taskmajor
Local-first task management engine for AI agents. MCP-compatible, highly customizable profiles (GTD, Energy-aware, Context tags), powered by an embedded TaskWarrior engine.
Install / Use
claude mcp add sznicolas -- npx -y github:sznicolas/taskmajorIf the server publishes to npm under a different name, use that package instead — check the repo README.
MCP Server
Model Context Protocol server
Quality Score
Category
Development & EngineeringSupported Platforms
Tags
Skill content
View source on GitHubTaskMajor
Your tasks. Your rules. Your agents.
A local-first MCP server that gives AI agents native access to a full TaskWarrior-compatible engine.
No cloud. No binary to install. No lock-in. Just a pip install and your agents are in control.
Ask your agent:
"Add a task to review the API spec"
"What's on my plate for today?"
"Run my weekly review"
"Show me the roadmap for Project X"
Why TaskMajor?
Most task managers are cloud-first and agent-hostile. TaskMajor is neither.
- 🔒 100% local & private — all data stays on your machine, stored as plain text via the embedded TaskChampion engine
- ⚙️ Full TaskWarrior power — projects, priorities, due dates, tags, UDAs, dependencies, contexts — all accessible by your agent
- 🎭 Profile System — the feature no other task MCP has: define exactly what your agent can see and do, adapted to your methodology
- ⚡ Zero friction — one command installs and starts a fully functional MCP server
🚀 Quick Start
pip install taskmajor # or: uv tool install taskmajor
taskmajor # starts at http://localhost:8888
Point your MCP client (Claude Desktop, Hermes, Continue, or any MCP-compatible agent) at http://localhost:8888.
That's it. No TaskWarrior binary to install — the engine is embedded.
🤖 Hermes user in a hurry? → 5-minute Hermes setup
Want a specific workflow?
taskmajor --profile standard # task CRUD + agenda (default)
taskmajor --profile productivity # GTD: reviews, energy levels, contexts
taskmajor --profile project-mgmt # sprints, effort tracking, blockers queue
→ Full installation options — uv, Docker, from source
📸 Demo


🎭 Profile System — TaskMajor's Killer Feature
Profiles are the heart of TaskMajor. Each profile defines a complete behavioral contract for your agent: what tools it can call, what data structures exist, what views it can read, and how it should reason about your tasks.
| Profile | Extends | What it unlocks |
|---|---|---|
| minimal | — | Bare CRUD: add, query, complete, delete |
| standard | minimal | Agenda views, date tools, metadata, organization prompts |
| productivity | standard | GTD: energy levels (low/medium/high), daily & weekly reviews, capture workflow, context tags (+@) |
| project-mgmt | standard | Sprint planning, effort analytics, blockers queue |
| your-profile | any built-in | Extend any profile without touching the package |
Custom profile in 3 lines:
# ~/.config/myproject/profiles/myprofile/manifest.yaml
extends: [productivity]
# add your own UDAs, contexts, resources, and agent instructions
taskmajor -c ~/.config/myproject
→ Profile System documentation
✨ Key Features
- 🔒 Private by design — plain-text storage, TaskChampion engine, runs entirely on your machine
- 🧠 Profile-shaped agents — agents see only what you expose; switch workflows with one flag
- ⚡ Energy-aware scheduling —
low / medium / highenergy UDA lets your agent suggest tasks that match your current state - 🏷️ Safe context tags —
+@work,+@home,+@focusfilter tasks without breaking TaskWarrior's native context system - 📅 Flexible date expressions — natural language dates resolved before any write operation
- 📋 Daily & weekly reviews — structured review resources and prompts built into the
productivityprofile - 🔄 Optional sync — periodic or manual sync to a TaskChampion server (local or remote)
- 🔌 Any MCP client — works with Claude Desktop, Hermes, Continue, or any MCP-compatible agent
- 🐳 Docker-ready — official Dockerfile for containerized deployments
- 📊 Built-in analytics — stats by project, priority, and status; effort and sprint tracking in
project-mgmt
📚 Documentation
| For | Start here | |---|---| | First install | Getting Started | | 5-minute walkthrough | Quick Start guide | | Hermes agent users | Hermes Quick Start | | Custom workflows | Profile System | | All profiles reference | Profile Reference | | Tools & resources API | API Reference | | Config & CLI flags | Configuration | | Docker / from source | Deployment | | Contributing | Contributing |
🌐 Community
- Issues & Features: GitHub Issues
- Discussions: GitHub Discussions
📄 License
MIT — see LICENSE
Built with FastMCP · pytaskwarrior · TaskChampion · Pydantic
Related Skills
Agent-Reach
79.0kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
ruflo
71.8k🌊 The original agent meta-harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, RAG integration, and native Claude Code / Codex / Hermes and many more Integrated
headroom
71.0kCompress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
CowAgent
46.9kOpen-source super AI assistant & Agent Harness. Plans tasks, runs tools and skills, self-evolves with memory and knowledge. Multi-agent, multi-model, multi-channel. Lightweight, extensible, one-line install. (formerly chatgpt-on-wechat)
