The Agent Skills Directory

[SAFE]: No malicious patterns, prompt injections, or data exfiltration attempts were detected. The skill's behavior is consistent with its stated purpose as an integrity guard.
[COMMAND_EXECUTION]: The skill utilizes subprocess.run to interact with the macOS /bin/launchctl utility. This is used for the documented, optional feature of installing background monitoring services and does not process untrusted external input.
[EXTERNAL_DOWNLOADS]: The skill does not perform any network requests or download external code; it operates entirely on local workspace data.
[SAFE]: The skill implements security best practices, such as refusing to operate on symlinks to prevent link-following attacks and using atomic file operations for restores.

soul-guardian