[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-runesleo--claude-code-workflow":3,"tool-runesleo--claude-code-workflow":64},[4,17,27,35,43,56],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":16},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,3,"2026-04-05T11:01:52",[13,14,15],"开发框架","图像","Agent","ready",{"id":18,"name":19,"github_repo":20,"description_zh":21,"stars":22,"difficulty_score":23,"last_commit_at":24,"category_tags":25,"status":16},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",140436,2,"2026-04-05T23:32:43",[13,15,26],"语言模型",{"id":28,"name":29,"github_repo":30,"description_zh":31,"stars":32,"difficulty_score":23,"last_commit_at":33,"category_tags":34,"status":16},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",107662,"2026-04-03T11:11:01",[13,14,15],{"id":36,"name":37,"github_repo":38,"description_zh":39,"stars":40,"difficulty_score":23,"last_commit_at":41,"category_tags":42,"status":16},3704,"NextChat","ChatGPTNextWeb\u002FNextChat","NextChat 是一款轻量且极速的 AI 助手，旨在为用户提供流畅、跨平台的大模型交互体验。它完美解决了用户在多设备间切换时难以保持对话连续性，以及面对众多 AI 模型不知如何统一管理的痛点。无论是日常办公、学习辅助还是创意激发，NextChat 都能让用户随时随地通过网页、iOS、Android、Windows、MacOS 或 Linux 端无缝接入智能服务。\n\n这款工具非常适合普通用户、学生、职场人士以及需要私有化部署的企业团队使用。对于开发者而言，它也提供了便捷的自托管方案，支持一键部署到 Vercel 或 Zeabur 等平台。\n\nNextChat 的核心亮点在于其广泛的模型兼容性，原生支持 Claude、DeepSeek、GPT-4 及 Gemini Pro 等主流大模型，让用户在一个界面即可自由切换不同 AI 能力。此外，它还率先支持 MCP（Model Context Protocol）协议，增强了上下文处理能力。针对企业用户，NextChat 提供专业版解决方案，具备品牌定制、细粒度权限控制、内部知识库整合及安全审计等功能，满足公司对数据隐私和个性化管理的高标准要求。",87618,"2026-04-05T07:20:52",[13,26],{"id":44,"name":45,"github_repo":46,"description_zh":47,"stars":48,"difficulty_score":23,"last_commit_at":49,"category_tags":50,"status":16},2268,"ML-For-Beginners","microsoft\u002FML-For-Beginners","ML-For-Beginners 是由微软推出的一套系统化机器学习入门课程，旨在帮助零基础用户轻松掌握经典机器学习知识。这套课程将学习路径规划为 12 周，包含 26 节精炼课程和 52 道配套测验，内容涵盖从基础概念到实际应用的完整流程，有效解决了初学者面对庞大知识体系时无从下手、缺乏结构化指导的痛点。\n\n无论是希望转型的开发者、需要补充算法背景的研究人员，还是对人工智能充满好奇的普通爱好者，都能从中受益。课程不仅提供了清晰的理论讲解，还强调动手实践，让用户在循序渐进中建立扎实的技能基础。其独特的亮点在于强大的多语言支持，通过自动化机制提供了包括简体中文在内的 50 多种语言版本，极大地降低了全球不同背景用户的学习门槛。此外，项目采用开源协作模式，社区活跃且内容持续更新，确保学习者能获取前沿且准确的技术资讯。如果你正寻找一条清晰、友好且专业的机器学习入门之路，ML-For-Beginners 将是理想的起点。",84991,"2026-04-05T10:45:23",[14,51,52,53,15,54,26,13,55],"数据工具","视频","插件","其他","音频",{"id":57,"name":58,"github_repo":59,"description_zh":60,"stars":61,"difficulty_score":10,"last_commit_at":62,"category_tags":63,"status":16},3128,"ragflow","infiniflow\u002Fragflow","RAGFlow 是一款领先的开源检索增强生成（RAG）引擎，旨在为大语言模型构建更精准、可靠的上下文层。它巧妙地将前沿的 RAG 技术与智能体（Agent）能力相结合，不仅支持从各类文档中高效提取知识，还能让模型基于这些知识进行逻辑推理和任务执行。\n\n在大模型应用中，幻觉问题和知识滞后是常见痛点。RAGFlow 通过深度解析复杂文档结构（如表格、图表及混合排版），显著提升了信息检索的准确度，从而有效减少模型“胡编乱造”的现象，确保回答既有据可依又具备时效性。其内置的智能体机制更进一步，使系统不仅能回答问题，还能自主规划步骤解决复杂问题。\n\n这款工具特别适合开发者、企业技术团队以及 AI 研究人员使用。无论是希望快速搭建私有知识库问答系统，还是致力于探索大模型在垂直领域落地的创新者，都能从中受益。RAGFlow 提供了可视化的工作流编排界面和灵活的 API 接口，既降低了非算法背景用户的上手门槛，也满足了专业开发者对系统深度定制的需求。作为基于 Apache 2.0 协议开源的项目，它正成为连接通用大模型与行业专有知识之间的重要桥梁。",77062,"2026-04-04T04:44:48",[15,14,13,26,54],{"id":65,"github_repo":66,"name":67,"description_en":68,"description_zh":69,"ai_summary_zh":69,"readme_en":70,"readme_zh":71,"quickstart_zh":72,"use_case_zh":73,"hero_image_url":74,"owner_login":75,"owner_name":76,"owner_avatar_url":77,"owner_bio":78,"owner_company":79,"owner_location":79,"owner_email":79,"owner_twitter":80,"owner_website":81,"owner_url":82,"languages":79,"stars":83,"forks":84,"last_commit_at":85,"license":79,"difficulty_score":23,"env_os":86,"env_gpu":87,"env_ram":87,"env_deps":88,"category_tags":93,"github_topics":79,"view_count":10,"oss_zip_url":79,"oss_zip_packed_at":79,"status":16,"created_at":94,"updated_at":95,"faqs":96,"releases":106},738,"runesleo\u002Fclaude-code-workflow","claude-code-workflow","Battle-tested Claude Code workflow template — memory management, context engineering, and task routing from 3 months of daily usage","claude-code-workflow 是一套专为 Claude Code 打造的实战级工作流模板。它旨在解决大模型在编程辅助中常见的“失忆”问题——即缺乏结构化管理时，AI 容易在会话间遗忘上下文，导致效率低下。借助 claude-code-workflow，Claude Code 能转变为持久化、可自我进化的开发伙伴。\n\n此方案特别适合追求高效编码体验的开发者。其核心亮点在于独特的三层架构设计：常驻规则层确保基础行为一致，按需加载文档层节省 Token 成本，热数据层实时记录项目状态。此外，claude-code-workflow 实现了智能任务路由，根据需求分配至 Opus 或 Sonnet 等不同模型层级，并内置强制验证机制，要求 AI 在完成任务前必须运行测试并确认结果，有效减少幻觉和错误。系统还能自动保存进度，避免意外关闭导致的工作丢失。这不仅仅是一个配置模板，更是一套经过三个月日常多项目验证的生产级解决方案。","# Claude Code Workflow\n\nA battle-tested workflow template for Claude Code — memory management, context engineering, and task routing from 3 months of daily usage across multiple projects.\n\n**Not a tutorial. Not a toy config. A production workflow that actually ships.**\n\n## Why This Exists\n\nClaude Code is powerful out of the box, but without structure it becomes a smart assistant that forgets everything between sessions. This template turns it into a **persistent, self-improving development partner** that:\n\n- Remembers past mistakes and applies lessons automatically\n- Manages context across long sessions without drifting\n- Routes tasks to the right model tier (Opus\u002FSonnet\u002FHaiku\u002FCodex\u002FLocal)\n- Forces verification before claiming completion (no more \"should work now\")\n- Auto-saves progress so closing the window doesn't lose work\n\n## Architecture: Three Layers\n\n```\n┌─────────────────────────────────────────────────────────┐\n│  Layer 0: Auto-loaded Rules (always in context)         │\n│  ┌─────────────┐ ┌────────────┐ ┌───────────────┐     │\n│  │ behaviors.md │ │skill-      │ │memory-flush.md│     │\n│  │              │ │triggers.md │ │               │     │\n│  └─────────────┘ └────────────┘ └───────────────┘     │\n├─────────────────────────────────────────────────────────┤\n│  Layer 1: On-demand Docs (loaded when needed)           │\n│  agents.md · content-safety.md · task-routing.md        │\n│  behaviors-extended.md · scaffolding-checkpoint.md ...   │\n├─────────────────────────────────────────────────────────┤\n│  Layer 2: Hot Data (your working memory)                │\n│  today.md · projects.md · goals.md · active-tasks.json  │\n└─────────────────────────────────────────────────────────┘\n```\n\n**Why three layers?** Context window is expensive. Loading everything wastes tokens and degrades quality. This system loads rules always (~2K tokens), docs only when relevant (~1-3K each), and keeps your daily state hot for instant recall.\n\n## What's Inside\n\n```\nclaude-code-workflow\u002F\n├── CLAUDE.md                     # Entry point — Claude reads this first\n├── README.md                     # You are here\n│\n├── rules\u002F                        # Layer 0: Always loaded\n│   ├── behaviors.md              # Core behavior rules (debugging, commits, routing)\n│   ├── skill-triggers.md         # When to auto-invoke which skill\n│   └── memory-flush.md           # Auto-save triggers (never lose progress)\n│\n├── docs\u002F                         # Layer 1: On-demand reference\n│   ├── agents.md                 # Multi-model collaboration framework\n│   ├── behaviors-extended.md     # Extended rules (knowledge base, associations)\n│   ├── behaviors-reference.md    # Detailed operation guides\n│   ├── content-safety.md         # AI hallucination prevention system\n│   ├── scaffolding-checkpoint.md # \"Do you really need to self-host?\" checklist\n│   └── task-routing.md           # Model tier routing + cost comparison\n│\n├── memory\u002F                       # Layer 2: Your working state (templates)\n│   ├── today.md                  # Daily session log\n│   ├── projects.md               # Cross-project status overview\n│   ├── goals.md                  # Week\u002Fmonth\u002Fquarter goals\n│   └── active-tasks.json         # Cross-session task registry\n│\n├── skills\u002F                       # Reusable skill definitions\n│   ├── session-end\u002FSKILL.md              # Auto wrap-up: save progress + commit + record\n│   ├── verification-before-completion\u002FSKILL.md  # \"Run the test. Read the output. THEN claim.\"\n│   ├── systematic-debugging\u002FSKILL.md     # 5-phase debugging (recall → root cause → fix)\n│   ├── planning-with-files\u002FSKILL.md      # File-based planning for complex tasks\n│   └── experience-evolution\u002FSKILL.md     # Auto-accumulate project knowledge\n│\n├── agents\u002F                       # Custom agent definitions\n│   ├── pr-reviewer.md            # Code review agent\n│   ├── security-reviewer.md      # OWASP security scanning agent\n│   └── performance-analyzer.md   # Performance bottleneck analysis agent\n│\n└── commands\u002F                     # Custom slash commands\n    ├── debug.md                  # \u002Fdebug — Start systematic debugging\n    ├── deploy.md                 # \u002Fdeploy — Pre-deployment checklist\n    ├── exploration.md            # \u002Fexploration — CTO challenge before coding\n    └── review.md                 # \u002Freview — Prepare code review\n```\n\n## Quick Start\n\n### 1. Copy to your Claude Code config\n\n```bash\n# Clone the template\ngit clone https:\u002F\u002Fgithub.com\u002Frunesleo\u002Fclaude-code-workflow.git\n\n# Copy to your Claude Code config directory\ncp -r claude-code-workflow\u002F* ~\u002F.claude\u002F\n\n# Or symlink if you want to keep it as a git repo\nln -sf ~\u002Fclaude-code-workflow\u002Frules ~\u002F.claude\u002Frules\nln -sf ~\u002Fclaude-code-workflow\u002Fdocs ~\u002F.claude\u002Fdocs\n# ... etc\n```\n\n### 2. Customize CLAUDE.md\n\nOpen `~\u002F.claude\u002FCLAUDE.md` and fill in:\n\n- **User Info**: Your name, project directory, social handles\n- **Sub-project Memory Routes**: Map your projects to memory paths\n- **SSOT Ownership Table**: Define where each type of info lives\n- **On-demand Loading Index**: Adjust doc paths if needed\n\n### 3. Start a session\n\n```bash\nclaude\n```\n\nClaude will automatically load your rules and start following the workflow. Try:\n\n- Start coding and notice the **task routing** (\"🔀 Route: bug fix → Sonnet\")\n- Hit a bug and watch **systematic debugging** kick in\n- Say \"that's all for now\" and see **session-end** auto-save everything\n- Come back tomorrow and find your context preserved in `today.md`\n\n## Key Concepts\n\n### SSOT (Single Source of Truth)\n\nEvery piece of information has ONE canonical location. The SSOT table in CLAUDE.md maps info types to files. Claude is trained to check SSOT before writing, preventing the \"same info in 5 places, all outdated\" problem.\n\n### Memory Flush\n\nClaude auto-saves progress on every task completion, every commit, and every exit signal. You can close the window mid-sentence and nothing is lost. No more \"I forgot to save my context.\"\n\n### Verification Before Completion\n\nThe most impactful rule: Claude cannot claim work is done without running the verification command and reading the output. Eliminates the #1 AI coding failure mode — \"should work now\" without actually checking.\n\n### Three-Tier Task Routing\n\nNot every task needs Opus. The routing system automatically matches task complexity to model tier:\n- **Opus**: Critical logic, security-sensitive, complex reasoning\n- **Sonnet**: Daily development, analysis, most coding tasks\n- **Haiku**: Simple queries, subagent tasks, quick lookups\n- **Codex**: Cross-verification, code review, second opinions\n- **Local**: Commit messages, formatting, offline work\n\n### Sunday Rule\n\nSystem optimization happens on Sundays. On other days, if you try to tweak your workflow instead of shipping, Claude will intercept and remind you to focus on output. Configurable to any cadence you prefer.\n\n## Customization Guide\n\n### Adding a new project\n\n1. Add to `memory\u002Fprojects.md`\n2. Add memory route in CLAUDE.md's \"Sub-project Memory Routes\"\n3. Create `PROJECT_CONTEXT.md` in the project root\n\n### Adding a new skill\n\nCreate `skills\u002Fyour-skill\u002FSKILL.md` with:\n\n```yaml\n---\nname: your-skill\ndescription: What it does\nallowed-tools:\n  - Read\n  - Write\n  - Bash\n---\n\n# Your Skill\n\n[Instructions for Claude when this skill is invoked]\n```\n\n### Adding a new agent\n\nCreate `agents\u002Fyour-agent.md` with:\n\n```yaml\n---\nname: your-agent\ndescription: What it does\ntools: Read, Grep, Glob, Bash\n---\n\n# Your Agent\n\n[Agent personality, review dimensions, output format]\n```\n\n### Adjusting model routing\n\nEdit `rules\u002Fbehaviors.md` → \"Task Routing\" section, and `docs\u002Ftask-routing.md` for detailed tier definitions.\n\n## Philosophy\n\nThis template encodes several principles learned from daily AI-assisted development:\n\n1. **Structure > Prompting**: A well-organized config file beats clever one-off prompts every time.\n2. **Memory > Intelligence**: An AI that remembers your past mistakes is more valuable than a smarter AI that starts fresh each session.\n3. **Verification > Confidence**: The cost of running `npm test` is always less than the cost of shipping a broken build.\n4. **Layered Loading > Flat Config**: Don't dump everything into context. Load rules always, docs on demand, data when needed.\n5. **Auto-save > Manual Save**: If it requires the user to remember, it will be forgotten. Make it automatic.\n\n## Requirements\n\n- [Claude Code](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fclaude-code) CLI (Claude Max or API subscription)\n- Optional: Codex CLI for cross-verification\n- Optional: Ollama for local model fallback\n\n## Prior Art & Credits\n\nThis template draws from:\n- [Manus](https:\u002F\u002Fmanus.im\u002F) file-based planning approach\n- OWASP Top 10 for security review patterns\n- Real-world experience from building [x-reader](https:\u002F\u002Fgithub.com\u002Frunesleo\u002Fx-reader) (650+ stars) and other open-source projects\n\n## Star History\n\n[![Star History Chart](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Frunesleo_claude-code-workflow_readme_ddcb6ef7fcfe.png)](https:\u002F\u002Fstar-history.com\u002F#runesleo\u002Fclaude-code-workflow&Date)\n\n## License\n\nMIT — Use it, fork it, make it yours.\n\n---\n\nBuilt by [@runes_leo](https:\u002F\u002Fx.com\u002Frunes_leo) — more AI tools at [leolabs.me](https:\u002F\u002Fleolabs.me) — [Telegram Community](https:\u002F\u002Ft.me\u002Frunesgang)\n","# Claude Code 工作流\n\n一个经过实战检验的 Claude Code 工作流模板 —— 基于跨多个项目的 3 个月日常使用经验，涵盖记忆管理、上下文工程和任务路由。\n\n**不是教程。不是玩具配置。是一个真正能交付的生产级工作流。**\n\n## 为何存在\n\nClaude Code 开箱即用功能强大，但如果没有结构支撑，它会变成一个在会话之间遗忘一切的智能助手。此模板将其转变为一个**持久、自我改进的开发伙伴**，能够：\n\n- 记住过去的错误并自动应用经验教训\n- 在长会话中管理上下文而不会偏离目标\n- 将任务路由到正确的模型层级（Opus\u002FSonnet\u002FHaiku\u002FCodex\u002FLocal）\n- 在声称完成前强制进行验证（不再出现“现在应该可以了”的情况）\n- 自动保存进度，关闭窗口不会丢失工作\n\n## 架构：三层结构\n\n```\n┌─────────────────────────────────────────────────────────┐\n│  Layer 0: Auto-loaded Rules (always in context)         │\n│  ┌─────────────┐ ┌────────────┐ ┌───────────────┐     │\n│  │ behaviors.md │ │skill-      │ │memory-flush.md│     │\n│  │              │ │triggers.md │ │               │     │\n│  └─────────────┘ └────────────┘ └───────────────┘     │\n├─────────────────────────────────────────────────────────┤\n│  Layer 1: On-demand Docs (loaded when needed)           │\n│  agents.md · content-safety.md · task-routing.md        │\n│  behaviors-extended.md · scaffolding-checkpoint.md ...   │\n├─────────────────────────────────────────────────────────┤\n│  Layer 2: Hot Data (your working memory)                │\n│  today.md · projects.md · goals.md · active-tasks.json  │\n└─────────────────────────────────────────────────────────┘\n```\n\n**为何分为三层？** 上下文窗口 (Context window) 成本高昂。加载所有内容会浪费 Token（令牌）并降低质量。该系统始终加载规则（约 2K Token），仅在相关时加载文档（每个约 1-3K），并保持您的每日状态热数据以便即时调用。\n\n## 内部结构\n\n```\nclaude-code-workflow\u002F\n├── CLAUDE.md                     # Entry point — Claude reads this first\n├── README.md                     # You are here\n│\n├── rules\u002F                        # Layer 0: Always loaded\n│   ├── behaviors.md              # Core behavior rules (debugging, commits, routing)\n│   ├── skill-triggers.md         # When to auto-invoke which skill\n│   └── memory-flush.md           # Auto-save triggers (never lose progress)\n│\n├── docs\u002F                         # Layer 1: On-demand reference\n│   ├── agents.md                 # Multi-model collaboration framework\n│   ├── behaviors-extended.md     # Extended rules (knowledge base, associations)\n│   ├── behaviors-reference.md    # Detailed operation guides\n│   ├── content-safety.md         # AI hallucination prevention system\n│   ├── scaffolding-checkpoint.md # \"Do you really need to self-host?\" checklist\n│   └── task-routing.md           # Model tier routing + cost comparison\n│\n├── memory\u002F                       # Layer 2: Your working state (templates)\n│   ├── today.md                  # Daily session log\n│   ├── projects.md               # Cross-project status overview\n│   ├── goals.md                  # Week\u002Fmonth\u002Fquarter goals\n│   └── active-tasks.json         # Cross-session task registry\n│\n├── skills\u002F                       # Reusable skill definitions\n│   ├── session-end\u002FSKILL.md              # Auto wrap-up: save progress + commit + record\n│   ├── verification-before-completion\u002FSKILL.md  # \"Run the test. Read the output. THEN claim.\"\n│   ├── systematic-debugging\u002FSKILL.md     # 5-phase debugging (recall → root cause → fix)\n│   ├── planning-with-files\u002FSKILL.md      # File-based planning for complex tasks\n│   └── experience-evolution\u002FSKILL.md     # Auto-accumulate project knowledge\n│\n├── agents\u002F                       # Custom agent definitions\n│   ├── pr-reviewer.md            # Code review agent\n│   ├── security-reviewer.md      # OWASP security scanning agent\n│   └── performance-analyzer.md   # Performance bottleneck analysis agent\n│\n└── commands\u002F                     # Custom slash commands\n    ├── debug.md                  # \u002Fdebug — Start systematic debugging\n    ├── deploy.md                 # \u002Fdeploy — Pre-deployment checklist\n    ├── exploration.md            # \u002Fexploration — CTO challenge before coding\n    └── review.md                 # \u002Freview — Prepare code review\n```\n\n## 快速开始\n\n### 1. 复制到您的 Claude Code 配置目录\n\n```bash\n# Clone the template\ngit clone https:\u002F\u002Fgithub.com\u002Frunesleo\u002Fclaude-code-workflow.git\n\n# Copy to your Claude Code config directory\ncp -r claude-code-workflow\u002F* ~\u002F.claude\u002F\n\n# Or symlink if you want to keep it as a git repo\nln -sf ~\u002Fclaude-code-workflow\u002Frules ~\u002F.claude\u002Frules\nln -sf ~\u002Fclaude-code-workflow\u002Fdocs ~\u002F.claude\u002Fdocs\n# ... etc\n```\n\n### 2. 自定义 CLAUDE.md\n\n打开 `~\u002F.claude\u002FCLAUDE.md` 并填写以下内容：\n\n- **用户信息**：您的姓名、项目目录、社交账号\n- **子项目记忆路由**：将您的项目映射到记忆路径\n- **SSOT 所有权表**：定义每种类型信息的存储位置\n- **按需加载索引**：根据需要调整文档路径\n\n### 3. 开始会话\n\n```bash\nclaude\n```\n\nClaude 将自动加载您的规则并开始遵循工作流。尝试：\n\n- 开始编码并观察**任务路由**（\"🔀 Route: bug fix → Sonnet\"）\n- 遇到 Bug 并观察**系统性调试**机制启动\n- 说\"that's all for now\"并查看**会话结束**自动保存所有内容\n- 明天回来并在 `today.md` 中找到保留的上下文\n\n## 核心概念\n\n### SSOT（单一事实来源）\n\n每条信息都有一个唯一的规范位置。CLAUDE.md 中的 SSOT 表将信息类型映射到文件。Claude 被训练为在写入前检查 SSOT，防止“同一条信息出现在 5 个地方且全部过时”的问题。\n\n### 记忆刷新 (Memory Flush)\n\nClaude 在每个任务完成、每次提交和每次退出信号时自动保存进度。您可以在句子中间关闭窗口，内容不会丢失。不再需要担心“我忘了保存我的上下文”。\n\n### 完成前验证\n\n最具影响力的规则：Claude 无法声称工作已完成，除非运行验证命令并读取输出结果。消除了 AI 编码最常见的失败模式 —— “现在应该可以了”但实际上并未检查。\n\n### 三层任务路由\n\n并非每个任务都需要 Opus。路由系统自动将任务复杂度匹配到模型层级：\n- **Opus**：关键逻辑、安全敏感、复杂推理\n- **Sonnet**：日常开发、分析、大多数编码任务\n- **Haiku**：简单查询、子代理任务、快速查找\n- **Codex**：交叉验证、代码审查、第二意见\n- **Local**：提交消息、格式化、离线工作\n\n### 周日规则\n\n系统优化安排在周日进行。在其他日子里，如果您试图调整工作流而不是交付成果，Claude 将拦截并提醒您专注于产出。可根据您喜欢的任何频率进行配置。\n\n## 自定义指南\n\n### 添加新项目\n\n1. 添加到 `memory\u002Fprojects.md`\n2. 在 `CLAUDE.md` 的“子项目记忆路由”中添加记忆路由\n3. 在项目根目录创建 `PROJECT_CONTEXT.md`\n\n### 添加新技能\n\n使用以下内容创建 `skills\u002Fyour-skill\u002FSKILL.md`：\n\n```yaml\n---\nname: your-skill\ndescription: What it does\nallowed-tools:\n  - Read\n  - Write\n  - Bash\n---\n\n# Your Skill\n\n[Instructions for Claude when this skill is invoked]\n```\n\n### 添加新智能体\n\n使用以下内容创建 `agents\u002Fyour-agent.md`：\n\n```yaml\n---\nname: your-agent\ndescription: What it does\ntools: Read, Grep, Glob, Bash\n---\n\n# Your Agent\n\n[Agent personality, review dimensions, output format]\n```\n\n### 调整模型路由\n\n编辑 `rules\u002Fbehaviors.md` 中的“任务路由”部分，以及 `docs\u002Ftask-routing.md` 以获取详细的层级定义。\n\n## 理念\n\n此模板编码了从日常 AI 辅助开发中学到的几个原则：\n\n1. **结构 > 提示**：一个组织良好的配置文件每次都胜过巧妙的一次性提示。\n2. **记忆 > 智能**：一个能记住你过去错误的 AI，比一个每次会话都从头开始的更聪明的 AI 更有价值。\n3. **验证 > 信心**：运行 `npm test` 的成本总是低于发布损坏构建的成本。\n4. **分层加载 > 扁平配置**：不要把所有东西都倒入上下文。规则始终加载，文档按需加载，数据需要时加载。\n5. **自动保存 > 手动保存**：如果需要用户去记住，它就会被遗忘。让它自动化。\n\n## 要求\n\n- [Claude Code](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fclaude-code) CLI (命令行界面) (Claude Max 或 API (应用程序接口) 订阅)\n- 可选：Codex CLI 用于交叉验证\n- 可选：Ollama 用于本地模型回退\n\n## 相关作品与致谢\n\n此模板借鉴了：\n- [Manus](https:\u002F\u002Fmanus.im\u002F) 基于文件的规划方法\n- OWASP Top 10 用于安全审查模式\n- 来自构建 [x-reader](https:\u002F\u002Fgithub.com\u002Frunesleo\u002Fx-reader) (650+ 星标) 和其他开源项目的实际经验\n\n## 星标历史\n\n[![Star History Chart](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Frunesleo_claude-code-workflow_readme_ddcb6ef7fcfe.png)](https:\u002F\u002Fstar-history.com\u002F#runesleo\u002Fclaude-code-workflow&Date)\n\n## 许可证\n\nMIT — 使用它，分叉它，让它成为你的。\n\n---\n\n由 [@runes_leo](https:\u002F\u002Fx.com\u002Frunes_leo) 构建 — 更多 AI 工具请访问 [leolabs.me](https:\u002F\u002Fleolabs.me) — [Telegram 社区](https:\u002F\u002Ft.me\u002Frunesgang)","# claude-code-workflow 快速上手指南\n\n`claude-code-workflow` 是一个经过实战验证的 Claude Code 工作流模板。它通过结构化的记忆管理、上下文工程和任务路由，将 Claude 转化为一个持久化、可自我改进的开发伙伴。\n\n## 1. 环境准备\n\n在开始之前，请确保您的开发环境满足以下要求：\n\n- **操作系统**: macOS \u002F Linux \u002F Windows (WSL)\n- **工具依赖**:\n  - [Git](https:\u002F\u002Fgit-scm.com\u002F) (用于克隆仓库)\n  - [Claude Code CLI](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fclaude-code) (需 Anthropic API 订阅或 Claude Max 账号)\n- **可选组件**:\n  - Codex CLI (用于交叉验证)\n  - Ollama (用于本地模型 fallback)\n\n## 2. 安装步骤\n\n### 2.1 克隆工作流模板\n\n从 GitHub 克隆项目到本地目录：\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Frunesleo\u002Fclaude-code-workflow.git\ncd claude-code-workflow\n```\n\n### 2.2 部署到配置目录\n\n将文件复制到您的 Claude Code 配置目录（默认为 `~\u002F.claude\u002F`）。\n\n**方案 A：直接复制（推荐）**\n```bash\n# 确保目标目录存在\nmkdir -p ~\u002F.claude\n\n# 复制所有文件\ncp -r .\u002F* ~\u002F.claude\u002F\n```\n\n**方案 B：符号链接（保持版本控制）**\n如果您希望保留 Git 历史并作为子模块更新：\n```bash\nln -sf ~\u002Fclaude-code-workflow\u002Frules ~\u002F.claude\u002Frules\nln -sf ~\u002Fclaude-code-workflow\u002Fdocs ~\u002F.claude\u002Fdocs\nln -sf ~\u002Fclaude-code-workflow\u002Fmemory ~\u002F.claude\u002Fmemory\nln -sf ~\u002Fclaude-code-workflow\u002Fskills ~\u002F.claude\u002Fskills\nln -sf ~\u002Fclaude-code-workflow\u002Fagents ~\u002F.claude\u002Fagents\nln -sf ~\u002Fclaude-code-workflow\u002Fcommands ~\u002F.claude\u002Fcommands\n```\n\n## 3. 基本使用\n\n### 3.1 初始化配置\n\n首次使用前，需要编辑入口配置文件以适配您的个人和项目信息：\n\n```bash\nnano ~\u002F.claude\u002FCLAUDE.md\n# 或使用您喜欢的编辑器\ncode ~\u002F.claude\u002FCLAUDE.md\n```\n\n请在文件中填写以下内容：\n- **User Info**: 您的姓名、项目目录、社交账号。\n- **Sub-project Memory Routes**: 将您的项目映射到记忆路径。\n- **SSOT Ownership Table**: 定义各类信息的唯一来源位置。\n- **On-demand Loading Index**: 根据需要调整文档加载路径。\n\n### 3.2 启动会话\n\n在终端中运行 Claude Code：\n\n```bash\nclaude\n```\n\n系统会自动加载规则并开始遵循工作流。您可以尝试以下操作来体验核心功能：\n\n- **查看任务路由**: 开始编码时，注意观察输出中的路由提示（例如 `🔀 Route: bug fix → Sonnet`）。\n- **系统调试**: 遇到 Bug 时输入 `\u002Fdebug` 触发系统化调试流程。\n- **自动保存**: 结束会话时说 \"that's all for now\"，系统会自动保存进度并提交代码。\n- **恢复上下文**: 第二天重新打开会话，之前的上下文会保存在 `today.md` 中。\n\n### 3.3 常用指令\n\n工作流内置了多个 Slash Commands 供快速调用：\n\n| 指令 | 说明 |\n| :--- | :--- |\n| `\u002Fdebug` | 启动系统化调试流程 |\n| `\u002Fdeploy` | 执行部署前检查清单 |\n| `\u002Fexploration` | 编码前的架构挑战与探索 |\n| `\u002Freview` | 准备代码审查 |","后端工程师小李负责重构一个遗留微服务系统，涉及跨模块数据迁移与长达数周的迭代任务。\n\n### 没有 claude-code-workflow 时\n- 每次重启会话都要重复解释项目背景，AI 经常遗忘之前的架构决策。\n- 长对话中上下文窗口耗尽，导致后续代码建议逐渐偏离实际需求。\n- 修复 Bug 后未经验证就提交，上线后才发现隐藏的逻辑漏洞。\n- 意外关闭终端前未保存进度，数小时的调试成果瞬间付诸东流。\n\n### 使用 claude-code-workflow 后\n- 自动沉淀历史错误与经验，新会话直接继承上下文记忆，无需重复输入。\n- 三层架构管理信息，关键规则常驻且按需加载文档，保持上下文精准高效。\n- 内置强制验证技能确保代码运行测试通过后才标记完成，杜绝“应该能跑”的幻觉。\n- 会话结束时自动保存状态并提交，随时中断都能无缝恢复工作进度。\n\nclaude-code-workflow 将临时的聊天助手转变为可信赖的持久化开发伙伴，显著提升复杂项目的交付质量。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Frunesleo_claude-code-workflow_2ca845ec.png","runesleo","Leo","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Frunesleo_23dc64fe.jpg","Building AI agents & automation for real-world workflows. Shipping tools, not demos. @runes_leo on X.",null,"runes_leo","https:\u002F\u002Fleolabs.me","https:\u002F\u002Fgithub.com\u002Frunesleo",536,65,"2026-04-05T21:10:08","Linux, macOS","未说明",{"notes":89,"python":87,"dependencies":90},"此工具为 Claude Code 的配置文件模板，非独立可执行程序。核心依赖为 Anthropic 官方 CLI 及网络 API。本地运行仅需基础 Shell 环境，无 Python 或深度学习库依赖。仅在可选使用 Ollama 本地模型时才需关注硬件配置。",[91,92],"Git","Claude Code CLI",[15,13],"2026-03-27T02:49:30.150509","2026-04-06T09:44:28.396923",[97,102],{"id":98,"question_zh":99,"answer_zh":100,"source_url":101},3132,"如何在 README 中标记项目已被 awesome-claude-code-workflows 收录？","可以在项目的 README 文件中添加以下 Markdown 代码块来显示徽章：\n```markdown\n[![Mentioned in awesome-claude-code-workflows](https:\u002F\u002Fawesome.re\u002Fmentioned-badge.svg)](https:\u002F\u002Fgithub.com\u002Fithiria894\u002Fawesome-claude-code-workflows)\n```\n","https:\u002F\u002Fgithub.com\u002Frunesleo\u002Fclaude-code-workflow\u002Fissues\u002F2",{"id":103,"question_zh":104,"answer_zh":105,"source_url":101},3133,"如果列表中的项目详情不准确，该如何更新？","如果发现收录信息有误或需要更新描述，可以直接前往 awesome-claude-code-workflows 仓库 [提交 PR](https:\u002F\u002Fgithub.com\u002Fithiria894\u002Fawesome-claude-code-workflows\u002Fpulls) 进行修改。",[]]