Diagnose Benchmark Failure

Use when a flowai benchmark fails and you need the cause from run artifacts before editing. Reads judge-evidence.md, the raw agent transcript, the sandbox SKILL.md, and scenario mod.ts; classifies the failure against a known taxonomy; produces an evidence-grounded report (no fixes). Do NOT trigger for passing benchmarks or generic skill iteration.

korchasa Updated 3 repo stars

File contents

korchasa/flowai commit 07c659b3ef

Frequently asked questions

npx get-skill add korchasa/diagnose-benchmark-failure