msal2020/groundtruth

Catch your AI coding agent when it lies — verify its claims against reality (runs tests, audits the diff, checks registries).

View on GitHub

Trust Signals

Scorecard Score
not yet scored
Maintenance Recency
Stale
License
None
namedescriptionrequireddefault
claimThe agent's claim to verify (e.g. the PR description). Optional — the diff battery runs regardless.no""
baseBase ref to diff against. Defaults to the PR base branch.no""
commentPost/update a sticky PR comment with the receipts (needs pull-requests: write).notrue
fail-on-overclaimFail the job when overclaims are found.notrue

no outputs