using-superpowers

Pass

Audited by Gen Agent Trust Hub on May 2, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill uses highly imperative language ("ABSOLUTELY MUST", "not negotiable") to mandate a specific workflow involving tool invocation. While this is a meta-instruction aimed at overriding default agent behavior, the skill explicitly includes a safeguard stating that user instructions always take precedence over the skill's logic.
  • [COMMAND_EXECUTION]: Documentation within the reference files provides example bash commands for repository environment detection, such as identifying git directories and current branches. These are read-only, non-destructive commands provided for platform compatibility guidance.
Audit Metadata
Risk Level
SAFE
Analyzed
May 2, 2026, 12:08 AM