skills/diego3g/craft/craft-states/Gen Agent Trust Hub

craft-states

Pass

Audited by Gen Agent Trust Hub on Nov 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides instructions for the agent to execute shell commands, specifically using headless Chrome or Playwright, to render and capture screenshots of design boards. This is an intended function for visual verification of HTML-based mockups.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes external design data from multiple sources, which constitutes a potential attack surface for indirect prompt injection.\n
  • Ingestion points: The agent reads and interprets HTML files, CSS stylesheets, design system tokens, and Figma component nodes across its various adapters.\n
  • Boundary markers: The skill does not currently include explicit instructions for the agent to ignore or isolate potentially malicious instructions embedded within these design assets.\n
  • Capability inventory: The skill enables shell command execution for rendering screenshots and provides tools for modifying design nodes via MCP adapters (e.g., Paper MCP).\n
  • Sanitization: There is no defined process for sanitizing or validating the content of design files before the agent processes them.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 15, 2026, 11:18 PM