figma-use

Pass

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [DYNAMIC_EXECUTION]: The skill enables the agent to run raw Figma Plugin API scripts via specialized actions to perform complex batch operations and custom logic that exceed standard tool capabilities.
  • [INDIRECT_PROMPT_INJECTION]: The agent processes untrusted design data and metadata from the Figma canvas to inform its actions.
  • Ingestion points: Canvas metadata and design context are retrieved using get_design_context and get_metadata in SKILL.md.
  • Boundary markers: There are no explicit delimiters or instructions to ignore malicious content embedded within Figma layers or properties.
  • Capability inventory: The skill possesses significant modification rights, including component instantiation and arbitrary script execution in SKILL.md.
  • Sanitization: While design specifications are validated against schemas before persistence, the input pipeline lacks sanitization for the instructions derived from the canvas.
  • [DATA_EXFILTRATION]: The skill facilitates the transfer of design tokens and variables from Figma to local configuration files such as src/figma/tokens.ts and uses a WebSocket bridge on ports 9223-9232 for operational communication.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 4, 2026, 06:21 AM
Security Audit — agent-trust-hub — figma-use