using-superpowers

Pass

Audited by Gen Agent Trust Hub on Jul 14, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill uses imperative language such as 'ABSOLUTELY MUST' and 'not negotiable' to ensure the agent prioritizes external skill files over its internal planning logic. This is intended to enforce a specific operational workflow within the Codex CLI environment rather than for malicious bypass purposes.
  • [DATA_EXPOSURE]: The skill directs the agent to locate and read files from './.agents/skills/' and '~/.agents/skills/'. While this involves file system access to ingested data, these are standard directories for skill discovery in the specified environment and do not constitute unauthorized data exposure.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 14, 2026, 06:45 AM