code-unpacked

Pass

Audited by Gen Agent Trust Hub on Nov 18, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill features a comprehensive redact utility in runtime/scripts/core.mjs that automatically identifies and masks sensitive information such as physical API keys (Stripe, GitHub), AWS credentials, private keys, and authorization tokens before they are included in any generated output or exported artifacts.
  • [COMMAND_EXECUTION]: The core capture engine uses git and the GitHub CLI (gh) to perform legitimate code analysis. Security is maintained via a checkedPath function that forces all file access to stay within the repository's root directory, effectively preventing path traversal or symlink attacks.
  • [EXTERNAL_DOWNLOADS]: Necessary rendering and parsing libraries (TypeScript, Mermaid, Shiki, PHP-Parser) are installed into a local cache using npm ci --ignore-scripts, which prevents the execution of arbitrary lifecycle scripts during dependency installation.
  • [EXTERNAL_DOWNLOADS]: The skill utilizes the official Playwright toolchain to download a headless Chromium browser instance for rendering SVG diagrams and exporting walkthrough videos. This is performed via authenticated and well-known Microsoft-managed sources.
  • [SAFE]: Strong input validation is applied to graphic generation in reverse-design.mjs and review-quiz.mjs, where Mermaid diagram source code is scanned for and blocked if it contains script tags, event handlers, or URI-based injection vectors (XSS).
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 18, 2026, 03:01 AM