skills-audit

Scanning source files for hardcoded credentials before publishing involves running an audit that flags potential secrets and credential-like patterns to prevent leakage. This audit enforces checks for hardcoded paths, identity references, credentials, and private project references.

Cross-skill leakage occurs when internal references or private URLs from one skill expose data in another during repository mirroring. Preventing cross-skill leakage requires auditing skill source files to detect and neutralize internal references before publishing to a public repo.

Yes, you can audit only changed or new skills since the last sync instead of the whole repository. The audit scope can target a single skill, multiple named skills, or exclusively all changed and new skills since the last repository sync.

Checking plan-file naming conventions for compliance before making skills public is handled by an audit that enforces specific naming rules. The audit evaluates skill source files and returns a structured report highlighting any naming convention violations or publication blockers.

No, the audit does not require any external dependencies or linting tools to scan for personal references. It operates independently to check skill source files for personal paths, names, and private project mentions without needing additional components.