Last commit on September 2, 2026·Created on June 9, 2026
tigerless-labs/autoharness
“Distills active conversation lessons into a durable, reusable skill library”
Combined rank
#217
across all skills
Stars
1.4k
7d change unavailable
Forks
84
Watchers
58
Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of real attention, adoption, and active maintenance.
TL;DR
This skill identifies class-level techniques and corrected approaches within a session to prevent the loss of non-trivial insights. It helps users maintain a persistent knowledge base by filtering out one-off narratives and environment-specific failures in favor of durable preferences.
The process involves comparing new lessons against an existing skill index to avoid duplication and reconciling contradictions so that the most recent, accurate version of a rule prevails.
WHO IT'S FOR
AI agent / automation builders
creating self-improving skill libraries for agents
Claude Code power users
automating the capture of non-trivial coding techniques
Platform / DevEx teams
standardizing AI coding rules across a team
Technical leads / architects
governing the evolution of AI agent knowledge
Want to install it?
Claude Marketplace · Claude Plugin · Clone to Claude · Manual CopyRepository instructions