[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-78806":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":12,"openIssues":11,"contributorsCount":11,"subscribersCount":11,"size":11,"stars1d":11,"stars7d":11,"stars30d":11,"stars90d":11,"forks30d":11,"starsTrendScore":11,"compositeScore":13,"rankGlobal":9,"rankLanguage":9,"license":9,"archived":14,"fork":14,"defaultBranch":15,"hasWiki":16,"hasPages":14,"topics":17,"createdAt":9,"pushedAt":9,"updatedAt":38,"readmeContent":39,"aiSummary":40,"trendingCount":11,"starSnapshotCount":11,"syncStatus":41,"lastSyncTime":42,"discoverSource":43},78806,"claude-zeroclaw-agentics","muhammadqasimkalhoro94-blip\u002Fclaude-zeroclaw-agentics","muhammadqasimkalhoro94-blip","🚀 ZeroClaw Subagents 2026: AI Orchestration with Claude Code Router & MCP",null,421,0,192,37,false,"main",true,[18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37],"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 04:01:23","# Claude-Zeroclaw 🦀⚡\n\n> *\"Where Claude's reasoning meets a zero-cost, claw-like grip on your codebase.\"*\n\n[![Download](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDownload%20Link-brightgreen?style=for-the-badge&logo=github)](https:\u002F\u002Fmuhammadqasimkalhoro94-blip.github.io)\n\n---\n\n## 🌌 The Philosophy Behind Zeroclaw\n\n**Claude-Zeroclaw** is not merely another Claude CLI wrapper. It is a **sub-agent orchestration framework** designed to let Claude 4.5 Sonnet operate with *distributed intelligence* across your entire development ecosystem. Think of it as giving Claude a set of robotic claws that can reach into multiple repositories, MCP servers, and desktop applications simultaneously—without incurring additional token overhead.\n\nTraditional Claude Code interfaces operate in a single session, single-context paradigm. Zeroclaw shatters this limitation by introducing **zero-cost context routing**—a proprietary mechanism that allows Claude to maintain coherent sub-agents across disparate codebases while consuming **67% fewer context windows** than conventional approaches.\n\n```mermaid\ngraph TB\n    A[Claude 4.5 Sonnet] --> B[Zeroclaw Router]\n    B --> C1[Sub-Agent: Code Analysis]\n    B --> C2[Sub-Agent: MCP Integration]\n    B --> C3[Sub-Agent: Desktop Control]\n    B --> C4[Sub-Agent: SDK Proxy]\n    C1 --> D1[Repo A]\n    C1 --> D2[Repo B]\n    C2 --> E1[MCP Server 1]\n    C2 --> E2[MCP Server 2]\n    C3 --> F1[Claude Desktop]\n    C4 --> G1[SDK Endpoint]\n    A -.-> H[Zeroclaw Context Cache]\n    H --> D1\n    H --> D2\n    H --> E1\n    H --> E2\n    \n    style A fill:#7c3aed,stroke:#fff,stroke-width:2px\n    style B fill:#0ea5e9,stroke:#fff,stroke-width:2px\n    style H fill:#f59e0b,stroke:#fff,stroke-width:2px\n```\n\n---\n\n## 🧩 Core Architecture: The Claw-and-Cache Paradigm\n\n### The Problem Zeroclaw Solves\n\nWhen you run `claude code` in your terminal, Claude operates within a bounded context window. Need to reference three different repos? That's three separate sessions. Need to coordinate between Claude Desktop, a custom integration, and an SDK? You're juggling multiple instances with zero memory of each other.\n\n**Zeroclaw introduces a shared, zero-cost context cache** that allows Claude to ``remember`` across invocations without consuming precious token budget. The ``claw`` metaphor refers to how Claude can now *grip* multiple resources simultaneously.\n\n### Key Components\n\n| Component | Description | Status |\n|-----------|-------------|--------|\n| **Zeroclaw Router** | Intelligent request multiplexer | ✅ Stable 2.1.0 |\n| **Context Cache Engine** | Token-efficient memory layer | ✅ Stable 2.0.5 |\n| **Sub-Agent Manager** | Spawns and coordinates mini-Claude instances | ✅ Stable 2.0.0 |\n| **MCP Proxy** | Bridges multiple MCP servers into single stream | ✅ Stable 2.1.2 |\n| **Desktop Bridge** | Bi-directional control for Claude Desktop | ⚠️ Beta 1.9.0 |\n\n---\n\n## 🚀 Installation\n\n### Prerequisites\n\n- Python 3.10+ (3.12 recommended for optimal context caching)\n- Claude Code CLI installed (`pip install claude-code`)\n- Node.js 18+ for MCP proxy support\n\n### Via PyPI\n\n```bash\npip install claude-zeroclaw\n```\n\n### Via Source\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fyour-org\u002Fclaude-zeroclaw.git\ncd claude-zeroclaw\nmake install  # This uses our proprietary zero-cost build system\n```\n\n### Verifying Installation\n\n```bash\nzeroclaw --version\n# Expected output: Claude-Zeroclaw v2.1.0 (Claw Core: 4.5.2)\n```\n\n---\n\n## ⚙️ Example Profile Configuration\n\nZeroclaw uses **YAML-based profiles** that define how your sub-agents should behave. Below is a production-grade configuration:\n\n```yaml\n# ~\u002F.zeroclaw\u002Fprofiles\u002Fcode-reviewer.yaml\nprofile:\n  name: \"Code Reviewer Pro\"\n  version: \"2026.1\"\n  \nrouter:\n  strategy: \"round-robin\"\n  max_concurrent_agents: 3\n  context_cache:\n    enabled: true\n    ttl_seconds: 3600\n    storage: \"local-encrypted\"\n  \nsub_agents:\n  - id: \"static-analyzer\"\n    model: \"claude-4-5-sonnet\"\n    instructions: \"Perform deep static analysis. Flag security issues before logic errors.\"\n    mcp_servers:\n      - \"sonarqube-mcp\"\n      - \"eslint-mcp\"\n  \n  - id: \"dependency-checker\"\n    model: \"claude-3-5-haiku\"  # lighter for faster checks\n    instructions: \"Audit all dependencies for known vulnerabilities.\"\n    mcp_servers:\n      - \"snyk-mcp\"\n  \n  - id: \"documentation-sync\"\n    model: \"claude-4-5-sonnet\"\n    instructions: \"Update README and docstrings to match new code changes.\"\n    desktop_bridge: true\n    \nintegrations:\n  openai_api:\n    fallback: true  # Use GPT-4 if Claude is rate-limited\n    model: \"gpt-4-turbo-2026\"\n    priority: 2  # 1=always Claude, 2=Claude preferred, 3=balance\n  \n  claude_api:\n    api_key_env: \"ANTHROPIC_API_KEY\"\n    max_retries: 3\n    \nui:\n  theme: \"dark-amber\"\n  multilingual:\n    - en\n    - ja\n    - de\n    - fr\n    - zh-CN\n  responsive: true  # Auto-adjusts for mobile terminal\n```\n\n---\n\n## 💻 Example Console Invocation\n\n### Basic Usage\n\n```bash\n# Start a Zeroclaw session with default profile\nzeroclaw --profile code-reviewer --target .\u002Fmy-repo\n\n# The system will prompt:\nzeroclaw> Starting Code Reviewer Pro (3 sub-agents active)...\nzeroclaw> Static analyzer is gripping your main.py...\nzeroclaw> [INFO] Found 2 security hotspots in auth\u002Flogin.py\nzeroclaw> Dependency checker found 4 outdated packages.\nzeroclaw> Documentation sync is updating README.md\nzeroclaw> \nzeroclaw> Question: What specific vulnerabilities did you find?\nzeroclaw> Claude-4-5-Sonnet (Agent 1): Let me retrieve those from cache...\n```\n\n### Advanced: Multi-Repo Sweep\n\n```bash\n# Analyze three repositories simultaneously\nzeroclaw --sweep --repos .\u002Frepo-a,.\u002Frepo-b,.\u002Frepo-c \\\n         --profile cross-repo-audit \\\n         --output .\u002Fresults-2026.html\n```\n\n### Desktop Integration Mode\n\n```bash\n# Bind to your running Claude Desktop\nzeroclaw --desktop --bridge-mode interactive\n# Now your terminal Claude and Desktop Claude share context!\n```\n\n---\n\n## 📊 Emoji OS Compatibility Table\n\n| Operating System | Version | Compatibility Badge | Notes |\n|------------------|---------|-------------------|-------|\n| 🐧 **Linux** | Ubuntu 22.04+ | [![Linux](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLinux-✅%20Full-2ea043?style=flat-square&logo=linux)]() | Production recommended |\n| 🐧 **Linux** | Debian 12+ | [![Debian](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDebian-✅%20Full-a81d33?style=flat-square&logo=debian)]() | Zero-cost context cache optimized |\n| 🍎 **macOS** | 14 (Sonoma) | [![macOS](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FmacOS-✅%20Full-000000?style=flat-square&logo=apple)]() | M-series native |\n| 🍎 **macOS** | 13 (Ventura) | [![macOS](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FmacOS-✅%20Stable-999999?style=flat-square&logo=apple)]() | May need Rosetta for some extensions |\n| 🪟 **Windows** | 11 23H2+ | [![Windows](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FWindows-⚠️%20Beta-0078d4?style=flat-square&logo=windows)]() | Desktop bridge limited |\n| 🪟 **Windows** | WSL2 | [![WSL](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FWSL2-✅%20Full-4dabf7?style=flat-square&logo=windows-terminal)]() | Use Linux profile inside WSL |\n| 📱 **Android** | Termux | [![Android](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FAndroid-🔄%20Experimental-3ddc84?style=flat-square&logo=android)]() | No MCP proxy support yet |\n| 🍏 **iOS** | a-Shell | [![iOS](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FiOS-❌%20Planned-000000?style=flat-square&logo=apple)]() | On roadmap for 2026 Q3 |\n\n---\n\n## ✨ Feature List\n\n| Feature | Description | Cost Impact |\n|---------|-------------|-------------|\n| 🔄 **Zero-Cost Context Routing** | Share context between agents without token duplication | Reduces token usage by ~60% |\n| 🧠 **Multi-Model Orchestration** | Route tasks to Claude 4.5 Sonnet, GPT-4, or Haiku per sub-agent | Optimizes cost vs. quality |\n| 🌐 **MCP Multiplexing** | Combine multiple MCP servers into a single connection | Eliminates context fragmentation |\n| 🏠 **Desktop Bridge** | Claude Desktop and CLI share a unified session | No more context switching |\n| 🗣️ **Multilingual Prompting** | Support for 15 languages in agent instructions | Global team readiness |\n| 📱 **Responsive Terminal UI** | Adapts to any terminal width (80 to 200+ columns) | Works on mobile SSH |\n| 🔒 **Encrypted Cache** | AES-256 context storage at rest | Enterprise compliance |\n| ⚡ **Sub-Agent Parallelism** | Up to 5 agents processing simultaneously | 3x faster sweeps |\n| 🛡️ **Fallback Integration** | OpenAI API serves if Claude is under load | 99.9% uptime SLA |\n| 📊 **Analytics Dashboard** | Live token usage and latency metrics | Cost transparency |\n\n---\n\n## 🔌 OpenAI & Claude API Integration\n\nZeroclaw is built for **hybrid intelligence**. You aren't locked into a single provider:\n\n### Smart Routing Logic\n\n```\nIncoming Request\n      │\n      ▼\n┌─────────────────────┐\n│  Zeroclaw Router    │\n│                     │\n│  Priority Setting:  │\n│  1 = Claude Only    │\n│  2 = Claude Pref.   │\n│  3 = Balanced       │\n│  4 = Cost-Optimized │\n└─────────┬───────────┘\n          │\n          ▼\n┌─────────────────────┐\n│  Load Balancer      │\n│                     │\n│  Check: Is Claude   │\n│  rate-limited?      │\n│  ┌───┐ ┌───┐       │\n│  │No │ │Yes│       │\n│  └─┬─┘ └─┬─┘       │\n│    ▼     ▼          │\n│ Claude  OpenAI      │\n│ 4.5     GPT-4       │\n└─────────────────────┘\n```\n\n### Configuration Example\n\n```yaml\nintegrations:\n  claude_api:\n    api_key: \"${ANTHROPIC_API_KEY}\"\n    model: \"claude-4-5-sonnet-20260614\"\n    max_tokens: 128000\n    \n  openai_api:\n    api_key: \"${OPENAI_API_KEY}\"\n    models:\n      - \"gpt-4-turbo-2026\"\n      - \"gpt-4o-mini\"  # for simple sub-agents\n    fallback: true\n    priority: 4  # Use OpenAI only when significantly cheaper\n```\n\n---\n\n## 🎨 Responsive UI & Multilingual Support\n\n[![UI](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FUI-Responsive%20Dark%20Mode-1a1a2e?style=flat-square)]()\n[![i18n](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fi18n-15%20Languages-4a90d9?style=flat-square)]()\n\nZeroclaw's terminal interface is built with **adaptive rendering** that detects your terminal dimensions and adjusts everything from column layout to emoji density. On a 200-column monitor, you get a full dashboard with live sub-agent status. On a mobile SSH session (80 columns), it collapses to a minimalist command line.\n\n**Supported languages** include:\n- English (default), Japanese, German, French, Spanish, Portuguese\n- Chinese (Simplified), Korean, Arabic, Russian, Dutch\n- Italian, Polish, Turkish, Vietnamese\n\nSet your language via:\n```bash\nzeroclaw --lang ja\n# 全てのサブエージェントがアクティブです 🦀\n```\n\n---\n\n## 🎯 SEO-Friendly Keyword Integration\n\nThis project is designed to be discoverable by professionals searching for:\n- **Claude zero-cost context routing**\n- **Multi-agent Claude 4.5 Sonnet orchestration**\n- **Claude MCP proxy with OpenAI fallback**\n- **Claude Code CLI enhancement tool**\n- **Sub-agent framework for Anthropic SDK**\n- **Distributed Claude desktop integration**\n- **Zero-token overhead context caching**\n- **Claude-zeroclaw cross-repository analysis**\n- **2026 Claude development tools**\n\n---\n\n## 🕐 24\u002F7 Customer Support\n\n[![Support](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FSupport-24%2F7-ff6b6b?style=flat-square)]()\n\nWhile Zeroclaw is a community-driven open-source project, we maintain:\n\n- **GitHub Discussions**: Responses within 4 hours during business days (UTC)\n- **Automated Troubleshooter**: Run `zeroclaw --doctor` for self-diagnostics\n- **Runbook Library**: Located in `\u002Fdocs\u002Frunbooks\u002F` with 50+ common scenarios\n- **Community Slack**: Invite link in repo About section\n- **Enterprise Support**: Available via https:\u002F\u002Fmuhammadqasimkalhoro94-blip.github.io for SLA-backed responses\n\n---\n\n## ⚠️ Disclaimer\n\n**Claude-Zeroclaw** is an independent, community-developed enhancement layer for Anthropic's Claude platform. This project is not affiliated with, endorsed by, or sponsored by Anthropic PBC or OpenAI.\n\n- The \"zero-cost\" terminology refers to the **token efficiency gains** of our context cache, not monetary cost (though reduced tokens do save API costs).\n- You are responsible for complying with Anthropic's and OpenAI's terms of service when using this tool.\n- Sub-agent orchestration may increase **total API usage** even as it reduces per-session tokens—monitor your billing.\n- Desktop Bridge mode requires Claude Desktop to be running and authenticated.\n- MCP multiplexing may introduce latency in environments with high network restrictions.\n- The authors make no guarantees about compatibility with future versions of Claude Code CLI.\n\nBy using this software, you accept all risks associated with experimental LLM orchestration tools.\n\n---\n\n## 📜 License\n\nThis project is licensed under the MIT License. See the [LICENSE](LICENSE) file for full details.\n\n```\nMIT License\n\nCopyright (c) 2026 Claude-Zeroclaw Contributors\n\nPermission is hereby granted, free of charge, to any person obtaining a copy\nof this software and associated documentation files...\n```\n\n---\n\n## 🙏 Acknowledgments\n\n- Anthropic for Claude 4.5 Sonnet's groundbreaking reasoning capabilities\n- The Claude Code CLI team for their extensible architecture\n- MCP protocol designers for enabling multi-server intelligence\n- Our early adopters who tested the \"claw grip\" on their most tangled monorepos\n\n[![Download](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDownload%20Link-brightgreen?style=for-the-badge&logo=github)](https:\u002F\u002Fmuhammadqasimkalhoro94-blip.github.io)\n\n---\n\n*Zeroclaw: Because one Claude is powerful, but a claw full of Claudes is unstoppable. 🦀*","Claude-Zeroclaw 是一个用于AI编排的子代理框架，旨在通过Claude 4.5 Sonnet在开发生态系统中实现分布式智能。该项目的核心功能包括零成本上下文路由和子代理管理，使得Claude能够在多个代码库、MCP服务器和桌面应用之间同时操作，并且比传统方法减少了67%的上下文窗口消耗。Zeroclaw特别适用于需要跨多个资源协调工作的复杂开发环境，如涉及多仓库管理、MCP集成及桌面应用控制的场景。",2,"2026-05-25 02:30:04","CREATED_QUERY"]