create-voice
Warn
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: MEDIUMDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill assembles and executes substantial JavaScript code blocks within the Figma environment using tools like
figma_executeanduse_figma. This runtime script generation is used to automate complex layout, styling, and annotation tasks within the design software. - [INDIRECT_PROMPT_INJECTION]: The skill processes external, user-supplied Markdown files which act as the primary source of truth for the documentation.
- Ingestion points: Data is read from user-specified
.mdfiles and associated hidden metadata blocks (render-meta). - Boundary markers: The instructions do not define clear boundaries or 'ignore' commands for instructions potentially embedded within the ingested Markdown content.
- Capability inventory: The skill leverages powerful design automation tools capable of script execution (
figma_execute) and visual data capture. - Sanitization: External content is interpolated into JavaScript execution strings without visible sanitization or escaping mechanisms.
- [COMMAND_EXECUTION]: The skill utilizes a variety of tools to programmatically interact with Figma, including searching for components, navigating file structures, and modifying local variables and styles.
Audit Metadata