meta-audit

Pass

Audited by Gen Agent Trust Hub on Apr 21, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses dynamic context injection (shell commands prefixed with an exclamation mark) to execute local find operations when the skill is loaded. This is used solely for discovering the file paths of agent definitions and local settings for the audit task. Evidence: !find .claude/agents -name ".md" -not -name "settings"and!find .claude/agents -maxdepth 1 -name "settings.local.json" -type f in SKILL.md.
  • [SAFE]: The skill is configured for a read-only audit purpose. It lacks network communication tools or domains, and its allowed-tools list (Glob, Read, TodoWrite) is appropriate for its stated purpose of analysis and reporting without system modification.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 21, 2026, 01:17 AM