From superpowers
Verifies reproducibility claims for experiments: requires exact commands, config, seed, commit hash, dataset version, and metric tables before accepting performance improvements.
How this skill is triggered — by the user, by Claude, or both
Slash command
/superpowers:reproducibility-checkThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
No performance claim without attached evidence.
No performance claim without attached evidence.
Before saying "this is better" or "the bug is fixed", verify:
Read references/evidence-checklist.md when you need the full checklist.
State:
npx claudepluginhub shunyangliu/superpowers_dl --plugin superpowersAudits, runs, or verifies experimental evidence for CS/AI/ML papers. Produces Evidence Inventory with evidence type annotations and Protocol Risk assessments.
Enforces fresh verification of results before making claims: run analysis, read output, check assumptions, then report. Prevents asserting findings without evidence.
Captures provenance records (environment, code, data, seeds, commands) so that results, experiments, or analyses can be reproduced exactly by others or future sessions.