anydesign

Pass

Audited by Gen Agent Trust Hub on Jun 30, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill interacts with external websites to retrieve design assets, HTML, and CSS variables. This functionality is implemented using the WebFetch tool and included scripts like scripts/extract_css_vars.py and scripts/capture_site.py, which are essential for its core purpose of design analysis.
  • [COMMAND_EXECUTION]: The instructions require the agent to execute several bundled Python scripts (capture_site.py, extract_css_vars.py, lint_design_md.py, etc.). These scripts perform automated browser tasks and data processing to assist in generating design documentation.
  • [PROMPT_INJECTION]: By ingesting and analyzing untrusted content from external URLs (HTML, CSS, and Figma metadata), the skill has an inherent attack surface for indirect prompt injection. While the agent follows structured templates, it could potentially encounter adversarial content embedded within the design data of analyzed websites.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 30, 2026, 05:32 PM
Security Audit — agent-trust-hub — anydesign