super-execute

Pass

Audited by Gen Agent Trust Hub on Mar 7, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill acts as a process orchestrator that enforces security and quality through recursive review loops and hard gates, preventing the implementation of unvetted code.\n- [PROMPT_INJECTION]: The skill's ingestion of external plans and PR comments constitutes an indirect prompt injection surface.\n
  • Ingestion points: The skill processes external implementation plan files and PR bot feedback retrieved via the 'gh-address-comments' skill.\n
  • Boundary markers: Prompt templates use distinct headers (e.g., 'PLAN CONTEXT:', 'YOUR TASK:') to isolate instructions from the data being processed.\n
  • Capability inventory: The orchestrated subagents have permissions for file system modification, command execution for testing purposes, and GitHub API interactions.\n
  • Sanitization: Security is maintained through mandatory multi-layered review gates (Task, Phase, and PR levels) that verify all implementation outputs against the validated plan before proceeding.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 7, 2026, 05:32 PM