figma-integration
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run a local validation script
python3 scripts/validate_tokens.py. This is a routine part of the development workflow for verifying token integrity. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data by reading Figma frames, variables, and screenshots through a Figma MCP server.
- Ingestion points: External Figma data (frames, variables, screenshots) and the local file
workflows/figma-integration.md. - Boundary markers: None identified in the provided instructions to delimit external content from instructions.
- Capability inventory: Execution of a local Python script (
scripts/validate_tokens.py) and standard file read/write operations within the repository. - Sanitization: The skill includes a validation step (
validate_tokens.py), but it is focused on token schema/structure rather than sanitizing against prompt injection instructions embedded in design data.
Audit Metadata