Skills
Navigate
skills/srinitude/skills/visual-design-system-extractor/Gen Agent Trust Hub

visual-design-system-extractor

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill performs network operations to retrieve metadata from trusted and well-known services:
  • scripts/google_fonts_api.py fetches font metadata from Google's official fonts API (fonts.google.com).
  • scripts/sync_mise_primitives.py downloads JSON schemas for the mise development tool from its official repository on GitHub (raw.githubusercontent.com/jdx/mise).
  • [COMMAND_EXECUTION]: The skill uses the subprocess module to perform necessary orchestration and testing tasks:
  • scripts/run_agentic_request.py executes specified runner tools to facilitate model-owned requests.
  • scripts/screenshot_preview.py utilizes the Playwright library to launch a headless browser for capturing visual evidence of the design system.
  • scripts/sync_mise_primitives.py runs local commands like mise --version to verify environment state.
  • [INDIRECT_PROMPT_INJECTION]: The skill provides an interface for processing external design references (images and site URLs). It mitigates injection risks through structured data contracts (YAML schemas), deterministic validation gates, and instructions that explicitly require the agent to ground its findings in visible evidence rather than speculative description.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 02:01 AM