create-structure
Pass
Audited by Gen Agent Trust Hub on May 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill relies on figma_execute and use_figma to run logic within the Figma API sandbox, which is standard for design automation.
- [REMOTE_CODE_EXECUTION]: JavaScript code is dynamically generated and executed within the Figma environment to handle complex measurements.
- [EXTERNAL_DOWNLOADS]: The skill retrieves design templates from the Figma cloud using the importComponentByKeyAsync method.
- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it ingests data from uspecs.config.json and agent-structure-instruction.md (Ingestion points) without explicit boundary markers (Boundary markers) while maintaining extensive Figma API control (Capability inventory: figma_execute) and lacking sanitization (Sanitization).
Audit Metadata