Claude Task Master

Claude Task Master (package name task-master-ai) turns the way you work with coding agents from improvised chat into managed projects. You feed it a PRD and it parses it into a dependency-tracked task list; the agent then works through tasks with next_task, set_task_status, complexity analysis, and auto-expansion of big tasks into subtasks. It ships as an MCP server, a Claude Code plugin with 49 slash commands and 3 specialized agents (orchestrator, executor, checker), and a CLI — and works with Cursor, Lovable, Windsurf, Roo, and VS Code.

With ~28,200 GitHub stars it is one of the most-starred agent-workflow tools ever, and the tool-set tiers (core/standard/all) show real care about token budgets — core mode cuts roughly 70% of tool-call tokens. You can also run it against Claude Code's own CLI with no API key at all.

The honest caveat: the license is MIT with a Commons Clause, which bars selling it as a hosted service or building competing products on it — read the terms before commercializing around it. Also, the repo's last push was April 2026, so it is in maintenance mode while the ecosystem moves fast. Still the reference implementation for PRD-driven agent task management.

1. Install: npm install -g task-master-ai (or add via MCP: claude mcp add task-master-ai -- npx -y task-master-ai@latest).
2. In your project, run task-master init and point it at your PRD (task-master parse-prd your-prd.txt).
3. Ask your agent to work the queue: What's the next task? — it uses get_tasks/next_task/set_task_status.
4. Use /tm:parse-prd, /tm:next, /tm:expand slash commands if you installed the Claude Code plugin.