The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill relies on external design data from Figma, which is ingested through tools like get_metadata and get_design_context. This represents an attack surface where malicious content within a design (such as layer names or text objects) could potentially influence the behavior of the agent or downstream skills in the pipeline.
Ingestion points: Design metadata and assets are retrieved from the Figma API via get_metadata, get_design_context, and download_assets (referenced in the text).
Boundary markers: No specific boundary markers or 'ignore' instructions are mandated to isolate untrusted design data from the agent's logic during processing.
Capability inventory: The skill utilizes picsart_drive_upload for storage operations and picsart_scene_patch for modifying scene state.
Sanitization: The documentation does not detail any sanitization or validation processes for strings and identifiers imported from the design file before they are incorporated into the internal motion.json representation.

design-layer-import