Skip to main content

Skillet

Build better Claude Code skills through systematic evaluation. Instead of guessing whether your skills work, capture real failures, measure performance, and iterate with data.

Why Skillet?

Capture Gaps

Record when Claude Code fails or produces suboptimal results. Turn frustrations into test cases.

Run Evaluations

Systematically test your skills against captured scenarios. Get quantitative feedback on improvements.

Iterate with Data

Make targeted improvements based on actual failure modes, not hunches.

Ship with Confidence

Know your skills work before deploying them to production.

Quick Start

Install Skillet:
pip install skillet
Then capture a gap, run an eval, and improve your skill.