{"slug":"ruvnet-harness-security-bench","source_name":"ruvnet/harness-security-bench","name":"Ruvnet/Harness Security Bench","description":"Run `@metaharness/darwin security bench` (upstream \"Darwin Shield\" / ADR-155) — evolves a champion security-detection harness against a 10-vuln / 9-decoy corpus and grades it on TPR/FPR/patch-pass/repro/unsafe vs four baselines (B0 static, B1 LLM-single-pass, B2 fixed-agent, B3 Darwin-champion). Closest reference implementation for ruflo's own ADR-155 nightly self-learning security harness (PR","version":1,"lift":{"pass_rate_delta_pts":45.45,"pass_rate_pct":77.3,"total_cases":22,"passed_cases":17,"tokens_delta_pct":27.3,"turns_delta_pct":0,"verdict":"mixed","benchmark_model":"gemini-3.6-flash","grading_method":"judged","completed_at":"2026-08-07T21:27:54.617758+00:00"},"skill_score":0.7727,"benchmark_models":[{"model":"gemini-3.6-flash","headline":true,"delta_pts":45.45,"with_pass_pct":77.3,"without_pass_pct":31.8,"tokens_delta_pct":27.3,"turns_delta_pct":0,"total_cases":22,"cases_aggregated":17,"verdict":"mixed","never_hurt":true,"completed_at":"2026-08-07T21:27:54.617758+00:00","run_id":"49f2611c-99f7-464c-9958-3724e79dfdc5","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":"MIT","install_count":0,"manifest_hash":"f63f3a0b5e7929e7bdcbf0cb83448461b70864676c25870d834965c553448165","raw_url":"https://app.decimal.ai/s/ruvnet-harness-security-bench/SKILL.md","scorecard_url":"https://app.decimal.ai/skills/ruvnet-harness-security-bench"}