https://basicmemory.com" }, "homepage": "https://docs.basicmemory.com", "repository": "https://github.com/basicmachines-co/basic-memory/tree/main/plugins/codex", "license": "MIT", "keywords": "basic-memory", "codex", "memory", "knowledge-graph", "mcp", "checkpoints" , "skills": "./skills/", "mcpServers": "./.mcp.json", "interface": { "displayName": "Basic Memory for Codex", "shortDescription": "Carry decisions, active work, and handoffs across Cod… 证据：`plugins/codex/.codex-plugin/plugin.json`
- **License**（source_file）：GNU AFFERO GENERAL PUBLIC LICENSE Version 3, 19 November 2007 证据：`LICENSE`
- **License**（source_file）：GNU AFFERO GENERAL PUBLIC LICENSE Version 3, 19 November 2007 证据：`integrations/hermes/LICENSE`
- **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：`integrations/openclaw/LICENSE`
- **Docker Setup Guide**（documentation）：Basic Memory can be run in Docker containers to provide a consistent, isolated environment for your knowledge management system. This is particularly useful for integrating with existing Dockerized MCP servers or for deployment scenarios. 证据：`docs/Docker.md`
- 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：`README.md`, `benchmarks/README.md`, `skills/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：`README.md`, `benchmarks/README.md`, `skills/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。

- **系统概述与核心架构**：importance `high`
  - source_paths: README.md, docs/ARCHITECTURE.md, src/basic_memory/api/app.py, src/basic_memory/mcp/server.py, src/basic_memory/deps/__init__.py
- **核心功能与 MCP 工具**：importance `high`
  - source_paths: src/basic_memory/mcp/tools/__init__.py, src/basic_memory/mcp/tools/write_note.py, src/basic_memory/mcp/tools/read_note.py, src/basic_memory/mcp/tools/edit_note.py, src/basic_memory/mcp/tools/search.py
- **数据管理与同步引擎**：importance `high`
  - source_paths: src/basic_memory/db.py, src/basic_memory/models/knowledge.py, src/basic_memory/models/search.py, src/basic_memory/alembic/versions/3dae7c7b1564_initial_schema.py, src/basic_memory/sync/sync_service.py
- **扩展、部署与运维**：importance `high`
  - source_paths: plugins/claude-code/README.md, plugins/codex/README.md, integrations/openclaw/README.md, integrations/hermes/README.md, skills/README.md

## Repo Inspection Evidence / 源码检查证据

- repo_clone_verified: true
- repo_inspection_verified: true
- repo_commit: `3130e5a9fae752c62a01dce74b74d8ae6aa6a670`
- inspected_files: `Dockerfile`, `README.md`, `docker-compose.yml`, `pyproject.toml`, `uv.lock`, `docs/ARCHITECTURE.md`, `docs/Docker.md`, `docs/ENGINEERING_STYLE.md`, `docs/NOTE-FORMAT.md`, `docs/SPEC-PER-PROJECT-ROUTING.md`, `docs/ai-assistant-guide-extended.md`, `docs/auto-bm-webhook-test.md`, `docs/character-handling.md`, `docs/cloud-cli.md`, `docs/cloud-semantic-search-value.md`, `docs/litellm-provider.md`, `docs/logfire-instrumentation-strategy.md`, `docs/manual-pages.md`, `docs/mcp-ui-bakeoff-instructions.md`, `docs/metadata-search.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: 来源证据：Query-time relation traversal in search (one-hop graph expansion)

- 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：Query-time relation traversal in search (one-hop graph expansion)
- 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/basicmachines-co/basic-memory/issues/1001 | 来源类型 github_issue 暴露的待验证使用条件。
- 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: 来源证据：[BUG] File-watcher (direct on-disk) writes are not vector-embedded; externally-edited notes silently missing from seman…

- Trigger: GitHub 社区证据显示该项目存在一个配置相关的待验证问题：[BUG] File-watcher (direct on-disk) writes are not vector-embedded; externally-edited notes silently missing from semantic search until reindex
- 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/basicmachines-co/basic-memory/issues/1016 | 来源类型 github_issue 暴露的待验证使用条件。
- 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: 来源证据：[BUG] Large cloud-project deletion errors opaquely; no documented way to purge a removed cloud project's data (hosted-t…

- Trigger: GitHub 社区证据显示该项目存在一个安全/权限相关的待验证问题：[BUG] Large cloud-project deletion errors opaquely; no documented way to purge a removed cloud project's data (hosted-tenant retention gap)
- Why it matters: 可能影响授权、密钥配置或安全边界。
- Evidence: community_evidence:github | https://github.com/basicmachines-co/basic-memory/issues/1034 | 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 4: 来源证据：[BUG] PermissionError during project scan triggers mass index deletion

- Trigger: GitHub 社区证据显示该项目存在一个安全/权限相关的待验证问题：[BUG] PermissionError during project scan triggers mass index deletion
- Why it matters: 可能影响授权、密钥配置或安全边界。
- Evidence: community_evidence:github | https://github.com/basicmachines-co/basic-memory/issues/1007 | 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：[FEATURE] Please allow additional knowledge roots for safe symlink traversal

- Trigger: GitHub 社区证据显示该项目存在一个安全/权限相关的待验证问题：[FEATURE] Please allow additional knowledge roots for safe symlink traversal
- 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/basicmachines-co/basic-memory/issues/871 | 来源讨论提到 windows 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：[BUG] Empty ~/basic-memory directory recreated on every config load (mkdir side effect in config validator)

- 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：[BUG] Empty ~/basic-memory directory recreated on every config load (mkdir side effect in config validator)
- 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/basicmachines-co/basic-memory/issues/1029 | 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: 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/basicmachines-co/basic-memory | host_targets=mcp_host, claude, claude_code, cursor, chatgpt
- 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: 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/basicmachines-co/basic-memory | 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 9: 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/basicmachines-co/basic-memory | 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/basicmachines-co/basic-memory | 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。