ClaudeForge
A CLAUDE.md Generator and Maintenance tool for for Claude Code to create high-quality CLAUDE.md instruction files — aligned with Anthropic’s best practices for Claude Code.
- 评测生成时间(北京时间)
- 本报告引擎
- v3.10.0
- 当前引擎
- v3.17.3
本报告与当前引擎使用不同规则;原分数不会自动更新,不同版本的分数不宜直接对比。
进入后确认来源与额度,提交才会创建任务。
代码版本未记录或记录无效,无法核对本报告的 commit 与送评文本;不会用当前项目代码回填历史记录。
综合采用结论
存在需要人工复核的风险或证据不足
- 基础评测完成+25/25确定性评分与静态安全扫描已完成
- README 有效证据+25/2515,289 个去重后的有效字符
- 独立证据来源+20/205 类非重复证据,重复文件不叠加
- 仓库元数据+10/10已取得仓库状态与采用数据
- 活跃记录+5/5已取得最近提交时间
- AI 复核+15/15已完成结构化 AI 证据复核
本报告评了什么
综合分用于静态采用筛选,不是任务成功率,也不代表使用 Skill 后的能力提升。
- 静态检查范围
- 记录 6 个文件 · 44,848 字符
- 只检查采集到的受限文本与已知风险模式;未读取的脚本、依赖和运行环境不在结论内。
- AI 内容复核
- 已使用
- 复核送评文本的说明和设计,不是安装测试或实际任务执行。记录缺失时不推断已使用或未使用。
- 真实任务效果
- 本报告未提供实测证据
- 没有使用/不使用 Skill 的任务对照、重复运行或耗时与 token 成本结果。采用前仍需在隔离环境验证自己的任务。
脚本引用覆盖
脚本引用覆盖记录不可得;不推断没有引用或已经检查。
ClaudeForge 组件架构与交互
README 描述了多个组件(技能、命令、代理、钩子)及其关系,适合用架构图展示。
左右滑动查看完整图示
- • Slash commands 部分描述 /enhance-claude-md 调用 claudeforge-skill
- • Agent 部分说明守护代理通过 Skill 工具调用技能
- • Hooks 部分列出事件如 PostToolUse、InstructionsLoaded
- Features 部分列出交互式初始化、智能分析、增强、后台维护等
- Quick Start 提供插件安装命令:/plugin marketplace add alirezarezvani/ClaudeForge
- Components Deep Dive 详述 claudeforge-skill 的质量评分标准(0-100)
- Requirements 明确要求 Claude Code 2.0+,Git 推荐
- License 部分声明 MIT 许可
- 问题与用途描述
- 有效 README
- 安装或接入步骤
- 可执行示例
- 明确的问题定义:自动化 CLAUDE.md 创建与维护
- 发现高风险的一键下载执行或安装命令
- 缺少限制、权限或边界
- 高级功能依赖 Claude Code 2.1.4+,未说明降级行为
- 示例多为模拟对话,缺少真实输出验证
- 未明确说明各组件在无 Claude Code 环境下的行为
Claude Code 项目需要自动化 CLAUDE.md 生成与维护、团队希望统一 CLAUDE.md 格式与质量、需要跨工具(AGENTS.md)兼容的项目、需要后台守护代理持续同步文档的项目
不使用 Claude Code 的项目、对文档行数无硬性限制要求的项目
也有自己的公开项目?先免登录查已有报告;没有报告时,可保留地址继续登录评测。
查找我项目的报告 →unsafe-install-commandREADME.md:91high confidencecurl -fsSL https://raw.githubusercontent.com/alirezarezvani/ClaudeForge/main/install.sh | bash修复:固定版本与校验和,先下载审查再执行,避免管道直接交给 Shell。
- 01固定版本与校验和,先下载审查再执行,避免管道直接交给 Shell。
- 02补充限制、权限或边界
方法、证据与局限展开收起
GitHub Repository API
6 个文件 · 44,848 字符
v3.10.0 · AI 内容复核:已使用(deepseek-chat)
- 静态评测不会安装或执行项目代码
- 安全扫描基于高信号文件与已知模式,不能替代人工审计
- 流行度只反映采用程度,不代表安全或工程质量
README
ClaudeForge
Automated CLAUDE.md creation, enhancement, and maintenance for Claude Code projects
ClaudeForge is a comprehensive toolkit that eliminates the tedious process of manually creating and maintaining CLAUDE.md files. With intelligent analysis, automated generation, and background maintenance, your CLAUDE.md files stay perfectly synchronized with your codebase.
🆕 What's New
- Installable Claude Code plugin — manifest at
.claude-plugin/plugin.json; install with/plugin marketplace add alirezarezvani/ClaudeForge && /plugin install claudeforge - Hard 150-line cap per CLAUDE.md — enforced deterministically by
hooks/hooks.jsononPostToolUse(Edit|Write)andInstructionsLoaded(everyload_reason); larger projects spread content across chained sub-files via@pathimports /sync-claude-md— walks every CLAUDE.md, prunes stale references, splits when over the cap, repairs root ↔ sub chains/sync-claude-md --weekly— orchestrates three forked task-style skills in parallel:claude-md-drift-audit,claude-md-link-check,claude-md-dependency-rescan- Karpathy behavioural guidelines auto-embedded in every generated CLAUDE.md and installed as a standalone
~/.claude/skills/karpathy-guidelines/skill scoped to code-file globs AGENTS.md/.cursorrules/.windsurfrulesinterop —/enhance-claude-mddetects sibling instruction files and chains them via@-imports instead of overwritingCLAUDE.local.mdpersonal tier — per-developer overrides exempt from the cap, gitignored automatically- Layered hook config —
hooks/hooks-config.json(committed defaults) +hooks/hooks-config.local.json(gitignored) lets developers opt out per machine - Lifecycle hooks:
SessionStart,PreToolUse,PostToolUse,InstructionsLoaded,Stop(one-line drift summary at session end) - Guardian agent runs
model: haikuwith a fail-closed contract (Skill-tool only, never auto-commits, aborts on missing validated output)
👉 Upgrading from v1.x? See docs/MIGRATION_V2.md.
✨ Features
- 🚀 Interactive Initialization - Explores your repository, detects project context, and creates customized CLAUDE.md files through conversational workflow
- ✅ Intelligent Analysis - Scans and evaluates existing CLAUDE.md files with quality scoring (0-100) and actionable recommendations
- 🔧 Smart Enhancement - Adds missing sections and improves structure automatically
- 🛡️ Background Maintenance - Guardian agent keeps CLAUDE.md synchronized with codebase changes
- 📦 Modular Architecture - Supports complex projects with context-specific files (backend/, frontend/, database/)
- 🎯 100% Native Format - All generated files follow official Claude Code format with project structure diagrams, setup instructions, and architecture sections
- 🛠️ Tech Stack Customization - Tailors guidelines to TypeScript, Python, Go, React, Vue, FastAPI, and more
- 👥 Team Size Adaptation - Adjusts complexity based on team size (solo, small, medium, large)
📦 What's Included
Skills
claudeforge-skill(skill/SKILL.md) — core analysis / validation / generation engine; runs onmodel: haiku. Scoped viapaths:toCLAUDE.md,CLAUDE.local.md,AGENTS.md,.cursorrules,.windsurfrules, and.claude/rules/*.mdso it auto-loads only when those files are touched.karpathy-guidelines(skill/karpathy-guidelines/SKILL.md) — Think Before Coding, Simplicity First, Surgical Changes, Goal-Driven Execution. Embedded into every generated CLAUDE.md and installed as a standalone skillpaths:-scoped to ~23 source-file extensions. Adapted with attribution from the MIT-licensed forrestchang/andrej-karpathy-skills.claude-md-drift-audit(skill/claude-md-drift-audit/SKILL.md, forked +agent: Explore) — walks the last N days of git history and flags every CLAUDE.md line that references deleted paths, renamed paths, or removed dependencies. Read-only./claude-md-drift-audit [days=7].claude-md-link-check(skill/claude-md-link-check/SKILL.md, forked +agent: Explore) — verifies every@pathchain import and every relative markdown link inside every CLAUDE.md resolves. Read-only./claude-md-link-check [path-glob].claude-md-dependency-rescan(skill/claude-md-dependency-rescan/SKILL.md, forked +agent: Explore) — diffs declared dependencies (package.json/requirements.txt/pyproject.toml/go.mod/Cargo.toml) against the Tech Stack section of every CLAUDE.md. Read-only./claude-md-dependency-rescan [manifest].
Slash commands
/enhance-claude-md(command/enhance-claude-md.md) — multi-phase init/enhance workflow withargument-hint,when_to_use,allowed-tools, anddisallowedTools(blocksWebFetch/WebSearch). Delegates deep codebase scans to the Explore subagent./sync-claude-md(command/sync-claude-md.md) — inventory → prune stale refs → enforce the 150-line cap → repair root ↔ sub chain. New--weeklyflag orchestrates the three audit skills in parallel before doing sync work./claude-to-agents(command/claude-to-agents.md) — convert the project's CLAUDE.md tree into anAGENTS.mdfor Codex / Gemini Code Assist / any tool honouring the AGENTS.md convention. Three modes:--symlink(one source of truth, default on macOS/Linux),--copy(snapshot),--inline-chain(flattens the@pathchain into one self-contained file — recommended for modular projects since Codex/Gemini don't auto-resolve@imports). Backs up an existing AGENTS.md before overwrite.
Agent
claude-md-guardian(agent/claude-md-guardian.md) — background maintenance. Runsmodel: haikuwith a fail-closed contract: Skill-tool only, aborts on missing validated output, never auto-commits, respects the local hook config.
Hooks
hooks/hooks.json— wiresPostToolUse(Write|Edit),InstructionsLoaded(all fiveload_reasonvalues), andStopevents to scripts underhooks/.hooks/validate-claude-md.py— enforces the 150-line cap deterministically at load time and write time; exempts*.local.md; exits with stderr feedback on violation.hooks/audit-claude-md.py—Stophook; one-line CLAUDE.md health summary printed at session end (total tracked / over cap / near cap).hooks/hooks-config.json+hooks/hooks-config.local.json— layered config; per-developer overrides without forking the shipped manifest.
🚀 Quick Start
Installation
Claude Code Plugin (recommended)
ClaudeForge ships as a Claude Code plugin. From any Claude Code session:
/plugin marketplace add alirezarezvani/ClaudeForge
/plugin install claudeforge
This installs every component (skills, slash commands, guardian agent) and registers /enhance-claude-md and /sync-claude-md for any project. Works the same at the user level (available everywhere) or scoped to a single project.
macOS / Linux
curl -fsSL https://raw.githubusercontent.com/alirezarezvani/ClaudeForge/main/install.sh | bash
Windows (PowerShell)
iwr https://raw.githubusercontent.com/alirezarezvani/ClaudeForge/main/install.ps1 -useb | iex
Manual Installation
git clone https://github.com/alirezarezvani/ClaudeForge.git
cd ClaudeForge
./install.sh # or .\install.ps1 on Windows
Usage
After installation, restart Claude Code and run:
/enhance-claude-md
That's it! The command will:
- ✅ Explore your project structure
- ✅ Detect tech stack and project type
- ✅ Show discoveries and ask for confirmation
- ✅ Create customized CLAUDE.md file(s)
- ✅ Apply best practices and native format
💡 Use Cases
New Project (No CLAUDE.md)
/enhance-claude-md
# Claude explores your repository and responds:
#
# "Based on my exploration, here's what I discovered:
#
# 📦 Project Type: Full-Stack Application
# 🛠️ Tech Stack: TypeScript, React, Node.js, PostgreSQL, Docker
# 👥 Team Size: Small (5 developers)
# 🚀 Development Phase: MVP
#
# 📋 Recommended Structure:
# - Root CLAUDE.md (navigation hub, ~100 lines)
# - backend/CLAUDE.md (API guidelines, ~150 lines)
# - frontend/CLAUDE.md (React guidelines, ~175 lines)
#
# Would you like me to create these files?"
#
# You: "Yes"
#
# ✅ CLAUDE.md created (108 lines)
# ✅ backend/CLAUDE.md created (156 lines)
# ✅ frontend/CLAUDE.md created (182 lines)
Existing Project (Needs Improvement)
/enhance-claude-md
# Claude analyzes your current CLAUDE.md:
#
# "Quality Score: 65/100
#
# Missing Sections:
# - Project Structure (ASCII diagram)
# - Setup & Installation
# - Architecture
#
# Would you like me to enhance your CLAUDE.md with these sections?"
#
# You: "Yes"
#
# ✅ CLAUDE.md enhanced (+2 sections, quality score: 65 → 88)
Background Maintenance
# You start a new Claude Code session
# Guardian agent automatically checks for changes
#
# ✅ CLAUDE.md updated:
# - Tech Stack: Added 2 dependencies (react-query, tailwindcss)
# - Project Structure: Updated diagram with new components/ directory
# - Setup & Installation: New environment variables
#
# Changes: 3 sections, 12 lines
📚 Documentation
| Document | Description |
|---|---|
| Quick Start Guide | 5-minute tutorial to get started |
| Installation Guide | Detailed installation instructions and troubleshooting |
| Architecture Overview | How components work together |
| GitHub Workflows | CI/CD automation and quality gates |
| Branching Strategy | Branch flow and protection rules |
| Troubleshooting | Common issues and solutions |
| Contributing Guide | How to contribute to ClaudeForge |
📖 Examples
See the examples/ directory for:
- Basic usage scenarios
- Modular architecture setup
- Integration with existing projects
- Advanced customization
🔧 Components Deep Dive
Skill: claudeforge-skill
Core Capabilities:
- Analysis - Scans existing CLAUDE.md files for quality and completeness
- Validation - Checks against Anthropic guidelines and best practices
- Generation - Creates new CLAUDE.md files from scratch
- Enhancement - Adds missing sections and improves structure
- Template Selection - Chooses appropriate templates based on project context
Quality Scoring (0-100):
- Length appropriateness (25 pts)
- Section completeness (25 pts)
- Formatting quality (20 pts)
- Content specificity (15 pts)
- Modular organization (15 pts)
Slash Command: /enhance-claude-md
Multi-phase workflow (Discovery → Analysis → Task). Discovery delegates the deep codebase walk to the Explore subagent so it doesn't bloat the calling session. Phase 1 also detects sibling AGENTS.md / .cursorrules / .windsurfrules and chains them via @-imports rather than overwriting. Phase 3 invokes claudeforge-skill via the Skill tool.
Slash Command: /sync-claude-md (with --weekly)
Default mode: inventory every CLAUDE.md, prune stale references, enforce the 150-line cap by splitting into sub-files, repair the root ↔ sub chain. With --weekly, Phase 0 issues the three forked audit skills (claude-md-drift-audit, claude-md-link-check, claude-md-dependency-rescan) in parallel via the Skill tool, aggregates their findings under ## Weekly Audit Summary, then proceeds. Each forked skill runs in an isolated subagent context (context: fork, agent: Explore) so audit work doesn't bloat the calling session.
Agent: claude-md-guardian
Runs model: haiku, fork_safe: true. Hook frontmatter uses Anthropic's canonical keyed-object schema (events: SessionStart, PreToolUse, PostToolUse, InstructionsLoaded). Fail-closed contract: invokes claudeforge-skill exclusively through the Skill tool (never inlines SKILL.md), aborts on missing validated output, never auto-commits, respects hooks/hooks-config.local.json.
🎯 Requirements
- Claude Code 2.0 or later
- Git (recommended for change detection)
- Operating Systems: macOS, Linux, Windows (PowerShell)
🤝 Contributing
We welcome contributions! Please see our Contributing Guide for details.
Quick Contribution Steps:
- Fork the repository
- Create a feature branch (
git checkout -b feature/amazing-feature) - Commit your changes (
git commit -m 'Add amazing feature') - Push to the branch (
git push origin feature/amazing-feature) - Open a Pull Request
🐛 Issues & Support
- Bug Reports: GitHub Issues
- Feature Requests: GitHub Discussions
- Documentation: docs/
📄 License
This project is licensed under the MIT License - see the LICENSE file for details.
Copyright © 2025 Alireza Rezvani
🙏 Acknowledgments
- Built for the Claude Code community.
- The behavioural-guardrail skill adapts the four principles from the MIT-licensed forrestchang/andrej-karpathy-skills (inspired by Andrej Karpathy's commentary on LLM coding pitfalls). Original prose; attribution preserved in
skill/karpathy-guidelines/SKILL.md. - Several plugin conventions (layered hook config,
Stopaudit hook, command discovery metadata,paths:scoping on skills, fail-closed contracts) are adapted from the MIT-licensed shanraisshan/claude-code-best-practice. Patterns implemented in original code with attribution inCHANGELOG.md. - Anthropic's Claude Code documentation drove the load-event integrations (
InstructionsLoaded, all fiveload_reasonmatchers) and thecontext: forktask-style skills.
🚦 Project Status
Version: 2.1.0 (see CHANGELOG.md)
Status: ✅ Stable & Production-Ready
Requires: Claude Code 2.1.4+ for hooks/InstructionsLoaded/paths: features
📊 Quick Stats
- 5 skills (
claudeforge-skill,karpathy-guidelines, plus three forked audit skills) - 3 slash commands (
/enhance-claude-md,/sync-claude-mdwith--weekly,/claude-to-agents) - 1 agent (
claude-md-guardian, fail-closed contract) - 3 hook scripts wired across
PostToolUse,InstructionsLoaded,Stop - 5 Python modules under
skill/(analyzer, validator, generator, template_selector, workflow) - 7 reference CLAUDE.md templates under
skill/examples/ - 150 — hard line cap per CLAUDE.md, enforced at load time and write time
🌟 Star History
If you find ClaudeForge helpful, please consider giving it a star on GitHub!
Made with ❤️ for the Claude Code community