codebase-audit

To verify README documentation matches a codebase, perform a claims-first audit that extracts verifiable statements and systematically attempts to falsify them using static analysis and test evidence. This process identifies discrepancies between documented features and actual code behavior.

A falsification-first approach for checking code against documentation attempts to disprove each extracted claim by searching the codebase for counter-evidence before looking for supporting evidence. This method prioritizes finding documentation inaccuracies and inconsistencies over simple validation.

You can automatically extract and verify technical claims from JSDoc comments by running a claims-first audit. The process identifies verifiable claims from specified documentation files and uses dependency tracing and runtime probes to confirm accuracy against the current code.

The best way to audit setup instructions against source code is to run an automated claims extraction and falsification process. This systematically verifies stated features and behavioral descriptions by probing the codebase for counter-evidence, generating a detailed remediation report for discrepancies.

A codebase audit works for identifying technical debt from outdated documentation by systematically tracing dependencies and probing runtime behavior to falsify documented claims. It categorizes extracted statements and highlights inconsistencies, directly exposing documentation drift and technical debt.

Documentation claims are disproved during a code audit using static analysis, test evidence, runtime probes, and dependency tracing. The falsification-first methodology searches the codebase for counter-evidence against extracted claims before validating supporting evidence to ensure accuracy.