superteam-non-interactive

Pass

Audited by Gen Agent Trust Hub on May 19, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a 'fail-fast' contract for CI/CD environments, specifically instructing the agent to halt rather than guess when required input is missing.
  • [SAFE]: It utilizes standard environment variables (e.g., SUPERTEAM_ALLOW_PUBLISH) as explicit authorization gates before performing sensitive operations like pushing to a repository or opening pull requests.
  • [SAFE]: The instructions explicitly forbid skipping security-critical steps, such as adversarial design reviews, even when running in a headless or automated mode.
  • [SAFE]: All referenced resources are sibling files within the local directory structure (skills/superteam/), and no external network dependencies or remote scripts are invoked.
Audit Metadata
Risk Level
SAFE
Analyzed
May 19, 2026, 06:30 AM
Security Audit — agent-trust-hub — superteam-non-interactive