Metacog
Stops your agents repeating the same mistakes.
2 followers
Stops your agents repeating the same mistakes.
2 followers
AI coding agents can't feel when they're stuck. They'll retry the same broken fix five times because they can see each error but not the pattern. Metacog is a Claude Code hook that watches for these loops, and six other failure modes, silently. When things go wrong, it nudges the agent to reflect. When problems resolve, it learns what fixed them. Most memory systems prune rules that stop firing. Metacog treats that silence as proof the rule worked. Zero dependencies. One-command install.
Thanks for checking this out. I've been watching it watch my Claude Code sessions and actively iterating on what it does. It started as six senses detecting failure patterns, and it's already grown to seven (scope drift detection), plus session retrospectives and user interaction tracking to give feedback to the user at session end, all from observing where it helped and where it fired false positives. If you have feedback, including the negatives, I genuinely want it. This thing gets better by being honest about what can be improved.