superpowers-doctor

Diagnose malformed YAML and dead references by running 27 tiered checks that scan every installed skill, classifying structural defects as critical, error, warning, or info. The scan covers both local and overlay skill repositories to catch configuration regressions before downstream failures occur.

Yes, you can automatically fix trigger collisions and orphaned installations using graduated remediation modes. The tool supports --fix-safe, --fix, and --fix --purge-orphans flags, creating verified backups before applying any changes to ensure your skill ecosystem remains recoverable.

CRLF and UTF-8 BOM issues arise from platform-specific line ending and encoding mismatches, particularly on WSL or NTFS setups. The diagnostic tool detects these encoding defects alongside stale or dirty checkouts during post-clone audits to prevent rendering and parsing failures.

Yes, the doctor detects content and reference drift between overlay and source skill repositories. It surfaces reviewer and agent rendering drift during post-install audits or scheduled maintenance, ensuring installed skills match their source definitions and preventing silent configuration degradation.

Python3 and node are optionally required to exercise specific tests like TODO validation and reviewer-dispatch checks. The core 27 tiered integrity checks run independently, but enabling these dependencies provides deeper platform-aware diagnostics across the installed skill ecosystem.

Run integrity checks after bulk skill edits, after cloning repositories on Windows, or as a pre-release audit to catch corruption and configuration regressions. Scheduled maintenance also benefits from these scans to detect drift, orphaned installs, and trigger collisions before they cause failures.