{"slug":"jgravelle-benchmark-methodology","source_name":"jgravelle/benchmark-methodology","name":"Jgravelle/Benchmark Methodology","description":"Rules for measuring or quoting any benchmark number in this repo. Load before running the bench tier, writing a delta into a PR or CHANGELOG, or touching anything under benchmarks/.","version":1,"lift":{"pass_rate_delta_pts":68.18,"pass_rate_pct":81.8,"total_cases":22,"passed_cases":18,"tokens_delta_pct":-13.9,"turns_delta_pct":0,"verdict":"mixed","benchmark_model":"gemini-3.6-flash","grading_method":"judged","completed_at":"2026-10-06T07:50:07.251899+00:00"},"skill_score":0.8182,"benchmark_models":[{"model":"gemini-3.6-flash","headline":true,"delta_pts":68.18,"with_pass_pct":81.8,"without_pass_pct":13.6,"tokens_delta_pct":-13.9,"turns_delta_pct":0,"total_cases":22,"cases_aggregated":19,"verdict":"mixed","never_hurt":true,"completed_at":"2026-10-06T07:50:07.251899+00:00","run_id":"487398ee-4a81-4326-ad92-561c58d42a42","latest_attempt":{"id":"487398ee-4a81-4326-ad92-561c58d42a42","runner_model":"gemini-3.6-flash","overall_verdict":"mixed","completed_at":"2026-10-06T07:50:07.251899+00:00","gate":null},"version_number":1,"is_latest_version":true,"gate":null}],"trust":{"skill_safety":"passed","safety_status":"clean","intent_verdict":"safe","content_status":"clean","indexable":true},"license":"NOASSERTION","install_count":0,"total_uses":0,"total_users":0,"manifest_hash":"da1e2f2360194fa9ada913d8cc1044738e948c8a276ebe559d7ee9a0e0b1c3c5","install_name":"jgravelle-benchmark-methodology","version_id":"ff9836d2-9b90-4e81-93ca-4477f0f28108","body_sha256":"da1e2f2360194fa9ada913d8cc1044738e948c8a276ebe559d7ee9a0e0b1c3c5","render_hash":"0f56745cb688393df4e94e153a27ea3d04d1b3545139b7ac67d061affcb58c9f","raw_url":"https://app.decimal.ai/s/jgravelle-benchmark-methodology/SKILL.md","scorecard_url":"https://app.decimal.ai/skills/jgravelle-benchmark-methodology"}