design-md-chrome-extractor

Pass

Audited by Gen Agent Trust Hub on May 16, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references a GitHub repository (https://github.com/bergside/design-md-chrome.git) for local development and installation.
  • [COMMAND_EXECUTION]: The documentation includes commands for cloning repositories, changing directories, and running test suites using Node.js.
  • [PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted data from external websites to generate design guidelines for AI agents, which presents an indirect prompt injection surface.
  • Ingestion points: Style extraction from arbitrary URLs (described in SKILL.md functions).
  • Boundary markers: No delimiters or ignore-instruction warnings are present in the output templates.
  • Capability inventory: Generates DESIGN.md and SKILL.md files for use in coding environments.
  • Sanitization: Logic focuses on visual token normalization rather than detecting or escaping malicious text strings in CSS properties.
Audit Metadata
Risk Level
SAFE
Analyzed
May 16, 2026, 07:52 PM