Benchmark
Skill keep-starknet-strange/starknet-skills/.claude/skills/benchmark
Run evaluation benchmarks and report precision/recallFrom its SKILL.md
npx -y skills add keep-starknet-strange/starknet-skills --skill benchmarkAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
1.1 KB, as published. Nobody here has run it
Benchmark Skill
Quick Start
- Run auditor benchmark:
python3 scripts/quality/benchmark_cairo_auditor.py --cases evals/cases/cairo_auditor_benchmark.jsonl --output /tmp/bench-auditor.md --min-precision 0.90 --min-recall 0.90 - Run contract benchmark:
python3 scripts/quality/benchmark_contract_skills.py --cases evals/cases/contract_skill_benchmark.jsonl --output /tmp/bench-contracts.md --min-precision 0.95 --min-recall 0.95 - If
$ARGUMENTScontains--save, append--saveto both commands to copy scorecards intoevals/scorecards/. - Report precision/recall and fail conditions per suite.
- Follow the full workflow: Benchmark Workflow
When to Use
- You changed skills, references, datasets, or evaluation logic.
- You need comparable KPI numbers before/after a PR.
When NOT to Use
- You only need schema/structure validation (use validate skill instead).
- Benchmark inputs or dependencies are missing.