figma-code-workflows
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external, untrusted content such as HTML documents and project files, which could potentially contain embedded instructions meant to influence the agent's behavior during the import process.
- Ingestion points: Data enters the agent's context through local HTML files, project archives, and the content of running applications (rendered pages).
- Boundary markers: While the instructions emphasize clarifying user intent and validating the project locally, they do not prescribe the use of explicit delimiters or instructions to ignore embedded commands within the imported data.
- Capability inventory: The skill utilizes tools capable of reading local file systems, converting HTML structures, and uploading data to external Figma services (e.g.,
html_to_figma,send_code_to_figma). - Sanitization: The instructions do not specify a requirement for the agent to sanitize or filter the content of the files or application data before processing them with the integration tools.
Audit Metadata