Skills
Navigate
skills/jahfaliabdulrahman-dev/hermes-skills/repo-front-door/Gen Agent Trust Hub

repo-front-door

Pass

Audited by Gen Agent Trust Hub on Oct 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The script scripts/render_card.py uses subprocess.run to launch a local browser (Chrome or Chromium) in headless mode. This is used to render HTML templates into social preview images.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes repository data and renders HTML templates using headless Chrome with the --allow-file-access-from-files flag. This configuration allows the browser to access local files via the file:// protocol, which is used legitimately here for loading fonts and images, but represents an attack surface if the input HTML were malicious.\n
  • Ingestion points: Data from the repository via gh CLI and local file reads.\n
  • Boundary markers: Instructions are provided in plain text without specific delimiters for untrusted data.\n
  • Capability inventory: File system write access and execution of a local browser via subprocess.run in scripts/render_card.py.\n
  • Sanitization: No explicit sanitization of repository data before inclusion in the HTML template is described.\n- [EXTERNAL_DOWNLOADS]: The skill recommends using shields.io for badges and jsdelivr.net for content delivery of repository assets in the README.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 3, 2026, 04:02 AM