The Incoherence Detector skill provides a specialized 22-step workflow designed to audit, identify, and resolve discrepancies where documentation has become stale or misaligned with the actual codebase. It offers a systematic approach to detecting contradictions, categorizing them by severity, and facilitating a reconciliation phase to update either the code or the docs. While this skill is now deprecated in favor of the more robust 'doc-accuracy' skill, it remains essential for teams maintaining legacy workflows that depend on the specific scripts/incoherence.py reconciliation logic.
主な機能
01Structured 22-step audit and reconciliation process
02Comprehensive report generation with actionable resolution suggestions
03Automated detection of doc-vs-code contradictions
0435 GitHub stars
05Severity-based reporting for critical and high-priority issues
06Legacy support for scripts/incoherence.py reconciliation scripts
ユースケース
01Auditing legacy codebases for stale or outdated technical documentation
02Finding and fixing contradictions between implementation details and README files
03Executing manual reconciliation steps for complex documentation audits