# memanto - 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 memanto 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

- **AI 研究者或研究型 Agent 构建者**：README 明确围绕研究、实验或论文工作流展开。 证据：`README.md` Claim：`clm_0004` supported 0.86
- **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_0005` 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：`examples/claudecode-skills-memanto/lifecycle-hooks/skills/memanto-companion/SKILL.md` Claim：`clm_0006` 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：`examples/claudecode-skills-memanto/lifecycle-hooks/skills/memanto-companion/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：`examples/claudecode-skills-memanto/lifecycle-hooks/.claude-plugin/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` Claim：`clm_0003` supported 0.86

## 怎么开始

- `pip install memanto` 证据：`README.md` Claim：`clm_0007` supported 0.86

## 继续前判断卡

- **当前建议**：仅建议沙盒试装
- **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**：仅建议沙盒试装
- **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**：真实输出质量不能在安装前相信。
- **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：AI 研究者或研究型 Agent 构建者**（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：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_0005` 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：`examples/claudecode-skills-memanto/lifecycle-hooks/skills/memanto-companion/SKILL.md` Claim：`clm_0006` 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：`examples/claudecode-skills-memanto/lifecycle-hooks/skills/memanto-companion/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：`examples/claudecode-skills-memanto/lifecycle-hooks/.claude-plugin/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` Claim：`clm_0003` supported 0.86

### 现在还不能相信

- **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：`examples/claudecode-skills-memanto/lifecycle-hooks/.claude-plugin/plugin.json`, `examples/claudecode-skills-memanto/lifecycle-hooks/skills/memanto-companion/SKILL.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。 证据：`examples/claudecode-skills-memanto/lifecycle-hooks/.claude-plugin/plugin.json`
- **项目输出质量是否满足用户具体任务？**（unverified）：安装前预览只能展示流程和边界，不能替代真实评测。
- **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：`examples/claudecode-skills-memanto/lifecycle-hooks/.claude-plugin/plugin.json`, `examples/claudecode-skills-memanto/lifecycle-hooks/skills/memanto-companion/SKILL.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`, `examples/claudecode-skills-memanto/lifecycle-hooks/.claude-plugin/plugin.json`
- **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_0008` 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：`examples/claudecode-skills-memanto/lifecycle-hooks/.claude-plugin/plugin.json` Claim：`clm_0009` 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_0010` 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：`examples/claudecode-skills-memanto/lifecycle-hooks/skills/memanto-companion/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：`examples/claudecode-skills-memanto/lifecycle-hooks/.claude-plugin/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` Claim：`clm_0003` supported 0.86

### 上下文规模

- 文件总数：272
- 重要文件覆盖：40/272
- 证据索引条目：79
- 角色 / Skill 条目：1

### 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 memanto 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 memanto 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 memanto 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 1 个角色 / Skill / 项目文档条目。

- **memanto-companion**（skill）：Inspect and manage the cross-session engineering memory that Memanto maintains for your Claude Code skills. Use when the user asks what Memanto remembers, wants to see their engineering profile, manually recall context for a skill, or store a decision. The automatic lifecycle hooks handle capture/injection on their own — this skill is the manual control surface. 激活提示：当用户任务与“memanto-companion”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：`examples/claudecode-skills-memanto/lifecycle-hooks/skills/memanto-companion/SKILL.md`

## 证据索引

- 共索引 79 条证据。

- **What Is MEMANTO?**（documentation）：A companion memory agent that lets your agents focus and improve while you keep ownership of everything they learn. 证据：`README.md`
- **Claude Code Skills × Memanto — Cross-Session Engineering Memory**（documentation）：Claude Code Skills × Memanto — Cross-Session Engineering Memory 证据：`examples/claudecode-skills-memanto/lifecycle-hooks/README.md`
- **CrewAI + Memanto Example**（documentation）：This directory contains a real-world example of CrewAI agents using Memanto as their shared, persistent memory layer. Two agents collaborate through a semantic memory database that survives across sessions, agents, and runs. 证据：`examples/crewai-memory/README.md`
- **Memanto + LangGraph Integrations**（documentation）：This directory contains several out-of-the-box examples demonstrating how to integrate Memanto's persistent memory capabilities into LangGraph agents. 证据：`examples/langgraph-memanto/README.md`
- **LangGraph + Memanto: Give Your Graph a Permanent Brain**（documentation）：LangGraph + Memanto: Give Your Graph a Permanent Brain 证据：`examples/langgraph-memanto/memanto_base_store/README.md`
- **Claude Code + Memanto Integration**（documentation）：This package provides native integration of Memanto's https://memanto.ai persistent, cross-session memory capabilities into Claude Code https://docs.anthropic.com/en/docs/agents-and-tools/claude-code/overview and the mattpocock/skills https://github.com/mattpocock/skills ecosystem. 证据：`integrations/claudecode/README.md`
- **CrewAI + Memanto: Persistent Multi-Agent Memory**（documentation）：CrewAI + Memanto: Persistent Multi-Agent Memory 证据：`integrations/crewai/README.md`
- **Hermes + Memanto: Persistent Memory Agent**（documentation）：Hermes + Memanto: Persistent Memory Agent 证据：`integrations/hermes-agents/README.md`
- **LangGraph + Memanto Integration**（documentation）：This package provides native LangGraph tools and a standalone memory layer for integrating Memanto's persistent, cross-session memory capabilities into LangGraph agents. 证据：`integrations/langgraph/README.md`
- **Memanto MCP Server**（documentation）：Persistent semantic memory for any MCP-compatible agent. 证据：`integrations/mcp/README.md`
- **@moorcheh-ai/memanto**（documentation）：TypeScript SDK for Memanto https://github.com/moorcheh-ai/memanto — memory that AI agents love. 证据：`sdks/typescript/README.md`
- **Package**（package_manifest）：{ "name": "@moorcheh-ai/memanto", "version": "0.2.3", "description": "TypeScript SDK for Memanto — memory that AI agents love.", "license": "MIT", "type": "module", "main": "./dist/index.cjs", "module": "./dist/index.js", "types": "./dist/index.d.ts", "exports": { ".": { "types": "./dist/index.d.ts", "import": "./dist/index.js", "require": "./dist/index.cjs" } }, "files": "dist", "README.md" , "scripts": { "openapi:fetch": "node scripts/fetch-openapi.mjs", "openapi:generate": "openapi-ts", "openapi:regenerate": "npm run openapi:fetch && npm run openapi:generate", "build": "tsup", "typecheck": "tsc --noEmit", "test": "vitest run", "test:watch": "vitest" }, "devDependencies": { "@hey-api/open… 证据：`sdks/typescript/package.json`
- **Memanto Memory**（documentation）：Before running ANY skill, always execute: 证据：`integrations/claudecode/CLAUDE.md`
- **Contributing to MEMANTO**（documentation）：Thank you for your interest in contributing to MEMANTO — the universal memory layer for agentic AI. This guide will help you get set up and understand how we work. 证据：`CONTRIBUTING.md`
- **Memanto Companion**（skill_instruction）：Cross-session engineering memory for Claude Code skills runs automatically via lifecycle hooks SessionStart , UserPromptExpansion , Stop . This skill is the manual control surface for when the user wants to inspect or steer it. 证据：`examples/claudecode-skills-memanto/lifecycle-hooks/skills/memanto-companion/SKILL.md`
- **Plugin**（structured_config）：{ "$schema": "https://code.claude.com/schemas/plugin-manifest.json", "name": "memanto-skills", "displayName": "Memanto Engineering Memory", "description": "Cross-session engineering memory for Claude Code skills, powered by Memanto. Recalls past architectural decisions before a skill runs and distills new ones after — automatically, via lifecycle hooks.", "version": "0.1.0", "author": { "name": "Moorcheh AI", "url": "https://memanto.ai" }, "homepage": "https://memanto.ai", "repository": "https://github.com/moorcheh-ai/memanto", "license": "MIT", "keywords": "memanto", "memory", "skills", "hooks", "claude-code", "mattpocock" , "skills": "./skills/memanto-companion" } 证据：`examples/claudecode-skills-memanto/lifecycle-hooks/.claude-plugin/plugin.json`
- **License**（source_file）：Copyright 2026 EdgeAI Innovations Inc. 证据：`LICENSE`
- **License**（source_file）：Copyright 2026 EdgeAI Innovations Inc. 证据：`integrations/hermes-agents/LICENSE`
- **License**（source_file）：Copyright 2026 EdgeAI Innovations Inc. 证据：`integrations/mcp/LICENSE`
- **MEMANTO Agent Integration Guide**（documentation）：For AI Coding Assistants, Chatbots, and Automation Agents 证据：`docs/AGENT_INTEGRATION_GUIDE.md`
- **Agent Memory Best Practices Guide**（documentation）：For : AI agents using MEMANTO for persistent memory Last Updated : March 2026 Architecture : Session-Based CLI + API 证据：`docs/AGENT_MEMORY_BEST_PRACTICES.md`
- **MEMANTO CLI Installation & Usage Guide**（documentation）：MEMANTO CLI Installation & Usage Guide 证据：`docs/CLI_INSTALLATION.md`
- **MEMANTO CLI User Guide**（documentation）：Status : Production Ready Last Updated : December 2025 证据：`docs/CLI_USER_GUIDE.md`
- **MEMANTO Production Deployment Guide**（documentation）：MEMANTO Production Deployment Guide 证据：`docs/DEPLOYMENT_GUIDE.md`
- **Getting Started with MEMANTO**（documentation）：Audience : Developers building AI agents who need persistent memory Time to complete : 15-30 minutes Prerequisites : Basic Python knowledge, Docker optional 证据：`docs/GETTING_STARTED.md`
- **Hybrid Timeline Guide: Combining Automatic + Semantic**（documentation）：Hybrid Timeline Guide: Combining Automatic + Semantic 证据：`docs/HYBRID_TIMELINE_GUIDE.md`
- **MEMANTO Session-Based Architecture**（documentation）：Status : Design Specification Date : December 2025 Author : Dr. Majid Fekri, CTO Moorcheh.ai 证据：`docs/SESSION_ARCHITECTURE.md`
- **Timeline Visualization Examples**（documentation）：For : Developers building dashboards and visualization tools Status : Reference Examples 证据：`docs/TIMELINE_VISUALIZATION_EXAMPLES.md`
- **MEMANTO Quick Start Guide**（documentation）：5-Minute Guide to Session-Based API 证据：`docs/V2_QUICK_START.md`
- **Memanto Memory Agent Provider**（documentation）：Memanto https://memanto.ai is a memory agent — typed long-term memory with confidence and provenance, semantic recall, and RAG-style answers, backed by the Moorcheh https://moorcheh.ai vector platform. Each Memanto namespace is a first-class agent memanto agent create/activate , so this provider maps one Hermes identity to one Memanto agent. 证据：`integrations/hermes-agents/hermes_memanto/PLUGIN_README.md`
- **Openapi**（structured_config）：{ "openapi": "3.1.0", "info": { "title": "Memanto - Memory that AI Agents Love!", "description": "A memory layer service for agentic AI systems using Moorcheh SDK", "version": "0.0.0.dev0" }, "paths": { "/health": { "get": { "tags": "Health" , "summary": "Health Check", "description": "Health check endpoint", "operationId": "health check health get", "responses": { "200": { "description": "Successful Response", "content": { "application/json": { "schema": { "$ref": " /components/schemas/HealthResponse" } } } } } } }, "/ready": { "get": { "tags": "Health" , "summary": "Readiness Check", "description": "Readiness check for Kubernetes", "operationId": "readiness check ready get", "responses":… 证据：`sdks/typescript/openapi.json`
- **Tsconfig**（structured_config）：{ "compilerOptions": { "target": "ES2022", "module": "ESNext", "moduleResolution": "Bundler", "lib": "ES2022", "DOM" , "strict": true, "esModuleInterop": true, "skipLibCheck": true, "forceConsistentCasingInFileNames": true, "declaration": true, "declarationMap": true, "sourceMap": true, "outDir": "./dist", "rootDir": "./src", "resolveJsonModule": true, "isolatedModules": true, "types": "node" }, "include": "src/ / " , "exclude": "node modules", "dist", "test" } 证据：`sdks/typescript/tsconfig.json`
- **Config**（source_file）：DEFAULT AGENT ID = "skills-dev-profile" DEFAULT RECALL LIMIT = 8 DEFAULT MIN SIMILARITY: float None = None class ConfigError RuntimeError ⋮---- @dataclass frozen=True class SkillsConfig ⋮---- api key: str agent id: str = DEFAULT AGENT ID recall limit: int = DEFAULT RECALL LIMIT min similarity: float None = DEFAULT MIN SIMILARITY ⋮---- @classmethod def from env cls - SkillsConfig ⋮---- api key = os.environ.get "MOORCHEH API KEY" or "" .strip ⋮---- agent id = recall limit = int env "MEMANTO RECALL LIMIT", DEFAULT RECALL LIMIT min similarity = float env "MEMANTO MIN SIMILARITY", DEFAULT MIN SIMILARITY ⋮---- def int env name: str, default: int - int ⋮---- raw = os.environ.get name ⋮---- def flo… 证据：`examples/claudecode-skills-memanto/lifecycle-hooks/memanto_skills/config.py`
- **Main**（source_file）：def run session agent id, user id, task, thread id ⋮---- api key = os.getenv "MOORCHEH API KEY" ⋮---- client = SdkClient api key=api key tools = create memanto tools client, agent id app = create research graph "poolside/laguna-xs.2:free", tools config = {"configurable": {"thread id": thread id}} ⋮---- events = app.stream ⋮---- def main ⋮---- agent id = "langgraph-researcher-001" user id = "user-123" task1 = ⋮---- task2 = 证据：`examples/langgraph-memanto/cross_session_recall/main.py`
- **-----------------------------------------------------------------------**（source_file）：MEMORY CATEGORIES = MEMORY TYPES LITERAL = str class MemantoMemory ⋮---- ----------------------------------------------------------------------- Session lifecycle ⋮---- def ensure agent self - dict ⋮---- """Create the agent if it doesn't exist yet.""" ⋮---- def activate session self - str def deactivate session self - None ⋮---- tags = metadata.get "tags", if metadata else ⋮---- types = memory type if memory type else None res = self.client.recall ⋮---- def answer self, question: str - str ⋮---- res = self.client.answer agent id=self.agent name, question=question ⋮---- Batch helpers ⋮---- """Store up to 100 memories at once. Each tuple is text, memory type, metadata or None . """ payload =… 证据：`examples/langgraph-memanto/custom_memory_saver/memanto_memory.py`
- **Config**（source_file）：DEFAULT AGENT ID = "skills-dev-profile" DEFAULT RECALL LIMIT = 8 DEFAULT MIN SIMILARITY: float None = None class ConfigError RuntimeError ⋮---- @dataclass frozen=True class SkillsConfig ⋮---- api key: str agent id: str = DEFAULT AGENT ID recall limit: int = DEFAULT RECALL LIMIT min similarity: float None = DEFAULT MIN SIMILARITY ⋮---- @classmethod def from env cls - SkillsConfig ⋮---- api key = os.environ.get "MOORCHEH API KEY" or "" .strip ⋮---- agent id = recall limit = int env "MEMANTO RECALL LIMIT", DEFAULT RECALL LIMIT min similarity = float env "MEMANTO MIN SIMILARITY", DEFAULT MIN SIMILARITY ⋮---- def int env name: str, default: int - int ⋮---- raw = os.environ.get name ⋮---- def flo… 证据：`integrations/claudecode/claudecode_memanto/config.py`
- **Config**（source_file）：class TransportType str, Enum ⋮---- STDIO = "stdio" SSE = "sse" STREAMABLE HTTP = "streamable-http" VALID AGENT PATTERNS = {"support", "project", "tool"} class MCPServerSettings BaseSettings ⋮---- model config = SettingsConfigDict moorcheh api key: SecretStr = Field default agent id: str None = Field agent pattern: str = Field agent auto create: bool = Field session duration hours: int None = Field expose admin tools: bool = Field transport: TransportType = Field host: str = Field port: int = Field log level: str = Field ⋮---- @field validator "agent pattern" @classmethod def validate pattern cls, v: str - str ⋮---- allowed = ", ".join sorted VALID AGENT PATTERNS ⋮---- @field validator "log… 证据：`integrations/mcp/memanto_mcp/config.py`
- **Backend**（source_file）：class Backend str, Enum ⋮---- CLOUD = "cloud" ON PREM = "on-prem" def parse backend value: str None - Backend def get active llm model cloud default: str - str None ⋮---- state path = Path.home / ".memanto" / "on-prem" / "state.json" ⋮---- state = json.loads state path.read text 证据：`memanto/app/clients/backend.py`
- **Onprem**（source_file）：DEFAULT URL = "http://localhost:8080" def import raw client - Any def import docker runtime helpers - tuple Any, Any class DocumentsAdapter ⋮---- def init self, raw: Any - None def getattr self, name: str - Any def upload file self, namespace name: str, file path: str Path - dict ⋮---- src = Path file path .resolve ⋮---- upload root = ensure upload dir host file = upload root / src.name .resolve ⋮---- container path = host path to container upload path host file, upload root resp = self. raw.files.upload ⋮---- resp = {} ⋮---- class OnPremClient ⋮---- def init self, base url: str None = None, timeout: int None = None - None ⋮---- client cls = import raw client ⋮---- def health self - Any cla… 证据：`memanto/app/clients/onprem.py`
- **Backend selection: "cloud" default or "on-prem".**（source_file）：memanto env = Path.home / ".memanto" / ".env" ⋮---- config file = Path.home / ".memanto" / "config.yaml" ⋮---- data = yaml.safe load f memanto = data.get "memanto", {} answer = memanto.get "answer", {} ans model = answer.get "model" ⋮---- ans temp = answer.get "temperature" ⋮---- ans limit = answer.get "answer limit" ⋮---- summary = memanto.get "summary", {} sum model = summary.get "model" ⋮---- cli = memanto.get "cli", {} smart parse = cli.get "smart parse" ⋮---- backend = memanto.get "backend" ⋮---- state path = Path.home / ".memanto" / "on-prem" / "state.json" ⋮---- state = json.loads state path.read text op url = state.get "url" ⋮---- op embed = state.get "embedding provider" ⋮---- clas… 证据：`memanto/app/config.py`
- **Constants**（source_file）：MemoryType = Literal SourceType = str StatusType = Literal "active", "superseded", "deleted", "provisional" ProvenanceType = Literal ValidationMode = Literal "strict", "lenient", "off" ActorType = Literal "user", "agent", "system" ProvenanceSource = Literal "user", "agent", "tool", "system" VALID MEMORY TYPES = { VALID PROVENANCE TYPES = { ALLOWED UPDATE FIELDS = { VALID PATTERNS = {"support", "project", "tool"} 证据：`memanto/app/constants.py`
- **Memory**（source_file）：router = APIRouter ⋮---- memory = MemoryRecord ⋮---- service = MemoryWriteService client context = {"user confirmed": request.user confirmed} result = service.store memory memory, context ⋮---- memory records = ⋮---- result = service.batch store memories memory records, context ⋮---- result = service.update memory ⋮---- service = MemoryReadService client result = service.search memories ⋮---- scopes = result = service.search multi scope ⋮---- result = service.generate answer ⋮---- memory = service.get memory memory id, namespace ⋮---- success = service.delete memory memory id, namespace 证据：`memanto/app/legacy/memory.py`
- **Create FastAPI app**（source_file）：def validate startup dependencies - None ⋮---- backend = parse backend settings.MEMANTO BACKEND ⋮---- url = f"{settings.MOORCHEH ONPREM URL.rstrip '/' }/health" ⋮---- resp = httpx.get url, timeout=5.0 ⋮---- api key = settings.MOORCHEH API KEY.strip ⋮---- client = MoorchehClient api key=api key ⋮---- @asynccontextmanager async def lifespan : FastAPI Create FastAPI app app = FastAPI ⋮---- @app.get "/" async def root 证据：`memanto/app/main.py`
- **Create memory record with scope fields and provenance**（source_file）：router = APIRouter config manager = ConfigManager class RecallRequest BaseModel ⋮---- query: str = Field ..., min length=1, description="Search query" limit: int None = Field default=None, ge=1, description="Max results" min similarity: float None = Field type: list str None = Field default=None, description="Memory type filters" class RecallAsOfRequest BaseModel ⋮---- as of: datetime = Field ⋮---- @field validator "as of", mode="before" @classmethod def parse as of cls, v: object - datetime ⋮---- dt = datetime.fromisoformat v.replace "Z", "+00:00" ⋮---- class RecallChangedSinceRequest BaseModel ⋮---- since: datetime = Field ⋮---- @field validator "since", mode="before" @classmethod def par… 证据：`memanto/app/routes/memory.py`
- **Ensure parent directories exist**（source_file）：class DailyAnalysisService ⋮---- pattern = f"{agent id} {date} summary.md" session files = list self.sessions dir.glob pattern ⋮---- combined content = ⋮---- full text = "\n\n---\n\n".join combined content client = get moorcheh client namespace = agent namespace agent id summary prompt = f""" ⋮---- result = client.answer.generate summary text = result.get "answer", "Failed to generate summary." ⋮---- summary path = Path output path Ensure parent directories exist ⋮---- summary path = self.summaries dir / f"{agent id} {date}.md" ⋮---- viz service = SummaryVisualizationService ⋮---- def generate conflict report self, agent id: str, date: str - dict str, Any ⋮---- conflicts dir = Path.home / "… 证据：`memanto/app/services/daily_analysis_service.py`
- **Sections in canonical order**（source_file）：MEMORY TYPE META = { MEMORY TYPE ORDER = class MemoryExportService ⋮---- def init self, exports dir: Path None = None ⋮---- generated at = generated at or datetime.now .strftime "%Y-%m-%d %H:%M:%S" total = sum len mems for mems in memories by type.values type counts = {t: len mems for t, mems in memories by type.items if mems} lines: list str = ⋮---- summary parts = f"{t}: {c}" for t, c in type counts.items ⋮---- Sections in canonical order ⋮---- memories = memories by type.get mem type, ⋮---- title = mem.get "title" or "Untitled" content = mem.get "content" or "" .strip confidence = mem.get "confidence" tags = mem.get "tags", created at = mem.get "created at", "" status = mem.get "status",… 证据：`memanto/app/services/memory_export_service.py`
- **Count memories per hour**（source_file）：class SummaryVisualizationService ⋮---- HEADING RE = re.compile CONFIDENCE RE = re.compile ⋮---- """ Parse session MD files for a given agent/date and return a complete Markdown visualization block. Args: agent id: Agent identifier. date: Date string YYYY-MM-DD . sessions dir: Directory containing session summary MD files. Returns: Markdown string with visual insights may be empty if no data . """ memories = self. parse session files agent id, date, sessions dir ⋮---- sections: list str = timeline = self. build activity timeline memories ⋮---- distribution = self. build type distribution memories ⋮---- confidence = self. build confidence overview memories ⋮---- viz markdown = self.generate… 证据：`memanto/app/services/summary_visualization_service.py`
- **Validate JWT token**（source_file）：logger = logging.getLogger name class LowerStr str ⋮---- def title self def capitalize self class MoorchehClient ⋮---- def init self, api key: str, base url: str = "https://api.moorcheh.ai/v1" def request self, method: str, endpoint: str, json data: Any = None - Any ⋮---- url = f"{self.base url}{endpoint}" headers = { req data = req = urllib.request.Request url, data=req data, headers=headers, method=method ⋮---- body = e.read .decode "utf-8" ⋮---- err json = json.loads body ⋮---- @property def documents self ⋮---- class Docs ⋮---- def init self, client def upload self, namespace name, documents def delete self, namespace name, ids def get self, namespace name, ids ⋮---- @property def names… 证据：`memanto/cli/client/direct_client.py`
- **Validate JWT token**（source_file）：logger = logging.getLogger name all = "SdkClient" MAX BATCH SIZE = 100 MAX TITLE LENGTH = 100 MAX CONTENT LENGTH = InputLimits.MAX TEXT LENGTH class SdkClient ⋮---- def init self, api key: str - None def get moorcheh self def get write service self def get read service self def get agent service self def get session service self def get daily analysis service self def get export service self def get validated session for agent self, agent id: str ⋮---- session service = self. get session service ⋮---- Validate JWT token token payload = session service.validate session self.session token ⋮---- Surface the same specific session errors as the service ⋮---- Load the persisted session record ses… 证据：`memanto/cli/client/sdk_client.py`
- **On-prem: no API key needed. Pass a placeholder; the underlying**（source_file）：app = typer.Typer console = Console config manager = ConfigManager agent app = typer.Typer help="Agent management commands" session app = typer.Typer help="Legacy aliases for agent activation commands" config app = typer.Typer help="Configuration commands" schedule app = typer.Typer help="Daily summary scheduling commands" memory app = typer.Typer help="Memory management commands" connect app = typer.Typer help="Connect MEMANTO to external tools" migrate app = typer.Typer ⋮---- def error message: str, hint: str None = None - NoReturn ⋮---- body = message ⋮---- def warn message: str - None def get client - SdkClient ⋮---- """Get configured SDK client or exit if not initialized.""" ⋮---- back… 证据：`memanto/cli/commands/_shared.py`
- **Run the backend-specific setup if not already configured.**（source_file）：@config app.command "show" def config show ⋮---- api key = config manager.get api key server cfg = config manager.get server config cli cfg = config manager.get cli config ans cfg = config manager.get answer config rec cfg = config manager.get recall config ⋮---- schedule time = config manager.get schedule time table = Table title="MEMANTO Configuration", show header=False ⋮---- backend = config manager.get backend ⋮---- op = config manager.get onprem config ⋮---- current = config manager.get backend body = f"Active backend: {BRIGHT} {current.value} /{BRIGHT} " ⋮---- target = name.strip .lower ⋮---- target backend = Backend target ⋮---- Run the backend-specific setup if not already configur… 证据：`memanto/cli/commands/config_cmd.py`
- **Batch mode**（source_file）：start = time.perf counter ⋮---- client = get client agent id = active agent id ⋮---- raw = sys.stdin.read ⋮---- conversation path = Path from conversation ⋮---- raw = conversation path.read text encoding="utf-8" messages = json.loads raw ⋮---- result = client.extract memories from conversation elapsed = time.perf counter - start candidates = result.get "candidates", ⋮---- successful = result.get "successful", 0 failed = result.get "failed", 0 total = result.get "total submitted", len candidates ⋮---- Batch mode ⋮---- batch path = Path batch ⋮---- raw = batch path.read text encoding="utf-8" memories = json.loads raw ⋮---- Validate each item has at least 'content' ⋮---- result = client.batch… 证据：`memanto/cli/commands/memory.py`
- **Step 1 — resolve target only if we will actually write.**（source_file）：PROVIDER BUNDLES: dict str, dict str, Any = { ⋮---- getters = { ⋮---- label = PROVIDER BUNDLES provider "label" ⋮---- resolved = get fn ⋮---- stored = get fn ⋮---- entered = typer.prompt f" Enter your {label} API key", hide input=True ⋮---- def generate narrative prompt: str, , provider label: str - tuple str, str, str ⋮---- """Call the active agent's LLM for a comparison narrative best-effort .""" method = ⋮---- ans cfg = config manager.get answer config model = ans cfg.get "model", "unknown" last error = "" ⋮---- client = get client result = client.answer narrative = result or {} .get "answer", "" or "" ⋮---- last error = str exc ⋮---- last error = str reactivate exc ⋮---- bundle = PROVID… 证据：`memanto/cli/commands/migrate.py`
- **Per-backend data directory**（source_file）：yaml = importlib.import module "yaml" def normalize duplicated api key key: str - str ⋮---- key = key.strip ⋮---- half = len key // 2 ⋮---- class ConfigManager ⋮---- def init self, config dir: Path None = None def get api key self - str None ⋮---- key = os.environ.get "MOORCHEH API KEY", "" .strip ⋮---- def set api key self, api key: str - None ⋮---- """Save Moorcheh API key to ~/.memanto/.env.""" ⋮---- def get supermemory api key self - str None ⋮---- key = ⋮---- def set supermemory api key self, api key: str - None ⋮---- """Save Supermemory API key to ~/.memanto/.env.""" ⋮---- def get mem0 api key self - str None def set mem0 api key self, api key: str - None ⋮---- """Save Mem0 API key to… 证据：`memanto/cli/config/manager.py`
- **Schedule Manager**（source_file）：class ScheduleManager ⋮---- TASK NAME = "MemantoNightlyJob" LEGACY TASK NAMES = "MemantoDailySummary", def init self def remove legacy tasks self - None ⋮---- current cron = subprocess.run legacy markers = f" ⋮---- def command self - str def enable self, time str: str = "23:55" - dict str, Any def disable self - dict str, Any def get status self - dict str, Any def enable windows self, time str: str = "23:55" - dict str, Any ⋮---- command = ⋮---- def disable windows self - dict str, Any ⋮---- command = "schtasks", "/delete", "/tn", self.TASK NAME, "/f" ⋮---- def status windows self - dict str, Any ⋮---- command = "schtasks", "/query", "/tn", self.TASK NAME, "/fo", "LIST" ⋮---- result = subp… 证据：`memanto/cli/schedule_manager.py`
- **sdks/typescript/.gitignore**（source_file）：node modules/ dist/ src/generated/ .tsbuildinfo .DS Store 证据：`sdks/typescript/.gitignore`
- **Openapi Ts.Config**（source_file）：import { defineConfig } from "@hey-api/openapi-ts"; 证据：`sdks/typescript/openapi-ts.config.ts`
- **Fetch Openapi**（source_file）：// Fetch the latest OpenAPI spec from a running memanto server and write it // to ./openapi.json so codegen runs against the committed baseline. // // Usage: // node scripts/fetch-openapi.mjs --url http://localhost:8000 // MEMANTO OPENAPI URL=... node scripts/fetch-openapi.mjs 证据：`sdks/typescript/scripts/fetch-openapi.mjs`
- **Doctor**（source_file）：import { spawn } from "node:child process"; export interface DoctorResult { uvxAvailable: boolean; uvxVersion?: string; hint?: string; } ⋮---- export async function doctor uvxPath = "uvx" : Promise function run cmd: string, args: string : Promise 证据：`sdks/typescript/src/doctor.ts`
- **Index**（source_file）：import { readFile } from "node:fs/promises"; import { basename } from "node:path"; import { ServerLifecycle, type ServerOptions } from "./lifecycle.js"; ⋮---- export interface MemantoOptions extends ServerOptions { agentId: string; autoCreate?: boolean; } export interface RememberInput { content: string; type?: string; title?: string; confidence?: number; tags?: string ; source?: string; provenance?: string; } export interface BatchRememberItem { content: string; type?: string; title?: string; confidence?: number; tags?: string ; source?: string; provenance?: string; } export interface RecallInput { query: string; limit?: number; minSimilarity?: number; type?: string ; } export interface An… 证据：`sdks/typescript/src/index.ts`
- 其余 19 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`, `examples/claudecode-skills-memanto/lifecycle-hooks/README.md`, `examples/crewai-memory/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`, `examples/claudecode-skills-memanto/lifecycle-hooks/README.md`, `examples/crewai-memory/README.md`

## 用户开工前应该回答的问题

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

## 验收标准

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

---

## Doramagic Context Augmentation

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

## Human Manual 骨架

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

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

- **项目概览与系统架构**：importance `high`
  - source_paths: README.md, memanto/app/clients/backend.py, memanto/app/clients/onprem.py, memanto/app/clients/moorcheh.py, memanto/app/main.py
- **记忆操作、记忆类型与 CLI 命令**：importance `high`
  - source_paths: memanto/cli/commands/memory.py, memanto/cli/commands/memory_mgmt.py, memanto/cli/commands/agent.py, memanto/cli/commands/session.py, memanto/cli/commands/core.py
- **集成生态与 SDK**：importance `high`
  - source_paths: memanto/cli/commands/connect.py, memanto/cli/connect/engine.py, memanto/cli/connect/agent_registry.py, integrations/claudecode/claudecode_memanto/__init__.py, integrations/crewai/crewai_memanto/tools.py
- **部署、配置与安全加固**：importance `high`
  - source_paths: memanto/cli/commands/config_cmd.py, memanto/cli/config/manager.py, memanto/app/clients/backend.py, memanto/app/clients/onprem.py, memanto/app/config.py

## Repo Inspection Evidence / 源码检查证据

- repo_clone_verified: true
- repo_inspection_verified: true
- repo_commit: `61c7656cf5bd06a2c617cc2c7532f80c87e2f220`
- inspected_files: `Dockerfile`, `README.md`, `docker-compose.yml`, `pyproject.toml`, `docs/AGENT_INTEGRATION_GUIDE.md`, `docs/AGENT_MEMORY_BEST_PRACTICES.md`, `docs/CLI_INSTALLATION.md`, `docs/CLI_USER_GUIDE.md`, `docs/DEPLOYMENT_GUIDE.md`, `docs/GETTING_STARTED.md`, `docs/HYBRID_TIMELINE_GUIDE.md`, `docs/SESSION_ARCHITECTURE.md`, `docs/TIMELINE_VISUALIZATION_EXAMPLES.md`, `docs/V2_QUICK_START.md`, `examples/claudecode-skills-memanto/lifecycle-hooks/.claude-plugin/plugin.json`, `examples/claudecode-skills-memanto/lifecycle-hooks/README.md`, `examples/claudecode-skills-memanto/lifecycle-hooks/demo_session_1.py`, `examples/claudecode-skills-memanto/lifecycle-hooks/demo_session_2.py`, `examples/claudecode-skills-memanto/lifecycle-hooks/demo_session_3.py`, `examples/claudecode-skills-memanto/lifecycle-hooks/hooks/__init__.py`

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：[BOUNTY $100] 🐜 The Great Agentic Memory Showdown: Memanto Benchmarking & Evaluation Challenge

- 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：[BOUNTY $100] 🐜 The Great Agentic Memory Showdown: Memanto Benchmarking & Evaluation Challenge
- 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/moorcheh-ai/memanto/issues/639 | 来源讨论提到 node 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 2: 来源证据：[BOUNTY $100] 🐜The Memanto Bug & Exploit Challenge

- 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：[BOUNTY $100] 🐜The Memanto Bug & Exploit Challenge
- 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/moorcheh-ai/memanto/issues/770 | 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 3: 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/moorcheh-ai/memanto | host_targets=claude_code, claude, cursor
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 4: 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/moorcheh-ai/memanto | 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 5: 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/moorcheh-ai/memanto | 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/moorcheh-ai/memanto | 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: 存在评分风险

- Trigger: no_demo
- Why it matters: 风险会影响是否适合普通用户安装。
- Evidence: risks.scoring_risks | https://github.com/moorcheh-ai/memanto | 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 8: 来源证据：`detect-conflicts` (on-prem) fails on any active day because the conflict query exceeds the embedding model's context w…

- 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：`detect-conflicts` (on-prem) fails on any active day because the conflict query exceeds the embedding model's context window
- 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/moorcheh-ai/memanto/issues/1329 | 来源讨论提到 docker 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 9: issue/PR 响应质量未知

- Trigger: issue_or_pr_quality=unknown。
- Host AI rule: 抽样最近 issue/PR，判断是否长期无人处理。
- Why it matters: 用户无法判断遇到问题后是否有人维护。
- Evidence: evidence.maintainer_signals | https://github.com/moorcheh-ai/memanto | 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 10: 发布节奏不明确

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