review-quiz

Pass

Audited by Gen Agent Trust Hub on Nov 18, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a robust secret redaction mechanism via the redact function in runtime/scripts/core.mjs to prevent the exposure of API keys, private keys, and other credentials in generated quiz artifacts.
  • [SAFE]: It enforces a strict exclusion list through the excluded function in runtime/scripts/core.mjs to prevent the agent from accessing sensitive directories such as node_modules, .git, or environment files like .env.
  • [SAFE]: The instructions in runtime/guide.md explicitly follow the principle of data/instruction separation, directing the agent to treat all ingested source code and PR comments as data rather than instructions.
  • [SAFE]: External dependencies, including Mermaid, Playwright, and Shiki, are managed through standard package registries (NPM) using npm ci in a managed runtime directory, which is a safe practice for specialized rendering tools.
  • [SAFE]: Network operations and PR interactions are performed through the authenticated GitHub CLI (gh), which ensures that actions are restricted to the repository context and the user's existing permissions.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 18, 2026, 03:00 AM