Taskmaster

Taskmaster is an open-source, AI-powered task management system built by Eyal Toledano and Ralph Khreish under the Hamster (Wheel Go Fast, Inc.) umbrella. It gives coding agents a structured plan before they execute, designed to work seamlessly with AI coding tools like Cursor, Windsurf, Claude Code, Codex, and others. The project is available as an npm package (task-master-ai) and can be used via MCP (Model Context Protocol) or the command line.

What It Is

Taskmaster is a task management layer for AI-driven development workflows. Rather than letting coding agents operate without direction, Taskmaster parses a Product Requirements Document (PRD) and generates a structured set of tasks with acceptance criteria, dependencies, and context that any coding agent can execute. The core insight, as stated in the project's own documentation, is that "agents with a plan produce better work than agents without one." It sits between the human developer's intent and the AI agent's execution, providing the structured context that keeps agents on track across long sessions.

How It Works

Taskmaster can be installed as an MCP server (recommended) or used directly from the command line. The MCP integration lets it run inside editors like Cursor, Windsurf, and VS Code without leaving the AI chat pane. Key workflow steps include:

• Initialize: Run task-master init to set up the project structure
• Parse a PRD: Feed a requirements document to generate a full task list
• Navigate tasks: Use next, show, and list commands to manage work
• Research: A built-in research command fetches fresh information with project context
• Sync with Hamster Studio: Briefs, tasks, and plans can be synced into the repo via the Hamster CLI for team use

The MCP server supports selective tool loading via the TASK_MASTER_TOOLS environment variable, with modes ranging from core (7 tools, ~5,000 tokens) to all (36 tools, ~21,000 tokens), letting teams optimize context window usage.

AI Provider Support

Taskmaster is model-agnostic and supports a wide range of AI providers. Users configure a main model, a research model, and a fallback model. Supported providers include:

• Anthropic (Claude)
• OpenAI
• Google Gemini
• Perplexity (research model)
• xAI (Grok)
• OpenRouter
• Mistral
• Groq
• Azure OpenAI
• Ollama (local)
• Claude Code CLI (no API key required)
• Codex CLI (OAuth via ChatGPT subscription)

Adoption Signal

The GitHub repository (eyaltoledano/claude-task-master) reports 27,656 stars and 2,601 forks as of its last recorded update. The Hamster homepage states the project reached 1.5 million downloads and 25,000 GitHub stars, attributing this growth to the core insight that structured plans improve agent output. The npm package is task-master-ai.

Update: task-master-ai@0.43.1

The latest published release is task-master-ai@0.43.1, published on March 31, 2026. The repository was last updated in June 2026, indicating active maintenance. The project has evolved from a solo-developer tool into the CLI backbone of the broader Hamster Studio platform, which adds team collaboration, briefs, and two-way sync with issue trackers like Linear and Jira on top of Taskmaster's task execution layer.

Licensing Note

Taskmaster is licensed under the MIT License with a Commons Clause condition. This allows free use, modification, and distribution for any purpose — including building commercial products with it — but prohibits selling Taskmaster itself or offering it as a hosted service.