create-figjam-content
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: Indirect prompt injection surface detected.
- Ingestion points: The 'Read board contents' and 'Read the connection graph' snippets in
references/figjam-snippets.mdingest text content from FigJam nodes (stickies, shapes, connectors, etc.). - Boundary markers: None present. The scripts return raw text data directly from the Figma API to the agent context.
- Capability inventory: The skill possesses the ability to create and modify board elements (nodes, sections, tables, code blocks) using various snippets in
references/figjam-snippets.md. - Sanitization: No sanitization or filtering of the text content read from the board is performed before it is returned.
- [COMMAND_EXECUTION]: The skill provides multiple JavaScript snippets in
references/figjam-snippets.mddesigned to be executed via theuse_figmatool. These snippets interact directly with the Figma Plugin API to perform board operations.
Audit Metadata