# palinode - 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 palinode 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：`skill/palinode-claude-code/SKILL.md`, `skill/palinode-memory/SKILL.md`, `skill/palinode-session/SKILL.md` 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：`skill/palinode-claude-code/SKILL.md`, `skill/palinode-memory/SKILL.md`, `skill/palinode-session/SKILL.md` 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：`claude-plugin/.claude-plugin/plugin.json`, `plugin/openclaw.plugin.json` 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`, `docs/QUICKSTART.md` Claim：`clm_0003` supported 0.86

## 怎么开始

- `git clone https://github.com/phasespace-labs/palinode && cd palinode` 证据：`README.md` Claim：`clm_0006` supported 0.86
- `pip install -e .` 证据：`README.md` Claim：`clm_0007` supported 0.86
- `curl http://localhost:6340/status` 证据：`README.md` Claim：`clm_0008` supported 0.86
- `curl -X POST http://localhost:6340/save \` 证据：`docs/QUICKSTART.md` Claim：`clm_0009` supported 0.86
- `curl -X POST http://localhost:6340/search \` 证据：`docs/QUICKSTART.md` Claim：`clm_0010` supported 0.86
- `curl -X POST http://localhost:6340/reindex` 证据：`docs/QUICKSTART.md` Claim：`clm_0011` 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、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_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：`skill/palinode-claude-code/SKILL.md`, `skill/palinode-memory/SKILL.md`, `skill/palinode-session/SKILL.md` 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：`skill/palinode-claude-code/SKILL.md`, `skill/palinode-memory/SKILL.md`, `skill/palinode-session/SKILL.md` 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：`claude-plugin/.claude-plugin/plugin.json`, `plugin/openclaw.plugin.json` 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`, `docs/QUICKSTART.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

- **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-plugin/.claude-plugin/plugin.json`, `plugin/openclaw.plugin.json`, `skill/palinode-claude-code/SKILL.md`, `skill/palinode-memory/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：`claude-plugin/.claude-plugin/plugin.json`, `plugin/openclaw.plugin.json`
- **项目输出质量是否满足用户具体任务？**（unverified）：安装前预览只能展示流程和边界，不能替代真实评测。
- **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`, `docs/QUICKSTART.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-plugin/.claude-plugin/plugin.json`, `plugin/openclaw.plugin.json`, `skill/palinode-claude-code/SKILL.md`, `skill/palinode-memory/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：`README.md`, `claude-plugin/.claude-plugin/plugin.json`, `docs/QUICKSTART.md`, `plugin/openclaw.plugin.json`
- **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：`README.md`, `SECURITY.md`, `docs/CHANGELOG.md`, `docs/INSTALL-CLAUDE-CODE.md` 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**：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。）
- **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。
- **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。
- **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_0012` 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：`claude-plugin/.claude-plugin/plugin.json`, `plugin/openclaw.plugin.json` Claim：`clm_0013` 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`, `docs/QUICKSTART.md` Claim：`clm_0014` 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：`skill/palinode-claude-code/SKILL.md`, `skill/palinode-memory/SKILL.md`, `skill/palinode-session/SKILL.md` 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：`claude-plugin/.claude-plugin/plugin.json`, `plugin/openclaw.plugin.json` 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`, `docs/QUICKSTART.md` Claim：`clm_0003` supported 0.86

### 上下文规模

- 文件总数：217
- 重要文件覆盖：40/217
- 证据索引条目：80
- 角色 / Skill 条目：3

### 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 palinode 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 palinode 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 palinode 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 3 个角色 / Skill / 项目文档条目。

- **palinode-claude-code**（skill）：Use Palinode persistent memory in Claude Code via MCP 13 tools for search, save, git diff, triggers, consolidation . Use when user says 'save this decision', 'remember why we chose X', 'search memory before we start', 'what do we know about this project', 'show what changed in memory', 'register a trigger for this topic', 'set up Palinode MCP'. Also fires when an agent needs to: load prior context at session start,… 激活提示：当用户任务与“palinode-claude-code”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：`skill/palinode-claude-code/SKILL.md`
- **palinode-memory**（skill）：Search, save, and manage persistent agent memory using Palinode git-versioned markdown + vector search . Use when user says 'remember this', 'save to memory', 'what do you know about X', 'search memory for context', 'what was decided about', 'register a trigger', 'show memory diff', 'run consolidation'. Also fires when an orchestrating agent needs to: retrieve prior context before a task, save a decision for future… 激活提示：当用户任务与“palinode-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：`skill/palinode-memory/SKILL.md`
- **palinode-session**（skill）：Automatically manage persistent memory during coding sessions via Palinode MCP. Fires when: starting a new task, completing a milestone, making a decision, finishing a session, or when 30+ minutes have passed since last save. Also fires on 'save to memory', 'remember this', 'what do we know about'. Do NOT fire on trivial file edits or routine commands. 激活提示：当用户任务与“palinode-session”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：`skill/palinode-session/SKILL.md`

## 证据索引

- 共索引 80 条证据。

- **Supported Platforms**（documentation）：Audit-grade memory for AI coding agents. Git-versioned and file-native — every fact your agent recalls, you can blame , diff , and rollback . MCP-first, so one memory works in every editor. 证据：`README.md`
- **Palinode Claude Code Plugin**（documentation）：Persistent memory for AI agents. Your agent's memory is a folder of plain markdown files, indexed for semantic search and compacted by an LLM. Every change is a git commit. If every service crashes, cat still works. 证据：`claude-plugin/README.md`
- **Example Memory Files**（documentation）：Copy these to your Palinode memory directory to get started: 证据：`examples/README.md`
- **Palinode — systemd user units**（documentation）：Version-controlled, templated systemd user-unit files for the three palinode production services. 证据：`deploy/systemd/README.md`
- **Compaction Demo — Watching a Memory File Consolidate**（documentation）：Compaction Demo — Watching a Memory File Consolidate 证据：`examples/compaction-demo/README.md`
- **Palinode Claude Code hooks**（documentation）：Drop-in hooks that auto-capture Claude Code sessions to Palinode. 证据：`examples/hooks/README.md`
- **Sample Memory**（documentation）：A pre-populated memory directory for trying Palinode. Copy it to your memory directory: 证据：`examples/sample-memory/README.md`
- **CLAUDE.md — Palinode Memory Integration**（documentation）：CLAUDE.md — Palinode Memory Integration 证据：`examples/CLAUDE.md`
- **Package**（package_manifest）：{ "name": "openclaw-palinode", "version": "0.1.0", "type": "module", "description": "Palinode memory system for OpenClaw", "main": "./dist/index.js", "files": "dist", "openclaw.plugin.json", "types" , "dependencies": { "@sinclair/typebox": "0.34.47" }, "openclaw": { "extensions": "./dist/index.js" }, "scripts": { "build": "tsc", "prepublishOnly": "tsc", "pretest": "python3 ../scripts/dump-parity-registry.py parity-registry.json", "test": "vitest run" }, "devDependencies": { "@types/node": "^22.0.0", "typescript": "^5.4.0", "vitest": "^2.1.9" } } 证据：`plugin/package.json`
- **openclaw-palinode — Installation & Host-Side Opt-in Flags**（documentation）：openclaw-palinode — Installation & Host-Side Opt-in Flags 证据：`plugin/INSTALL.md`
- **Palinode — Claude Code Integration**（skill_instruction）：Palinode gives Claude Code persistent, git-versioned memory via 13 MCP tools. Memories survive across sessions, are searchable by meaning, and consolidate weekly. 证据：`skill/palinode-claude-code/SKILL.md`
- **Palinode Memory**（skill_instruction）：Palinode stores typed memory as git-versioned markdown files with hybrid search BM25 + vector . This skill provides procedural guidance for using Palinode's tools effectively. 证据：`skill/palinode-memory/SKILL.md`
- **Palinode Session Memory**（skill_instruction）：This skill keeps your AI agent's memory fresh across coding sessions using Palinode MCP tools. 证据：`skill/palinode-session/SKILL.md`
- **Plugin**（structured_config）：{ "name": "palinode", "description": "Persistent memory for AI agents. Git-versioned markdown files as source of truth, hybrid SQLite-vec + FTS5 search, deterministic compaction. Every line in your agent's brain has a git blame.", "version": "0.6.1", "author": { "name": "Paul Kyle", "url": "https://github.com/Paul-Kyle" }, "homepage": "https://github.com/phasespace-labs/palinode", "repository": "https://github.com/phasespace-labs/palinode", "license": "MIT", "keywords": "memory", "persistence", "markdown", "git", "agent", "mcp", "sqlite-vec", "bge-m3", "compaction", "local-first" } 证据：`claude-plugin/.claude-plugin/plugin.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`
- **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：`claude-plugin/LICENSE`
- **Acknowledgments**（documentation）：Palinode builds on ideas from across the agent memory landscape: 证据：`docs/ACKNOWLEDGMENTS.md`
- **Changelog**（documentation）：All notable changes to Palinode. Format follows Keep a Changelog https://keepachangelog.com/en/1.1.0/ . 证据：`docs/CHANGELOG.md`
- **Palinode Deployment Guide**（documentation）：Lessons from running Palinode in production across multiple machines with Claude Code, Claude Desktop, and Cursor. 证据：`docs/DEPLOYMENT-GUIDE.md`
- **Palinode — Developer Setup**（documentation）：This document covers the local development workflow, including running the test suite and the specific setup required when working inside a git worktree . 证据：`docs/DEVELOPMENT.md`
- **palinode doctor — diagnostic guide**（documentation）：palinode doctor is a read-only diagnostic command that inspects a running Palinode installation and reports silent-misconfiguration bugs that would otherwise look like data loss. It is the tool you reach for when search returns less than you expected, when "save" feels like it didn't stick, or when something used to work and now does not. It is deliberately scoped: doctor diagnoses , it almost never repairs . The cases where it does repair are explicitly whitelisted and described below — they never touch user data. 证据：`docs/DOCTOR.md`
- **Consolidation Executor Determinism Specification**（documentation）：Consolidation Executor Determinism Specification 证据：`docs/EXECUTOR-SPEC.md`
- **Git-Powered Memory in Palinode**（documentation）：Palinode treats memory as code. Every fact, decision, and project detail your agent learns is saved as a markdown file and versioned in a dedicated Git repository. This provides complete provenance: you can trace any fact back to the exact session where it was recorded, and see how your agent's understanding evolved over time. 证据：`docs/GIT-MEMORY.md`
- **Harness Smoke Checklist**（documentation）：Per-harness smoke checklist for validating Palinode MCP connectivity. Each harness runs the same 5-call sequence for cross-comparison. Tracked at issue 345 parent 342 . 证据：`docs/HARNESS-SMOKE.md`
- **How Palinode Memory Works**（documentation）：A complete guide to how context flows through the system — from capture to recall to consolidation. 证据：`docs/HOW-MEMORY-WORKS.md`
- **Installing Palinode with Claude Code**（documentation）：Installing Palinode with Claude Code 证据：`docs/INSTALL-CLAUDE-CODE.md`
- **L4 behavioural testing — design sketch**（documentation）：L4 behavioural testing — design sketch 证据：`docs/L4-BEHAVIORAL-TESTING-DESIGN.md`
- **Palinode Pre-Launch Checklist**（documentation）：Checklist for v1.0 public launch readiness. 证据：`docs/LAUNCH-CHECKLIST.md`
- **MCP Config Homes — Canonical Locations Per Client**（documentation）：MCP Config Homes — Canonical Locations Per Client 证据：`docs/MCP-CONFIG-HOMES.md`
- **MCP Install Recipes — Per-Client Setup Workflows**（documentation）：MCP Install Recipes — Per-Client Setup Workflows 证据：`docs/MCP-INSTALL-RECIPES.md`
- **Palinode MCP Setup**（documentation）：Connect Palinode's MCP tools to your AI coding assistant. Two transport options: 证据：`docs/MCP-SETUP.md`
- **Using Palinode with Obsidian**（documentation）：The comprehensive guide moved to OBSIDIAN.md OBSIDIAN.md — quickstart, the wiki contract, the embedding tools, migration paths, and FAQ. 证据：`docs/OBSIDIAN-SETUP.md`
- **Palinode + Obsidian**（documentation）：Point Obsidian at your Palinode directory. You get a persistent, searchable memory that any MCP-compatible LLM can read and write — plus everything Obsidian gives you natively: graph view, backlinks, Bases, plugins. No sync job, no export step, no plugin to install. The same markdown files are the source of truth for both. 证据：`docs/OBSIDIAN.md`
- **Palinode Operations Guide**（documentation）：How to upgrade, recover from crashes, and maintain a healthy Palinode installation. 证据：`docs/OPERATIONS.md`
- **PARITY.md — Cross-Surface Contract**（documentation）：Status: Active ADR-010 accepted 2026-04-26 Source of truth: palinode/core/parity.py — the registry Forcing function: tests/test surface parity.py — CI-blocking Defaults: palinode/core/defaults.py — single-place values for thresholds, cooldowns, source headers 证据：`docs/PARITY.md`
- **Palinode Quickstart**（documentation）：- Python 3.11+ - Ollama with BGE-M3 on a reachable host default: localhost:11434 - Git 证据：`docs/QUICKSTART.md`
- **palinode v0.8.0 — Obsidian, Doctor, Reliable Save**（documentation）：palinode v0.8.0 — Obsidian, Doctor, Reliable Save 证据：`docs/RELEASE-NOTES-v0.8.0.md`
- **Validation strategy for cross-session features**（documentation）：Validation strategy for cross-session features 证据：`docs/VALIDATION-STRATEGY.md`
- **Why Local Memory Matters**（documentation）：AI coding agents are getting powerful automation features — scheduled tasks, looping prompts, MCP integrations, custom skills, computer control. But they all share one limitation: sessions are stateless by default. 证据：`docs/WHY-LOCAL-MEMORY.md`
- **Logging conventions**（documentation）：Palinode runs unattended on remote hosts. When something goes wrong, the only forensic surface is the log — usually journalctl on the host running the API, watcher, or MCP server. The recurring failure mode this convention exists to prevent is "the system did the wrong thing silently" : a failure that was caught, handled, and degraded gracefully, but never surfaced at a level or with the structure an operator would notice. 证据：`docs/logging.md`
- **Runbook: Ollama degraded or unreachable**（documentation）：Runbook: Ollama degraded or unreachable 证据：`docs/ollama-degraded.md`
- **Sentinel returned by generate description when the Ollama call timed out.**（source_file）：logger = logging.getLogger "palinode.api" def wrap user content for llm content: str - str ⋮---- safe = ⋮---- Sentinel returned by generate description when the Ollama call timed out. Distinguishable from "" total failure fallback and a real description. The save path writes description pending=True to the API response when it sees this; the watcher retries files where description is still absent. DESCRIPTION DEFERRED = object identity sentinel — never a string 464: per-/generate-summaries-run budget for CHAT fallback escalations. generate summaries api resets remaining to config.auto summary.llm fallback max per run at the top of each run; the fallback helper decrements it once per file th… 证据：`palinode/api/enrichment.py`
- **Pattern that matches an existing auto-footer block up to end-of-string or**（source_file）：logger = logging.getLogger "palinode.api" CATEGORY TO ENTITY PREFIX: dict str, str = { WIKI FOOTER MARKER = " " SAFE SLUG RE = re.compile r"^ A-Za-z0-9. - +$" def safe wiki slug slug: str - bool ⋮---- """Return True if slug is safe to embed inside ... markdown. Used by apply wiki footer to drop hostile entity slugs that would inject markdown structure bar , embedded pipes, newlines, etc. . """ ⋮---- def apply wiki footer content: str, entities: list str - str ⋮---- """Append or update a See also auto-footer for un-linked entities. When entities are provided but some of them are not already referenced as wikilinks in content , this function appends a detectable auto-generated footer so that… 证据：`palinode/api/memory_write.py`
- **Sort newest first so listing surfaces recent activity.**（source_file）：logger = logging.getLogger "palinode.api" router = APIRouter def normalize sources raw: list dict str, Any - list dict str, str ⋮---- normalized: list dict str, str = ⋮---- ref = entry.get "ref" quote = entry.get "quote" ⋮---- ref = ref.strip computed = quote hash quote supplied = entry.get "quote hash" ⋮---- @router.get "/read" def read api file path: str, meta: bool = False - dict str, Any ⋮---- candidates = file path ⋮---- resolved = "" content = "" ⋮---- content = open memory file text resolved candidate ⋮---- file path = candidate resolved = resolved candidate ⋮---- result = { ⋮---- class SaveRequest BaseModel ⋮---- content: str type: str slug: str None = None entities: list str None =… 证据：`palinode/api/routers/memory.py`
- **OpenAI / generic sk-...**（source_file）：logger = logging.getLogger "palinode.api" ⋮---- SECRET PATTERNS: tuple tuple re.Pattern str , str , ... = ⋮---- OpenAI / generic sk-... ⋮---- Slack bot/user/app tokens ⋮---- GitHub personal-access tokens ⋮---- AWS access key id ⋮---- Generic 32+ char hex/base64 tokens preceded by a header-name keyword. Covers api key=... , token: ... , Authorization: Bearer ... , Bearer ... common bare prefix in HTTP traces , etc. ⋮---- def redact secrets text: str - str ⋮---- """Apply SECRET PATTERNS to text . Returns text unchanged on no match.""" ⋮---- text = pattern.sub replacement, text ⋮---- class SecretRedactingFilter logging.Filter ⋮---- """Strip credentials from log messages and traceback text befo… 证据：`palinode/api/server.py`
- **Render**（source_file）：ALLOWED TAGS: set str = { ALLOWED ATTRS: dict str, set str = { ALLOWED URL SCHEMES: set str = {"http", "https", "mailto"} md = MarkdownIt "commonmark", {"html": False, "linkify": True, "typographer": False} ⋮---- def render markdown body: str - str ⋮---- raw html = md.render body ⋮---- def sanitize html html: str - str 证据：`palinode/api/ui/render.py`
- **Diffs**（source_file）：{% extends "base.html" %} {% block title %}Palinode · diffs{% endblock %} {% block crumb %} diffs {% endblock %} {% block content %} Recent changes Memory writes over the last {{ diffs.days }} days, from git — newest first. {% if diffs.groups %} {% for g in diffs.groups %} {{ g.day }} {% for c in g.commits %} {{ c.hash }} {% if c.time %} {{ c.time }} {% endif %} {{ c.message }} {% if c.files %} {% for f in c.files %} {{ f.path }} {% endfor %} {% endif %} {% endfor %} {% endfor %} {% else %} No memory changes in the last {{ diffs.days }} days or the store is not a git repo . {% endif %} {% endblock %} 证据：`palinode/api/ui/templates/diffs.html`
- **Fact**（source_file）：{% extends "base.html" %} {% block title %}Palinode · {{ title }}{% endblock %} {% block crumb %} memory / {{ category or "uncategorized" }} / {{ slug }} {% endblock %} {% block health pill %} {% if broken seal %} tampered {% else %} verified {% endif %} {% endblock %} {% block content %} {{ kicker }} {{ title }} id {{ memory id }} {% if confidence is not none %} confidence {{ confidence }} {% endif %} {% if recall count %} recalled {{ recall count }}× {% endif %} {% for chip in extra chips %} {{ chip.label }} {{ chip.value }} {% endfor %} {{ body html safe }} Provenance {% if broken seal %} ✕ seal broken {% else %} ✓ chain intact {% endif %} why the agent believes this — and whether it’s b… 证据：`palinode/api/ui/templates/fact.html`
- **Quality**（source_file）：{% extends "base.html" %} {% block title %}Palinode · quality{% endblock %} {% block crumb %} quality {% endblock %} {% block content %} Quality queues What needs attention — from lint . Each item links to its fact. {% for q in quality.queues %} {{ q.label }} {{ q.rows length }} — {{ q.blurb }} {% if q.rows %} {% if q.kind == 'contradictions' %} {% for r in q.rows %} {{ r.entity }} — {{ r.issue }} {% endfor %} {% else %} {% for r in q.rows %} {{ r.path }} {% if r.detail %} · {{ r.detail }} {% endif %} {% endfor %} {% endif %} {% else %} Nothing here — clean. {% endif %} {% endfor %} {% endblock %} 证据：`palinode/api/ui/templates/quality.html`
- **Embedding Tools**（source_file）：def dedup suggest content, file path, min similarity, top k, fmt ⋮---- content = f.read ⋮---- results = api client.dedup suggest ⋮---- output fmt = OutputFormat fmt if fmt else get default format ⋮---- fp = r.get "file path", "" pct = int r.get "similarity", 0 100 snippet = r.get "snippet" or "" .strip .replace "\n", " " :200 ⋮---- def orphan repair broken link, min similarity, top k, fmt ⋮---- results = api client.orphan repair ⋮---- def cluster neighbors file path, min similarity, top k, fmt ⋮---- results = api client.cluster neighbors ⋮---- def topic coverage query, min similarity, fmt ⋮---- result = api client.topic coverage ⋮---- best = result.get "best match", "" pct = int result.get… 证据：`palinode/cli/embedding_tools.py`
- **Print plan**（source_file）：@click.group "import" def import cmd - None ⋮---- memory dir str = os.environ.get "PALINODE DIR", config.memory dir memory dir = Path os.path.expanduser memory dir str ⋮---- source path = Path source vault into cat = into category.rstrip "/" if into category else None ⋮---- Print plan ⋮---- dest rel = plan.dest path.relative to memory dir if plan.dest path.is relative to memory dir else plan.dest path src rel = plan.source path.relative to source path if plan.source path.is relative to source path else plan.source path status = "" ⋮---- status = " EXISTS " + "" if overwrite else " — would skip" ⋮---- result = execute import plans, overwrite=overwrite ⋮---- rel = p.relative to memory dir if… 证据：`palinode/cli/import_vault.py`
- **--force-obsidian implies --obsidian**（source_file）：CLAUDE MD BLOCK = """\ WRAP POLICY HEAVY NOTE = """ HOOK SCRIPT = """\ SAVE COMMAND BODY = """\ PS COMMAND BODY = """\ WRAP COMMAND BODY = """\ WRAP HEAVY COMMAND BODY = """\ SETTINGS HOOK BLOCK = { MCP JSON BLOCK = { def slugify name: str - str ⋮---- s = re.sub r" ^a-zA-Z0-9 - +", "-", name.strip .lower s = re.sub r"-+", "-", s .strip "-" ⋮---- def ensure parent path: Path - None ⋮---- wrap policy note = WRAP POLICY HEAVY NOTE if wrap policy == "heavy" else "" block = CLAUDE MD BLOCK.format ⋮---- existing = path.read text ⋮---- def write hook script path: Path, force: bool - str ⋮---- mode = path.stat .st mode ⋮---- def merge settings path: Path, force: bool - str ⋮---- existing = json.loa… 证据：`palinode/cli/init.py`
- **Parse frontmatter to extract entities.**（source_file）：SKIP DIRS: frozenset str = frozenset { ⋮---- raw = glob.glob os.path.join memory dir, include , recursive=True ⋮---- raw = glob.glob os.path.join memory dir, " / .md" , recursive=True results: list str = ⋮---- rel = os.path.relpath filepath, memory dir parts = rel.split os.sep ⋮---- def obsidian sync apply: bool, include glob: Optional str , exclude glob: Optional str - None ⋮---- memory dir = os.path.expanduser ⋮---- files = walk memory files memory dir, include glob, exclude glob ⋮---- mode = "apply" if apply else "dry-run" ⋮---- would update: list str = unchanged: list str = errors: list tuple str, str = ⋮---- original = fh.read ⋮---- Parse frontmatter to extract entities. ⋮---- except E… 证据：`palinode/cli/obsidian_sync.py`
- **Privacy Policy**（documentation）：Palinode — Persistent memory for AI agents Effective date: April 12, 2026 Last updated: April 12, 2026 证据：`PRIVACY.md`
- **PROGRAM.md — Palinode Memory Manager**（documentation）：PROGRAM.md — Palinode Memory Manager 证据：`PROGRAM.md`
- **Security Policy**（documentation）：If you discover a security vulnerability in Palinode, please report it responsibly. 证据：`SECURITY.md`
- **My App — Status Pass 0: raw, 3 weeks of accretion**（documentation）：My App — Status Pass 0: raw, 3 weeks of accretion 证据：`examples/compaction-demo/pass-0-initial.md`
- **My App — Status Pass 1: after UPDATE + SUPERSEDE**（documentation）：My App — Status Pass 1: after UPDATE + SUPERSEDE 证据：`examples/compaction-demo/pass-1-after-update-supersede.md`
- **My App — Status Pass 2: after MERGE + ARCHIVE**（documentation）：My App — Status Pass 2: after MERGE + ARCHIVE 证据：`examples/compaction-demo/pass-2-after-merge-archive.md`
- **Decision: REST API for Checkout**（documentation）：What Was Decided Use REST endpoints for the checkout API instead of GraphQL. 证据：`examples/decisions/api-design.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`, `claude-plugin/README.md`, `examples/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`, `claude-plugin/README.md`, `examples/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。

- **概述与系统架构**：importance `high`
  - source_paths: README.md, PROGRAM.md, pyproject.toml, docs/WHY-LOCAL-MEMORY.md, docs/HOW-MEMORY-WORKS.md
- **记忆引擎：存储、索引、检索与压缩**：importance `high`
  - source_paths: palinode/core/store.py, palinode/core/embedder.py, palinode/core/ranker.py, palinode/core/parser.py, palinode/consolidation/executor.py
- **接口与多平台集成**：importance `high`
  - source_paths: palinode/mcp.py, palinode/api/server.py, palinode/api/routers/memory.py, palinode/api/routers/search.py, palinode/api/routers/consolidation.py
- **部署、运维与诊断**：importance `high`
  - source_paths: palinode/cli/doctor.py, palinode/diagnostics/runner.py, palinode/diagnostics/registry.py, palinode/diagnostics/checks/__init__.py, palinode/diagnostics/fixes.py

## Repo Inspection Evidence / 源码检查证据

- repo_clone_verified: true
- repo_inspection_verified: true
- repo_commit: `fade8823fede5b4b40443340da667fea8207227c`
- inspected_files: `README.md`, `pyproject.toml`, `docs/ACKNOWLEDGMENTS.md`, `docs/CHANGELOG.md`, `docs/DEPLOYMENT-GUIDE.md`, `docs/DEVELOPMENT.md`, `docs/DOCTOR.md`, `docs/EXECUTOR-SPEC.md`, `docs/GIT-MEMORY.md`, `docs/HARNESS-SMOKE.md`, `docs/HOW-MEMORY-WORKS.md`, `docs/INSTALL-CLAUDE-CODE.md`, `docs/L4-BEHAVIORAL-TESTING-DESIGN.md`, `docs/LAUNCH-CHECKLIST.md`, `docs/MCP-CONFIG-HOMES.md`, `docs/MCP-INSTALL-RECIPES.md`, `docs/MCP-SETUP.md`, `docs/OBSIDIAN-SETUP.md`, `docs/OBSIDIAN.md`, `docs/OPERATIONS.md`

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

## Doramagic Pitfall Constraints / 踩坑约束

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

### Constraint 1: 失败模式：installation: Add llms-install.md for agent-bootstrap (Ollama detection + venv-path discipline)

- Trigger: Developers should check this installation risk before relying on the project: Add llms-install.md for agent-bootstrap (Ollama detection + venv-path discipline)
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: Add llms-install.md for agent-bootstrap (Ollama detection + venv-path discipline). Context: Observed when using node, python
- Why it matters: Developers may fail before the first successful local run: Add llms-install.md for agent-bootstrap (Ollama detection + venv-path discipline)
- Evidence: failure_mode_cluster:github_issue | https://github.com/phasespace-labs/palinode/issues/61 | Add llms-install.md for agent-bootstrap (Ollama detection + venv-path discipline)
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 2: 失败模式：installation: Anthropic Connectors Directory submission (after MCPB bundle ships)

- Trigger: Developers should check this installation risk before relying on the project: Anthropic Connectors Directory submission (after MCPB bundle ships)
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: Anthropic Connectors Directory submission (after MCPB bundle ships). Context: Observed when using node
- Why it matters: Developers may fail before the first successful local run: Anthropic Connectors Directory submission (after MCPB bundle ships)
- Evidence: failure_mode_cluster:github_issue | https://github.com/phasespace-labs/palinode/issues/62 | Anthropic Connectors Directory submission (after MCPB bundle ships)
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 3: 失败模式：installation: Build in public — show Palinode building Palinode

- Trigger: Developers should check this installation risk before relying on the project: Build in public — show Palinode building Palinode
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: Build in public — show Palinode building Palinode. Context: Observed when using node
- Why it matters: Developers may fail before the first successful local run: Build in public — show Palinode building Palinode
- Evidence: failure_mode_cluster:github_issue | https://github.com/phasespace-labs/palinode/issues/33 | Build in public — show Palinode building Palinode
- 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: 失败模式：installation: CLAUDE.md integration — zero-friction adoption path

- Trigger: Developers should check this installation risk before relying on the project: CLAUDE.md integration — zero-friction adoption path
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: CLAUDE.md integration — zero-friction adoption path. Context: Observed when using node, python
- Why it matters: Developers may fail before the first successful local run: CLAUDE.md integration — zero-friction adoption path
- Evidence: failure_mode_cluster:github_issue | https://github.com/phasespace-labs/palinode/issues/32 | CLAUDE.md integration — zero-friction adoption path
- 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: 失败模式：installation: Directory submissions — MCP Registry, Smithery, mcp.so, Cline

- Trigger: Developers should check this installation risk before relying on the project: Directory submissions — MCP Registry, Smithery, mcp.so, Cline
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: Directory submissions — MCP Registry, Smithery, mcp.so, Cline. Context: Observed when using node
- Why it matters: Developers may fail before the first successful local run: Directory submissions — MCP Registry, Smithery, mcp.so, Cline
- Evidence: failure_mode_cluster:github_issue | https://github.com/phasespace-labs/palinode/issues/35 | Directory submissions — MCP Registry, Smithery, mcp.so, Cline
- 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: 失败模式：installation: Docs: public integration guide for OpenClaw and Hermes-agent

- Trigger: Developers should check this installation risk before relying on the project: Docs: public integration guide for OpenClaw and Hermes-agent
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: Docs: public integration guide for OpenClaw and Hermes-agent. Context: Observed during installation or first-run setup.
- Why it matters: Developers may fail before the first successful local run: Docs: public integration guide for OpenClaw and Hermes-agent
- Evidence: failure_mode_cluster:github_issue | https://github.com/phasespace-labs/palinode/issues/60 | Docs: public integration guide for OpenClaw and Hermes-agent
- 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: 失败模式：installation: Homebrew tap (brew install palinode)

- Trigger: Developers should check this installation risk before relying on the project: Homebrew tap (brew install palinode)
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: Homebrew tap (brew install palinode). Context: Observed during installation or first-run setup.
- Why it matters: Developers may fail before the first successful local run: Homebrew tap (brew install palinode)
- Evidence: failure_mode_cluster:github_issue | https://github.com/phasespace-labs/palinode/issues/8 | Homebrew tap (brew install palinode)
- 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: 失败模式：installation: Multi-platform MCP and integration docs (Claude Code, Cursor, Zed, VS Code)

- Trigger: Developers should check this installation risk before relying on the project: Multi-platform MCP and integration docs (Claude Code, Cursor, Zed, VS Code)
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: Multi-platform MCP and integration docs (Claude Code, Cursor, Zed, VS Code). Context: Observed when using node
- Why it matters: Developers may fail before the first successful local run: Multi-platform MCP and integration docs (Claude Code, Cursor, Zed, VS Code)
- Evidence: failure_mode_cluster:github_issue | https://github.com/phasespace-labs/palinode/issues/24 | Multi-platform MCP and integration docs (Claude Code, Cursor, Zed, VS Code)
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 9: 失败模式：installation: v0.8.13

- Trigger: Developers should check this installation risk before relying on the project: v0.8.13
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: v0.8.13. Context: Observed when using node
- Why it matters: Upgrade or migration may change expected behavior: v0.8.13
- Evidence: failure_mode_cluster:github_release | https://github.com/phasespace-labs/palinode/releases/tag/v0.8.13 | v0.8.13
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 10: 失败模式：installation: v0.8.15

- Trigger: Developers should check this installation risk before relying on the project: v0.8.15
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: v0.8.15. Context: Observed when using python
- Why it matters: Upgrade or migration may change expected behavior: v0.8.15
- Evidence: failure_mode_cluster:github_release | https://github.com/phasespace-labs/palinode/releases/tag/v0.8.15 | v0.8.15
- 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。