{"id":50428546,"url":"https://github.com/jc1122/perf-benchmark-skill","last_synced_at":"2026-06-14T13:02:19.564Z","repository":{"id":342408995,"uuid":"1173092309","full_name":"jc1122/perf-benchmark-skill","owner":"jc1122","description":"Linux performance benchmarking skill for Claude Code — CPU, cache, branch, memory, ASM analysis with 7-dimension rubric","archived":false,"fork":false,"pushed_at":"2026-06-10T20:45:18.000Z","size":250,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-06-10T22:18:03.032Z","etag":null,"topics":["assembly","benchmarking","claude-code","cpu","leaf-skill","linux","performance","profiling","skill"],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/jc1122.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-03-05T02:09:04.000Z","updated_at":"2026-06-10T20:45:12.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/jc1122/perf-benchmark-skill","commit_stats":null,"previous_names":["jc1122/perf-benchmark-skill"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/jc1122/perf-benchmark-skill","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jc1122%2Fperf-benchmark-skill","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jc1122%2Fperf-benchmark-skill/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jc1122%2Fperf-benchmark-skill/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jc1122%2Fperf-benchmark-skill/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/jc1122","download_url":"https://codeload.github.com/jc1122/perf-benchmark-skill/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jc1122%2Fperf-benchmark-skill/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":34226629,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-06-12T02:00:06.859Z","response_time":109,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["assembly","benchmarking","claude-code","cpu","leaf-skill","linux","performance","profiling","skill"],"created_at":"2026-05-31T12:02:14.384Z","updated_at":"2026-06-14T13:02:19.559Z","avatar_url":"https://github.com/jc1122.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# perf-benchmark-skill\n\nLinux performance benchmarking skill for coding agents that support Skills. It\nprofiles Python and C workloads, scores a 7-dimension rubric, and keeps\nalgorithmic issues ahead of cache, branch, and ASM tuning.\n\nFor repo-agnostic use, pass an explicit `--target` or `--binary`.\n\n## Installation\n\n```bash\nnpx skills add \u003cskill-source\u003e/perf-benchmark-skill\n```\n\n`\u003cskill-source\u003e` means the installable source or repository path that hosts this skill.\n\n## Scope\n\n- Algorithmic scaling, wall-time stability, CPU efficiency, cache behavior,\n  branch prediction, memory profile, and optional ASM review\n- Linux-only, using `/proc`, `/sys`, Valgrind, and `perf`\n- `SKILL.md` is the detailed agent-facing workflow and CLI reference\n\nPytest benchmark autodiscovery is a convenience for Python repos. For\nnon-pytest entrypoints, use `--target` or `--binary`.\n\nMulti-size explicit targets must include `{SIZE}`.\n\nFull Algorithmic Scaling scoring requires `deep` or `asm` because allocation churn comes from massif.\n\n## Usage\n\n```bash\npython scripts/perf_benchmark_pipeline.py \\\n  --root /path/to/repo \\\n  --out-dir /tmp/bench \\\n  --tier deep \\\n  --target \"python -m benchmark_entrypoint {SIZE}\" \\\n  --sizes 10000,100000 \\\n  --source-prefix path/to/source/ \\\n  --perf-record \\\n  --max-cv 5.0 \\\n  --findings-out /tmp/perf-findings.json \\\n  --baseline-ledger /tmp/perf-ledger.jsonl\n```\n\n`--perf-record` is opt-in native hotspot sampling via `perf record` and\n`perf report`. Use it when `perf` is available and you want flat sampled\nhotspots in addition to the rubric.\n\n`--max-cv` (default 5.0) sets the coefficient-of-variation noise gate:\ntiming-derived dimensions exceeding this threshold are scored `N/A (noise)`.\n`--findings-out` writes shared-schema PERF findings JSON (one per FAIL/WARN\ndimension, `signal: \"PERF\"`). `--baseline-ledger` maintains an append-only\nJSONL run history with vs-last and vs-best regression checks; can be used\nalongside `--baseline` for point-in-time comparison.\n\n## Outputs\n\n- `benchmark_report.md`: scorecard, findings, prescriptions\n- `benchmark_summary.json`: machine-readable scores and regression data\n- `perf_findings.json`: shared-schema PERF findings (when `--findings-out` set)\n- `baseline_ledger.jsonl`: append-only run history (when `--baseline-ledger` set)\n- `tier1/` to `tier4/`: raw profiler artifacts by depth\n\n## More Detail\n\nSee [SKILL.md](SKILL.md) for the full workflow, tier behavior, agent guidance,\nand reference links.\n\n## Related Skills\n\n### perf-optimization (v0.1.0)\n\nThe [`perf-optimization/`](perf-optimization/) directory contains a companion\nskill that consumes `perf-benchmark` findings and applies an iterative\nmeasure -\u003e change -\u003e re-measure ratchet to systematically resolve diagnosed\nbottlenecks. It selects the highest-impact candidate per iteration, makes one\nbounded change, re-runs profiling under identical conditions, and records\naccepted wins in an append-only ledger. Algorithmic scaling failures gate all\nconstant-factor work.\n\nSee [perf-optimization/SKILL.md](perf-optimization/SKILL.md) for workflow details\nand verification requirements.\n\n## License\n\nMIT\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjc1122%2Fperf-benchmark-skill","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fjc1122%2Fperf-benchmark-skill","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjc1122%2Fperf-benchmark-skill/lists"}