Claim Audit
Audits project records and documentation to detect, verify, or hedge unsourced quantitative claims.
Claim Audit scans project documentation, decision logs, and specifications to identify bare numerical claims that lack verifiable citations or explicit hedging. By checking that referenced file paths exist, validating citation patterns, and surfacing ungrounded assertions, it helps teams maintain trustworthy technical records. Users can resolve findings by providing verifiable sources, labeling claims as [UNVERIFIED], or pruning unsupported metrics, with built-in baseline tracking for easy adoption across existing repositories and CI pipelines.
Características Principales
01Validates file paths and citations against the filesystem to prevent dead reference links
02Generates content-hashed baselines to track new claims without getting blocked by legacy text
03Integrates into Claude Code PostToolUse hooks and CI workflows with strict exit codes
04Scans Markdown files for numerical claims lacking traceable source citations or hedging
05Automatically identifies and skips justified restatements of previously cited figures
0617 GitHub stars
Casos de Uso
01Verifying performance claims, benchmark statistics, and metrics in closeout reports
02Automating documentation quality gates in CI or via Claude Code hooks during editing sessions
03Auditing architectural decision logs and project specs before committing to implementation