# oh-no-my-claudecode - 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 oh-no-my-claudecode 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_0003` supported 0.86

## 它能做什么

- **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/marketplace.json`, `.claude-plugin/plugin.json` Claim：`clm_0001` 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_0002` supported 0.86

## 怎么开始

- `pip install oh-no-my-claudecode` 证据：`README.md` Claim：`clm_0004` supported 0.86
- `/plugin marketplace add adaline-ankit/oh-no-my-claudecode` 证据：`README.md` Claim：`clm_0005` supported 0.86
- `/plugin install oh-no-my-claudecode@onmc` 证据：`README.md` Claim：`clm_0006` supported 0.86
- `git clone https://github.com/adaline-ankit/oh-no-my-claudecode` 证据：`README.md` Claim：`clm_0007` supported 0.86
- `pip install -e ".[dev]"` 证据：`README.md` Claim：`clm_0008` supported 0.86

## 继续前判断卡

- **当前建议**：先做角色匹配试用
- **为什么**：这个项目更像角色库，核心风险是选错角色或把角色文案当执行能力；先用 Prompt Preview 试角色匹配，再决定是否沙盒导入。

### 30 秒判断

- **现在怎么做**：先做角色匹配试用
- **最小安全下一步**：先用 Prompt Preview 试角色匹配；满意后再隔离导入
- **先别相信**：角色质量和任务匹配不能直接相信。
- **继续会触碰**：角色选择偏差、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_0003` 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/marketplace.json`, `.claude-plugin/plugin.json` Claim：`clm_0001` 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_0002` 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_0004` supported 0.86

### 现在还不能相信

- **角色质量和任务匹配不能直接相信。**（unverified）：角色库证明有很多角色，不证明每个角色都适合你的具体任务，也不证明角色能产生高质量结果。
- **不能把角色文案当成真实执行能力。**（unverified）：安装前只能判断角色描述和任务画像是否匹配，不能证明它能在宿主 AI 里完成任务。
- **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/marketplace.json`, `.claude-plugin/plugin.json`, `AGENTS.md`, `CLAUDE.md`
- **可安全回滚不能默认相信。**（unverified）：除非项目明确提供卸载和恢复说明，否则必须先在隔离环境验证。
- **真实安装后是否与用户当前宿主 AI 版本兼容？**（unverified）：兼容性只能通过实际宿主环境验证。 证据：`.claude-plugin/marketplace.json`, `.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。

### 继续会触碰什么

- **角色选择偏差**：用户对任务应该由哪个专家角色处理的判断。 原因：选错角色会让 AI 从错误专业视角回答，浪费时间或误导决策。
- **命令执行**：包管理器、网络下载、本地插件目录、项目配置或用户主目录。 原因：运行第一条命令就可能产生环境改动；必须先判断是否值得跑。 证据：`README.md`
- **宿主 AI 配置**：Claude/Codex/Cursor/Gemini/OpenCode 等宿主的 plugin、Skill 或规则加载配置。 原因：宿主配置会改变 AI 后续工作方式，可能和用户已有规则冲突。 证据：`.claude-plugin/marketplace.json`, `.claude-plugin/plugin.json`, `AGENTS.md`, `CLAUDE.md`
- **本地环境或项目文件**：安装结果、插件缓存、项目配置或本地依赖目录。 原因：安装前无法证明写入范围和回滚方式，需要隔离验证。 证据：`.claude-plugin/marketplace.json`, `.claude-plugin/plugin.json`, `README.md`
- **宿主 AI 上下文**：AI Context Pack、Prompt Preview、Skill 路由、风险规则和项目事实。 原因：导入上下文会影响宿主 AI 后续判断，必须避免把未验证项包装成事实。

### 最小安全下一步

- **先跑 Prompt Preview**：先用交互式试用验证任务画像和角色匹配，不要先导入整套角色库。（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。
- **保留原始角色选择记录**：如果输出偏题，可以回到任务画像阶段重新选择角色，而不是继续沿着错误角色推进。
- **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：`.claude-plugin/marketplace.json`, `.claude-plugin/plugin.json` 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

- **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/marketplace.json`, `.claude-plugin/plugin.json` Claim：`clm_0001` 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_0002` supported 0.86

### 上下文规模

- 文件总数：343
- 重要文件覆盖：40/343
- 证据索引条目：76
- 角色 / Skill 条目：27

### 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 oh-no-my-claudecode 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 oh-no-my-claudecode 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 oh-no-my-claudecode 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 27 个角色 / Skill / 项目文档条目。

- **Agent Integration Guides**（project_doc）：Use onmc plug to wire onmc into your coding agent in one command. All writes are idempotent — re-running never duplicates stanzas. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/integrations/README.md`
- **ONMC launch assets**（project_doc）：Current campaign: v0.36 - memory-grounded autonomous coding loops with proof . 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/launch/README.md`
- **oh-no-my-claudecode onmc**（project_doc）：! CI https://github.com/adaline-ankit/oh-no-my-claudecode/actions/workflows/ci.yml/badge.svg https://github.com/adaline-ankit/oh-no-my-claudecode/actions ! Latest release https://img.shields.io/github/v/release/adaline-ankit/oh-no-my-claudecode https://github.com/adaline-ankit/oh-no-my-claudecode/releases/latest ! PyPI version https://badge.fury.io/py/oh-no-my-claudecode.svg https://pypi.org/project/oh-no-my-claudec… 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`README.md`
- **AGENTS.md**（project_doc）：This file gives Codex, cloud coding agents, and other repository-aware agents the same maintainer expectations that Claude Code receives from CLAUDE.md . 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`AGENTS.md`
- **CLAUDE.md**（project_doc）：This file is source-controlled project guidance for Claude Code contributors. Do not replace it with generated ONMC memory output for this repository. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`CLAUDE.md`
- **Contributing**（project_doc）：Run the smallest relevant test while developing, then run the full gate before asking for review. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`CONTRIBUTING.md`
- **Changelog**（project_doc）：All notable changes to this project are documented here. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`CHANGELOG.md`
- **Releasing oh-no-my-claudecode**（project_doc）：This document is the authoritative guide for cutting a release. The pipeline uses PyPI trusted publishing OIDC — no API token secrets are stored in GitHub. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/RELEASING.md`
- **Agent-Native Workflows**（project_doc）：ONMC is designed to make coding agents start with useful repo memory instead of a blank context window. This document defines the supported agent surfaces and the guardrails contributors should preserve. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/agent-native-workflows.md`
- **Architecture**（project_doc）：- local-first - useful without an LLM - deterministic where possible - provenance-driven - easy to publish as a small OSS package 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/architecture.md`
- **CLI Reference**（project_doc）：This file is generated from Typer help output. Run python scripts/generate-cli-reference.py after changing CLI commands. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/cli-reference.md`
- **Two Agents, One Brain**（project_doc）：A walkthrough of onmc's cross-agent shared memory — how Agent A records a dead-end and Agent B avoids repeating it, just by cloning the same repository. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/demo.md`
- **🧠 oh-no-my-claudecode onmc**（project_doc）：Git-portable, cross-agent memory for AI coding agents. Claude Code, Codex, and Cursor share one provenanced brain: memory-grounded autonomous loops with verifier gates, replay, evals, agent-config security auditing, and tamper-evident receipts. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/index.md`
- **Memory Model**（project_doc）：P0 memory is typed structured repo knowledge, not raw chat history. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/memory-model.md`
- **Obsidian Vault Export**（project_doc）：ONMC can turn repository memory into a local Obsidian knowledge graph. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/obsidian.md`
- **Prompt Compiler**（project_doc）：The prompt compiler is ONMC's first bridge between deterministic repo memory and optional LLM reasoning. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/prompt-compiler.md`
- **Roadmap**（project_doc）：- onmc setup onboarding wizard - installable Python package - repo-local .onmc/ state - SQLite memory store - task lifecycle foundation - task attempt logging - git-portable .agent-memory/ sync - Claude Code compaction hooks + continuation snapshots - typed public Python API - MCP server tools + resources - doc ingestion - git-history ingestion - hotspot and git-pattern extraction - task brief compilation - token-ef… 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/roadmap.md`
- **Shipped Capabilities**（project_doc）：This document is the fastest way for a human or coding agent to understand what ONMC currently ships. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/shipped-capabilities.md`
- **Task Lifecycle**（project_doc）：ONMC now treats tasks as first-class local records, not just free-form prompt text. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/task-lifecycle.md`
- **ONMC Visual Dashboard**（project_doc）：Provide a local, read-only dashboard that makes ONMC memory, tasks, repository hotspots, and health signals easy to inspect and share. The first release targets maintainers and coding-agent users who want a useful visual surface without operating a hosted service. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/ui-dashboard.md`
- **Claude Code integration**（project_doc）：Option A — install via plugin marketplace recommended 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/integrations/claude-code.md`
- **Codex / GitHub Coding Agent integration**（project_doc）：Codex / GitHub Coding Agent integration 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/integrations/codex.md`
- **Cursor integration**（project_doc）：Cursor =0.40 reads markdown rule files from .cursor/rules/ . onmc plug cursor writes .cursor/rules/onmc.md so Cursor always knows to run onmc brief and onmc guard before starting any task. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/integrations/cursor.md`
- **Reddit post variants**（project_doc）：Title: I built a git-portable memory layer for coding agents — the brain commits with the repo 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/launch/reddit.md`
- **Show HN: oh-no-my-claudecode — git-portable memory for coding agents**（project_doc）：Show HN: oh-no-my-claudecode — git-portable memory for coding agents 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/launch/show-hn.md`
- **Ecosystem-List Submission Materials**（project_doc）：Ecosystem-List Submission Materials 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/launch/submissions.md`
- **ONMC v0.36 launch kit**（project_doc）：Category: memory-grounded execution layer for coding agents 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/launch/v0.36-launch-kit.md`

## 证据索引

- 共索引 76 条证据。

- **Agent Integration Guides**（documentation）：Use onmc plug to wire onmc into your coding agent in one command. All writes are idempotent — re-running never duplicates stanzas. 证据：`docs/integrations/README.md`
- **ONMC launch assets**（documentation）：Current campaign: v0.36 - memory-grounded autonomous coding loops with proof . 证据：`docs/launch/README.md`
- **oh-no-my-claudecode onmc**（documentation）：! CI https://github.com/adaline-ankit/oh-no-my-claudecode/actions/workflows/ci.yml/badge.svg https://github.com/adaline-ankit/oh-no-my-claudecode/actions ! Latest release https://img.shields.io/github/v/release/adaline-ankit/oh-no-my-claudecode https://github.com/adaline-ankit/oh-no-my-claudecode/releases/latest ! PyPI version https://badge.fury.io/py/oh-no-my-claudecode.svg https://pypi.org/project/oh-no-my-claudecode/ ! Python 3.11+ https://img.shields.io/badge/python-3.11+-blue.svg https://www.python.org/downloads/ ! License: MIT https://img.shields.io/badge/License-MIT-yellow.svg LICENSE 证据：`README.md`
- **AGENTS.md**（documentation）：This file gives Codex, cloud coding agents, and other repository-aware agents the same maintainer expectations that Claude Code receives from CLAUDE.md . 证据：`AGENTS.md`
- **CLAUDE.md**（documentation）：This file is source-controlled project guidance for Claude Code contributors. Do not replace it with generated ONMC memory output for this repository. 证据：`CLAUDE.md`
- **Contributing**（documentation）：Run the smallest relevant test while developing, then run the full gate before asking for review. 证据：`CONTRIBUTING.md`
- **Marketplace**（structured_config）：{ "$schema": "https://json.schemastore.org/claude-code-plugin-marketplace.json", "name": "onmc", "description": "ONMC - memory-grounded autonomous coding loops and No-Mistakes PR gates for Claude Code", "owner": { "name": "ONMC Contributors", "email": "ankit@adaline.ai" }, "plugins": { "name": "oh-no-my-claudecode", "displayName": "ONMC - Memory-Grounded Coding Loops", "description": "Run Claude Code with repo memory, No-Mistakes PR gates, known-dead-end guards, verifier gates, hard budgets, and tamper-evident receipts.", "source": { "source": "github", "repo": "adaline-ankit/oh-no-my-claudecode" }, "version": "0.48.0", "author": { "name": "ONMC Contributors" }, "homepage": "https://github.… 证据：`.claude-plugin/marketplace.json`
- **Plugin**（structured_config）：{ "$schema": "https://json.schemastore.org/claude-code-plugin-manifest.json", "name": "oh-no-my-claudecode", "displayName": "ONMC - Memory-Grounded Coding Loops", "description": "Run Claude Code with repo memory, No-Mistakes PR gates, known-dead-end guards, verifier gates, hard budgets, and tamper-evident receipts. Also ships lifecycle hooks and 12 MCP memory tools.", "version": "0.48.0", "author": { "name": "ONMC Contributors", "url": "https://github.com/adaline-ankit/oh-no-my-claudecode" }, "homepage": "https://github.com/adaline-ankit/oh-no-my-claudecode", "repository": "https://github.com/adaline-ankit/oh-no-my-claudecode", "license": "MIT", "keywords": "autonomous-agents", "agent-memor… 证据：`.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`
- **Changelog**（documentation）：All notable changes to this project are documented here. 证据：`CHANGELOG.md`
- **Releasing oh-no-my-claudecode**（documentation）：This document is the authoritative guide for cutting a release. The pipeline uses PyPI trusted publishing OIDC — no API token secrets are stored in GitHub. 证据：`docs/RELEASING.md`
- **Agent-Native Workflows**（documentation）：ONMC is designed to make coding agents start with useful repo memory instead of a blank context window. This document defines the supported agent surfaces and the guardrails contributors should preserve. 证据：`docs/agent-native-workflows.md`
- **Architecture**（documentation）：- local-first - useful without an LLM - deterministic where possible - provenance-driven - easy to publish as a small OSS package 证据：`docs/architecture.md`
- **CLI Reference**（documentation）：This file is generated from Typer help output. Run python scripts/generate-cli-reference.py after changing CLI commands. 证据：`docs/cli-reference.md`
- **Two Agents, One Brain**（documentation）：A walkthrough of onmc's cross-agent shared memory — how Agent A records a dead-end and Agent B avoids repeating it, just by cloning the same repository. 证据：`docs/demo.md`
- **🧠 oh-no-my-claudecode onmc**（documentation）：Git-portable, cross-agent memory for AI coding agents. Claude Code, Codex, and Cursor share one provenanced brain: memory-grounded autonomous loops with verifier gates, replay, evals, agent-config security auditing, and tamper-evident receipts. 证据：`docs/index.md`
- **Memory Model**（documentation）：P0 memory is typed structured repo knowledge, not raw chat history. 证据：`docs/memory-model.md`
- **Obsidian Vault Export**（documentation）：ONMC can turn repository memory into a local Obsidian knowledge graph. 证据：`docs/obsidian.md`
- **Prompt Compiler**（documentation）：The prompt compiler is ONMC's first bridge between deterministic repo memory and optional LLM reasoning. 证据：`docs/prompt-compiler.md`
- **Roadmap**（documentation）：- onmc setup onboarding wizard - installable Python package - repo-local .onmc/ state - SQLite memory store - task lifecycle foundation - task attempt logging - git-portable .agent-memory/ sync - Claude Code compaction hooks + continuation snapshots - typed public Python API - MCP server tools + resources - doc ingestion - git-history ingestion - hotspot and git-pattern extraction - task brief compilation - token-efficient compact/caveman briefs and codegraph navigation - local visual dashboard for memory, tasks, codegraph, and health - portable single-file dashboard snapshots for demos and handoffs - Obsidian vault export with provenance notes and memory relationship links - optional LLM-p… 证据：`docs/roadmap.md`
- **Shipped Capabilities**（documentation）：This document is the fastest way for a human or coding agent to understand what ONMC currently ships. 证据：`docs/shipped-capabilities.md`
- **Task Lifecycle**（documentation）：ONMC now treats tasks as first-class local records, not just free-form prompt text. 证据：`docs/task-lifecycle.md`
- **ONMC Visual Dashboard**（documentation）：Provide a local, read-only dashboard that makes ONMC memory, tasks, repository hotspots, and health signals easy to inspect and share. The first release targets maintainers and coding-agent users who want a useful visual surface without operating a hosted service. 证据：`docs/ui-dashboard.md`
- **Claude Code integration**（documentation）：Option A — install via plugin marketplace recommended 证据：`docs/integrations/claude-code.md`
- **Codex / GitHub Coding Agent integration**（documentation）：Codex / GitHub Coding Agent integration 证据：`docs/integrations/codex.md`
- **Cursor integration**（documentation）：Cursor =0.40 reads markdown rule files from .cursor/rules/ . onmc plug cursor writes .cursor/rules/onmc.md so Cursor always knows to run onmc brief and onmc guard before starting any task. 证据：`docs/integrations/cursor.md`
- **Reddit post variants**（documentation）：Title: I built a git-portable memory layer for coding agents — the brain commits with the repo 证据：`docs/launch/reddit.md`
- **Show HN: oh-no-my-claudecode — git-portable memory for coding agents**（documentation）：Show HN: oh-no-my-claudecode — git-portable memory for coding agents 证据：`docs/launch/show-hn.md`
- **Ecosystem-List Submission Materials**（documentation）：Ecosystem-List Submission Materials 证据：`docs/launch/submissions.md`
- **ONMC v0.36 launch kit**（documentation）：Category: memory-grounded execution layer for coding agents 证据：`docs/launch/v0.36-launch-kit.md`
- **Init**（source_file）：all = " version ", "OnmcRepo", "init" ⋮---- version = version "oh-no-my-claudecode" ⋮---- version = "0+unknown" 证据：`src/oh_no_my_claudecode/__init__.py`
- **Init**（source_file）：all = "AskResult", "compile ask" 证据：`src/oh_no_my_claudecode/ask/__init__.py`
- **Init**（source_file）：all = "AuditFinding", "AuditReport", "run audit" 证据：`src/oh_no_my_claudecode/audit/__init__.py`
- **Init**（source_file）：all = "AutopilotResult", "BrainCounts", "run autopilot" 证据：`src/oh_no_my_claudecode/autopilot/__init__.py`
- **Models**（source_file）：@dataclass class BrainCounts ⋮---- memories: int skills: int dead ends: int ⋮---- @dataclass class AutopilotResult ⋮---- goal: str know brief summary: str know dead ends count: int know profile applied: bool loop result: object receipt path: Path None verified: bool tokens: int cost usd: float None brain before: BrainCounts brain after: BrainCounts memories added: int skills added: int dead ends recorded: int skill promoted name: str None captured count: int consolidated count: int stop reason: str know context: str = field default="" plan model: str None = field default=None execute model: str None = field default=None plan used: bool = field default=False plan tokens: int None = field def… 证据：`src/oh_no_my_claudecode/autopilot/models.py`
- **Init**（source_file）：all = "BenchResult", "BenchScenario", "run benchmark" 证据：`src/oh_no_my_claudecode/bench/__init__.py`
- **Init**（source_file）：all = "BenchmarkMetric", "BenchmarkReport", "run benchmark suite" 证据：`src/oh_no_my_claudecode/benchmark/__init__.py`
- **Init**（source_file）：all = "BlameAnchor", "BlameResult", "compile blame" 证据：`src/oh_no_my_claudecode/blame/__init__.py`
- **Init**（source_file）：all = 证据：`src/oh_no_my_claudecode/check/__init__.py`
- **Engine**（source_file）：WARN KINDS: frozenset MemoryKind = frozenset INFO KINDS: frozenset MemoryKind = frozenset WARN ARTIFACT TYPES: frozenset MemoryArtifactType = frozenset class CheckSeverity StrEnum ⋮---- WARN = "warn" INFO = "info" ⋮---- @dataclass class CheckFinding ⋮---- rel path: str severity: CheckSeverity kind: str memory id: str title: str summary: str ⋮---- @dataclass class CheckResult ⋮---- findings: list CheckFinding = field default factory=list ⋮---- @property def warn count self - int ⋮---- @property def info count self - int ⋮---- @property def has warnings self - bool def findings for self, rel path: str - list CheckFinding ⋮---- memories = storage.list memories artifacts = storage.list memory a… 证据：`src/oh_no_my_claudecode/check/engine.py`
- **Init**（source_file）：all = 证据：`src/oh_no_my_claudecode/claude_md/__init__.py`
- **Init**（source_file）：all = 证据：`src/oh_no_my_claudecode/codegraph/__init__.py`
- **Models**（source_file）：SymbolKind = Literal "func", "class" ⋮---- @dataclass slots=True, frozen=True class Symbol ⋮---- name: str kind: SymbolKind file: str lineno: int ⋮---- @dataclass slots=True class GraphNode ⋮---- symbols: list Symbol = field default factory=list imports: list str = field default factory=list is test: bool = False ⋮---- @dataclass slots=True class CodeGraph ⋮---- nodes: dict str, GraphNode = field default factory=dict ⋮---- symbols by name: dict str, list str = field default factory=dict dependents: dict str, list str = field default factory=dict file tests: dict str, list str = field default factory=dict file count: int = 0 def to dict self - dict str, object ⋮---- @classmethod def from dic… 证据：`src/oh_no_my_claudecode/codegraph/models.py`
- **Init**（source_file）：all = 证据：`src/oh_no_my_claudecode/conventions/__init__.py`
- **Init**（source_file）：all = 证据：`src/oh_no_my_claudecode/coverage/__init__.py`
- **Init**（source_file）：all = "DigestEntry", "DigestResult", "compile digest" 证据：`src/oh_no_my_claudecode/digest/__init__.py`
- **Init**（source_file）：all = 证据：`src/oh_no_my_claudecode/embeddings/__init__.py`
- **Init**（source_file）：all = 证据：`src/oh_no_my_claudecode/evals/__init__.py`
- **Models**（source_file）：@dataclass class EvalCase ⋮---- id: str query: str expected files: list str = field default factory=list expected deadend substrings: list str = field default factory=list note: str = "" ⋮---- @dataclass class EvalCaseResult ⋮---- """Result of running one :class: EvalCase . Attributes ---------- case id: Corresponds to :attr: EvalCase.id . files hit: True when expected files is non-empty and at least one entry in compile recall output matches. True vacuous when expected files is empty. deadend hit: True when expected deadend substrings is non-empty and at least one guard entry matches. True vacuous when expected deadend substrings is empty. recall entries: Count of :class: ~oh no my claudec… 证据：`src/oh_no_my_claudecode/evals/models.py`
- **Init**（source_file）：all = "EvolutionReport", "RunPoint", "compile evolution" 证据：`src/oh_no_my_claudecode/evolution/__init__.py`
- **Init**（source_file）：all = 证据：`src/oh_no_my_claudecode/federation/__init__.py`
- **Init**（source_file）：all = "GuardEntry", "GuardResult", "compile guard" 证据：`src/oh_no_my_claudecode/guard/__init__.py`
- **Init**（source_file）：all = 证据：`src/oh_no_my_claudecode/hooks/__init__.py`
- **Append compact profile lines mistakes-to-avoid in terse mode when verbose.**（source_file）：BOOT DIGEST MAX TOKENS = 400 MAX INVARIANTS = 3 MAX HOTSPOTS = 3 MAX ACTIVE TASKS = 2 MAX USER PREFS = 5 MAX BOOT SKILLS = 2 MAX PROFILE ITEMS = 2 ⋮---- terse = is terse default=True invariants = select kind hotspots = select kind active tasks = t for t in tasks if t.status == TaskStatus.ACTIVE prefs = m for m in user memories or if m.feedback score -0.5 top skills = select top skills skills or , max items= MAX BOOT SKILLS profile = compile profile safe prefs ⋮---- text = render boot digest terse Append compact profile lines mistakes-to-avoid in terse mode when verbose. profile text = render profile terse profile ⋮---- text = f"{text}\n{profile text}" if text else profile text Append compac… 证据：`src/oh_no_my_claudecode/hooks/boot_digest.py`
- **Init**（source_file）：all = "ImportResult", "run import" ⋮---- msg = f"as kind must be 'skill' or 'memory', got: {as kind!r}" ⋮---- base cwd = cwd or Path.cwd ⋮---- md path = path if path is not None else Path source ⋮---- dirs = omc.resolve omc dirs path, cwd=cwd skills = omc.parse dirs imported = 0 skipped = 0 item names: list str = ⋮---- existing = storage.get skill skill.id ⋮---- files = hermes.resolve hermes files path, cwd=cwd memories = hermes.parse files ⋮---- existing = storage.get memory memory.id ⋮---- files = markdown.resolve md paths path ⋮---- skills = markdown.parse as skills files ⋮---- memories = markdown.parse as memories files 证据：`src/oh_no_my_claudecode/importers/__init__.py`
- **Base**（source_file）：@dataclass class ImportResult ⋮---- source: str as kind: str imported: int skipped: int dry run: bool items: list str = field default factory=list 证据：`src/oh_no_my_claudecode/importers/base.py`
- **Auto-detect in cwd.**（source_file）：Auto-detect in cwd. 证据：`src/oh_no_my_claudecode/importers/hermes.py`
- **Markdown**（source_file）：IMPORT TAG = "imported:md" def skill from file path: Path - Skill ⋮---- body = path.read text encoding="utf-8" name = omc parser. derive name body, path.stem trigger = omc parser. derive trigger body, name now = utc now ⋮---- def memories from file path: Path - list MemoryEntry ⋮---- text = path.read text encoding="utf-8" entries: list MemoryEntry = ⋮---- kind = hermes parser. infer kind title if title else MemoryKind.DOC FACT summary = body :200 .strip ⋮---- entry = MemoryEntry ⋮---- def resolve md paths path: Path - list Path ⋮---- files = sorted path.glob " .md" ⋮---- msg = f"No .md files found in directory: {path}" ⋮---- msg = f"Path not found: {path}" ⋮---- def parse as skills files: l… 证据：`src/oh_no_my_claudecode/importers/markdown.py`
- **Docs**（source_file）：EXCLUDED DOC PATTERNS = OUTPUT FILES = {"CLAUDE.md", "AGENTS.md", ".cursorrules"} TOC PATTERNS = def discover doc paths repo root: Path, , globs: list str - list Path ⋮---- discovered: set Path = set ⋮---- def is primary doc path: Path - bool ⋮---- name = path.name ⋮---- def should ingest doc path path: Path - bool def extract doc memories repo root: Path, doc path: Path, , max chars: int - list MemoryEntry ⋮---- content = doc path.read text encoding="utf-8" sections = split markdown sections content relative = relative path repo root, doc path now = utc now memories: list MemoryEntry = ⋮---- text = body.strip ⋮---- clipped = text :max chars .strip kind = classify doc section heading, clipp… 证据：`src/oh_no_my_claudecode/ingest/docs.py`
- **Llm Extractor**（source_file）：COMMIT BATCH SIZE = 50 MAX LLM COMMITS = 500 MAX SOURCE FILES = 20 logger = logging.getLogger name class ExtractedKnowledgeItem BaseModel ⋮---- kind: str title: str summary: str confidence: float = Field ge=0.0, le=1.0 source commits: list str = Field default factory=list files mentioned: list str = Field default factory=list class ExtractedKnowledgeList RootModel list ExtractedKnowledgeItem ⋮---- limited lines = commit lines :MAX LLM COMMITS resolved batch size = batch size or get batch size total commits or len limited lines batches: list str = ⋮---- chunk = limited lines index : index + resolved batch size ⋮---- def get batch size total commits: int - int ⋮---- commit extracted: list Mem… 证据：`src/oh_no_my_claudecode/ingest/llm_extractor.py`
- 其余 16 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：`docs/integrations/README.md`, `docs/launch/README.md`, `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：`docs/integrations/README.md`, `docs/launch/README.md`, `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。

- **Introduction & Quick Start**：importance `high`
  - source_paths: README.md, pyproject.toml, src/oh_no_my_claudecode/__init__.py, src/oh_no_my_claudecode/cli.py, src/oh_no_my_claudecode/api.py
- **Autonomous Loops, Autopilot, Swarm & No-Mistakes Gate**：importance `high`
  - source_paths: src/oh_no_my_claudecode/loop/engine.py, src/oh_no_my_claudecode/loop/models.py, src/oh_no_my_claudecode/loop/adapters.py, src/oh_no_my_claudecode/loop/checkpoint.py, src/oh_no_my_claudecode/loop/receipt.py
- **Repository Memory, Briefs, Guards & Sync**：importance `high`
  - source_paths: AGENT-MEMORY-SPEC.md, src/oh_no_my_claudecode/memory/catalog.py, src/oh_no_my_claudecode/memory/consolidation.py, src/oh_no_my_claudecode/memory/staleness.py, src/oh_no_my_claudecode/ingest/pipeline.py
- **Cross-Agent Plug, MCP Server, Hooks & Skill Export**：importance `high`
  - source_paths: src/oh_no_my_claudecode/integrations/plug.py, src/oh_no_my_claudecode/integrations/gh_aw.py, src/oh_no_my_claudecode/mcp_server/server.py, src/oh_no_my_claudecode/mcp_server/tools.py, src/oh_no_my_claudecode/mcp_server/resources.py

## Repo Inspection Evidence / 源码检查证据

- repo_clone_verified: true
- repo_inspection_verified: true
- repo_commit: `90e772a3c5746d6a08881fbe7e995b8e7cc24032`
- inspected_files: `README.md`, `pyproject.toml`, `uv.lock`, `docs/RELEASING.md`, `docs/_config.yml`, `docs/agent-native-workflows.md`, `docs/architecture.md`, `docs/cli-reference.md`, `docs/demo.md`, `docs/index.md`, `docs/integrations/README.md`, `docs/integrations/claude-code.md`, `docs/integrations/codex.md`, `docs/integrations/cursor.md`, `docs/launch/README.md`, `docs/launch/reddit.md`, `docs/launch/show-hn.md`, `docs/launch/submissions.md`, `docs/launch/v0.36-launch-kit.md`, `docs/memory-model.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: v0.48.0

- Trigger: Developers should check this installation risk before relying on the project: v0.48.0
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: v0.48.0. Context: Observed during installation or first-run setup.
- Why it matters: Upgrade or migration may change expected behavior: v0.48.0
- Evidence: failure_mode_cluster:github_release | https://github.com/adaline-ankit/oh-no-my-claudecode/releases/tag/v0.48.0 | v0.48.0
- 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: 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/adaline-ankit/oh-no-my-claudecode | host_targets=claude_code, claude, mcp_host
- 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：configuration: v0.42.0

- Trigger: Developers should check this configuration risk before relying on the project: v0.42.0
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: v0.42.0. Context: Source discussion did not expose a precise runtime context.
- Why it matters: Upgrade or migration may change expected behavior: v0.42.0
- Evidence: failure_mode_cluster:github_release | https://github.com/adaline-ankit/oh-no-my-claudecode/releases/tag/v0.42.0 | v0.42.0
- 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：configuration: v0.43.0

- Trigger: Developers should check this configuration risk before relying on the project: v0.43.0
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: v0.43.0. Context: Source discussion did not expose a precise runtime context.
- Why it matters: Upgrade or migration may change expected behavior: v0.43.0
- Evidence: failure_mode_cluster:github_release | https://github.com/adaline-ankit/oh-no-my-claudecode/releases/tag/v0.43.0 | v0.43.0
- 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：configuration: v0.44.0

- Trigger: Developers should check this configuration risk before relying on the project: v0.44.0
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: v0.44.0. Context: Source discussion did not expose a precise runtime context.
- Why it matters: Upgrade or migration may change expected behavior: v0.44.0
- Evidence: failure_mode_cluster:github_release | https://github.com/adaline-ankit/oh-no-my-claudecode/releases/tag/v0.44.0 | v0.44.0
- 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：configuration: v0.45.0

- Trigger: Developers should check this configuration risk before relying on the project: v0.45.0
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: v0.45.0. Context: Source discussion did not expose a precise runtime context.
- Why it matters: Upgrade or migration may change expected behavior: v0.45.0
- Evidence: failure_mode_cluster:github_release | https://github.com/adaline-ankit/oh-no-my-claudecode/releases/tag/v0.45.0 | v0.45.0
- 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：configuration: v0.46.0

- Trigger: Developers should check this configuration risk before relying on the project: v0.46.0
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: v0.46.0. Context: Source discussion did not expose a precise runtime context.
- Why it matters: Upgrade or migration may change expected behavior: v0.46.0
- Evidence: failure_mode_cluster:github_release | https://github.com/adaline-ankit/oh-no-my-claudecode/releases/tag/v0.46.0 | v0.46.0
- 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：configuration: v0.47.0

- Trigger: Developers should check this configuration risk before relying on the project: v0.47.0
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: v0.47.0. Context: Source discussion did not expose a precise runtime context.
- Why it matters: Upgrade or migration may change expected behavior: v0.47.0
- Evidence: failure_mode_cluster:github_release | https://github.com/adaline-ankit/oh-no-my-claudecode/releases/tag/v0.47.0 | v0.47.0
- 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：configuration: v0.49.0

- Trigger: Developers should check this configuration risk before relying on the project: v0.49.0
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: v0.49.0. Context: Source discussion did not expose a precise runtime context.
- Why it matters: Upgrade or migration may change expected behavior: v0.49.0
- Evidence: failure_mode_cluster:github_release | https://github.com/adaline-ankit/oh-no-my-claudecode/releases/tag/v0.49.0 | v0.49.0
- 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：configuration: v0.50.0

- Trigger: Developers should check this configuration risk before relying on the project: v0.50.0
- Host AI rule: Before packaging this project, run the relevant install/config/quickstart check for: v0.50.0. Context: Source discussion did not expose a precise runtime context.
- Why it matters: Upgrade or migration may change expected behavior: v0.50.0
- Evidence: failure_mode_cluster:github_release | https://github.com/adaline-ankit/oh-no-my-claudecode/releases/tag/v0.50.0 | v0.50.0
- 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。