{
  "markdown": "# 🪖 ClaudeArmy\n\n![Version](https://img.shields.io/badge/version-0.4.1-blue) ![License](https://img.shields.io/badge/license-MIT-green) ![Node](https://img.shields.io/badge/node-18%2B-brightgreen)\n\nAn MCP server that lets Claude orchestrate multiple Claude Code agents working autonomously across different projects.\n\n![ClaudeArmy Demo](demo.gif)\n\n## What It Does\n\nClaudeArmy gives Claude (in the chat interface) the ability to spawn background Claude Code processes that work independently on coding tasks. Think of it as a chain of command:\n\n- **You** → give strategic direction\n- **Claude (chat)** → breaks it down and dispatches agents\n- **Claude Code agents** → execute autonomously, spawning their own sub-agents as needed\n\nThis means you can kick off work across multiple projects simultaneously, keep chatting normally, and check in on progress whenever you want.\n\n## Tools\n\n| Tool | Description |\n|------|-------------|\n| `dispatch_task` | Deploy a Claude Code agent to a project directory. Supports `mode: \"team\"` for multi-agent coordination |\n| `check_tasks` | Monitor status of all running/completed agents |\n| `get_task_output` | Retrieve what an agent did and its full output |\n| `get_agent_team` | Detailed view of a lead agent and all its sub-agents |\n| `cancel_task` | Stop a running agent gracefully |\n| `purge_tasks` | Clean up completed/failed tasks |\n\n## Setup\n\n1. Clone this repo\n2. `npm install`\n3. Add to your Claude Desktop config:\n\n   - **Linux:** `~/.config/Claude/claude_desktop_config.json`\n   - **macOS:** `~/Library/Application Support/Claude/claude_desktop_config.json`\n   - **Windows:** `%APPDATA%\\Claude\\claude_desktop_config.json`\n\n```json\n\"claude-army\": {\n  \"command\": \"node\",\n  \"args\": [\"/path/to/claude-army/src/index.js\"]\n}\n```\n\n> **Windows note:** Use forward slashes in the path value (e.g. `\"C:/Users/you/claude-army/src/index.js\"`).\n\n4. Restart Claude Desktop\n\n## Usage Examples\n\n**Single project task:**\n> \"Refactor GarbageFire's app.py into modular components\"\n\n**Multi-project parallel work:**\n> \"Refactor GarbageFire to be more modular, and add a Salesforce adapter to my Reporting Tool\"\n\nClaude will dispatch separate agents to each project directory and let them work autonomously.\n\n**Check progress:**\n> \"How are the agents doing?\"\n\n## Agent Teams\n\nUse `mode: \"team\"` on `dispatch_task` to have the lead agent automatically decompose work and spawn specialized sub-agents that run in parallel.\n\n**How it works:**\n- The lead agent analyzes the task and breaks it into focused subtasks\n- Each subtask is delegated to a sub-agent via Claude Code's `Task` tool\n- Sub-agents work in parallel, each with a well-defined scope\n- The lead agent synthesizes results once all sub-agents complete\n\n**Monitoring teams:**\n- `check_tasks` shows a sub-agent summary (e.g. \"3 sub-agents: 1 running, 2 completed\")\n- `get_agent_team` gives a detailed tree view of the lead agent and every sub-agent, including status, runtime, and output previews\n\n## Configuration\n\nEdit the constants at the top of `src/index.js`:\n\n- `CLAUDE_BINARY` - Path to your Claude Code binary\n- `MAX_CONCURRENT_TASKS` - Max simultaneous agents (default: 5)\n\n## Requirements\n\n- Claude Code installed and authenticated\n- Node.js 18+\n- Claude Desktop or any MCP-compatible client\n\n## Changelog\n\n### v0.4.1\n- Bump `@modelcontextprotocol/sdk` to ^1.26.0, resolving 3 high severity supply chain vulnerabilities\n- Add `repository` field to package.json\n\n### v0.4.0\n- Agent team support: track sub-agents spawned via Claude Code Task tool\n- New `get_agent_team` tool: tree view of lead + all sub-agents\n- `check_tasks` now shows sub-agent summary for team operations\n- `dispatch_task` gains `mode` parameter: `solo` (default) or `team`\n- Cross-platform: fixed Windows binary (`claude.cmd`) and signal handling\n- Error messages now show platform-appropriate diagnostic commands\n- README: added setup paths for Linux, macOS, and Windows\n\n### v0.3.0\n- Fixed stream-json parsing to correctly handle assistant message content arrays\n- Staleness indicator: `check_tasks` now shows time since last agent activity\n- Better error surfacing: actionable messages for CLI not found, permission denied, and mid-run failures\n- Fixed output duplication in `get_task_output`\n- Demo GIF added to README\n\n### v0.2.0\n- Real-time progress tracking via Claude Code's stream-json output\n- `check_tasks` now shows recent agent activity (file reads, edits, bash commands) instead of just runtime\n- `get_task_output` includes a full progress timeline with timestamps\n- Structured event parsing for tool use, assistant messages, and results\n\n### v0.1.0\n- Initial release\n- Core tools: dispatch_task, check_tasks, get_task_output, cancel_task, purge_tasks\n- Fire-and-forget architecture with background process management\n- Max 5 concurrent agents\n- stdio transport for MCP communication\n\n## License\n\nMIT\n",
  "bytes": 4877,
  "sha": "f05948b4e6122f5fe0d6dee3ddcb4462f2b5176c24a7bb882bed949c5b35bfd4",
  "repo_slug": "robbhimself-1337/claude-army",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_robbhimself_1337_claude_army_36e3d2cb/readme"
}