The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill includes a utility (scripts/emit-placement.mjs) that generates JavaScript code for the Figma Plugin API. This code is constructed from a template and populated with layout coordinates and node IDs from a local plan. It is intended to be executed via the official use_figma tool to automate canvas operations.
[EXTERNAL_DOWNLOADS]: The scripts/verify-freshness.mjs script fetches a README file from Figma's official community repository on GitHub. This is used to verify that the skill's internal router table is synchronized with the latest community-contributed Figma skills. The fetched content is parsed as text and is not executed.
[COMMAND_EXECUTION]: The skill orchestrates several internal Node.js scripts (plan-layout.mjs, stage-assets.mjs, verify-board.mjs) using the Bash tool to perform geometric calculations and asset management tasks.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from multiple sources, including Figma canvas objects, local image file headers, and remote markdown files from GitHub. To mitigate potential injection risks, the skill implements a mandatory vetting gate: the emit-placement.mjs script refuses to generate placement code for any asset not explicitly marked as vetted: true by the agent or user.

figma-ops