The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses platform-specific tools (skill_view, skill_manage, skills_list, read_file, search_files, terminal, patch) to perform its auditing functions. These are standard operations for a meta-programming or quality-assurance skill.
[COMMAND_EXECUTION]: The skill instructs the agent to create and run temporary Python scripts for ad-hoc verification of skill edits. The provided pattern uses tempfile.NamedTemporaryFile and subprocess.run, which is a controlled way to verify logic changes during development.
[REMOTE_CODE_EXECUTION]: While the skill mentions installing skills from GitHub URLs, it emphasizes using native platform validators and auditing tools (hermes skills audit) to ensure safety before and after installation.
[DATA_EXPOSURE]: The skill includes a 'Public Skill Package Release Gate' section that explicitly instructs users to perform privacy scans for credentials, personal paths, and private IPs before sharing skills, which is a security best practice.

skill-auditor