#737 · Primary category： AI Agents & Automation
autoharness
Autoharness — a self-learning skill layer for Claude Code — distills skills from your real sessions, updates them as you work, and prunes the ones that stop getting used. No daemon, no benchmark.
Project last updated：07/28/26
GitHub Stars
1.3K
Forks
82
Contributors
2
License
MIT
Why we included this project
Anyone who spends the day in Claude Code knows the drill: restate the same assumptions, re-explain the same conventions, every session. AutoHarness watches that real work and distills the recurring scenarios into reusable skills, keeping the library tidy on its own: overlapping skills get merged into fewer ones instead of stacking near-duplicates, and anything that stops being used is pruned. It only touches the skills it generated itself, so your hand-written or installed ones stay untouched. The whole thing runs as plain Python with zero third-party dependencies, connected through Claude Code hooks and an MCP server, so there is no daemon to babysit and no evaluation pipeline to maintain. It is a quiet bet, and the numbers point toward the payoff: the same model went from 42% to 78% on CORE-Bench once it had the harness.
Articles for this project
No articles for this project yet.
To suggest a topic or contribute an article, contact us.
Related projects in this category
openclaw
Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞
hermes-agent
The agent that grows with you
n8n
Fair-code workflow automation platform with native AI capabilities. Combine visual building with custom code, self-host or cloud, 400+ integrations.
deepseek-harness
DeepSeek Harness: Everything is a Plugin.
AutoGPT
AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matters.