The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill involves the ingestion of board metadata via the get_figjam tool. This results in the processing of potentially untrusted text (e.g., sticky notes or labels) during runtime.
Ingestion points: get_figjam tool in SKILL.md used for inspecting board content.
Boundary markers: The SKILL.md contains an Anti-Patterns section explicitly telling the agent not to treat external content or tool responses as trusted instructions.
Capability inventory: The skill allows modification of board elements and asset uploads via use_figma and upload_assets.
Sanitization: No automated sanitization is provided in the JavaScript snippets, relying on agent reasoning.
[DYNAMIC_EXECUTION]: The skill provides numerous JavaScript templates for the agent to follow when using the use_figma tool. This is the intended behavior for an API-focused skill and allows the agent to perform complex operations like batch modifications and layout reflowing.

figma-use-figjam