The Agent Skills Directory

[SAFE]: The skill provides instructions for identifying over-engineered code and suggests simplifications using standard library features. It does not contain any executable code, sensitive file access, or network communication.
[SAFE]: While the skill ingests codebase data which is an untrusted source, its operation is confined to a one-shot report with no automated code modifications or external exfiltration capabilities. Ingestion point: local or provided codebase files. Boundary markers: none. Capability inventory: no subprocess or exec calls detected. Sanitization: none required for reporting-only scope.

ponytail-audit