The Agent Skills Directory

[SAFE]: The skill uses platform-specific tools for integration and enforces a secure asset persistence policy by requiring the agent to move temporary Figma URLs to permanent storage using upload_asset.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from Figma node trees which is then interpolated into React code, creating a potential attack surface. \n
Ingestion points: The figma_get_file_nodes tool retrieves text nodes and properties from external Figma files.\n
Boundary markers: No explicit instructions are provided to the agent to ignore or escape instructions embedded within design text nodes.\n
Capability inventory: The agent has the capability to write and modify site code via edit_site_page and upload content via upload_asset.\n
Sanitization: No explicit sanitization or escaping steps for text content are defined before code generation.

figma