What skills are you evaluating?

[hoainho]

Curious what people are throwing at eval-harness in the wild.

If youve baselined a skill (opencode skill, Claude agent, LangGraph node, anything) and run it through here, drop:

• What it does (one line)
• How many cases you wrote
• What kind of regressions youve caught (FIXTURE_STALE, SKILL_CHANGED, MODEL_CHANGED, flaky)

Im especially interested in unusual check kinds people are wishing existed — that feedback drives the v0.5.0 roadmap.

— @hoainho