The Agent Skills Directory

[SAFE]: Instructional content only. The skill consists solely of markdown guidelines and YAML configuration metadata. It contains no scripts, binaries, or automated tool definitions, posing no risk of direct code execution.
[SAFE]: Secure history management. The instructions specifically advocate for the use of --force-with-lease during history rewrites, which is a recognized development safety mechanism to prevent accidentally overwriting remote changes made by other contributors.
[SAFE]: Absence of network or sensitive operations. The skill does not include any commands or patterns for external data exfiltration, remote script downloads, or unauthorized access to sensitive file paths such as credentials or environment files.

minimal-evidence-checks