roast-my-ui

Pass

Audited by Gen Agent Trust Hub on Jul 3, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted external data such as component code and content from live URLs, creating a surface for indirect prompt injection where malicious instructions could be embedded in the design assets.
  • Ingestion points: Ingests HTML/CSS component code, live URL content via browser tools, and Figma data via MCP (SKILL.md).
  • Boundary markers: There are no instructions to use delimiters or ignore embedded natural language instructions within the code or content being analyzed.
  • Capability inventory: The skill executes a bundled script (scripts/color.py), can write a file (roast.md) to the local directory, and utilizes browser and Figma MCP tools (SKILL.md).
  • Sanitization: No sanitization or escaping of the ingested data is described before it is processed by the agent.
  • [COMMAND_EXECUTION]: The skill involves the execution of a bundled Python script (scripts/color.py) to verify color contrast ratios. The agent is instructed to pass color values and labels extracted directly from user-provided code as arguments to this script.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 3, 2026, 03:42 AM
Security Audit — agent-trust-hub — roast-my-ui