The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill is configured to connect to the official Figma MCP server at https://mcp.figma.com/mcp. This is a well-known service and the connection is a requirement for the skill's primary functionality.
[PERSISTENCE]: In references/figma-mcp-config.md, the skill provides instructions for persisting the FIGMA_OAUTH_TOKEN by adding export commands to shell profiles like ~/.bashrc or ~/.zshrc. While this is a standard method for managing development environment variables, it involves modifying files that govern shell persistence.
[INDIRECT_PROMPT_INJECTION]: The skill features an indirect prompt injection surface because it ingests external design data to drive code implementation decisions.
Ingestion points: Untrusted design data enters the agent's context through tools like get_design_context and get_metadata described in SKILL.md and references/figma-tools-and-prompts.md.
Boundary markers: There are no explicit instructions for the agent to use delimiters or ignore potentially malicious instructions embedded within the Figma node data.
Capability inventory: The agent is instructed to translate design data into production-ready React and Tailwind code, which involves file-writing capabilities in the user's project.
Sanitization: The skill lacks defined validation or sanitization protocols for the structured design representation fetched from the Figma API.

figma