30+ battle-tested AI Agent skills. Built by developers, for developers.
Every skill is MIT licensed, optimized for OpenClaw, and ready to use in minutes.
One command to install any skill
openclaw skills install skylv-agent-evaluator
Built from real agent workflows.
Works out of the box.
Free to use and modify.
Designed for OpenClaw agents.
Regular improvements.
Part of ClawHub marketplace.
Score AI agent behavior across 5 criteria: accuracy, efficiency, clarity, safety, helpfulness. Essential for multi-agent systems.
"Thinking about thinking" — detect cognitive biases, self-reflection, self-correction. Based on EvoMap GEP principles.
Auto-repair agent errors using pattern recognition. Handles crashes, tool failures, memory leaks.