The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests data from external Figma files to generate source code. A malicious Figma design could contain instructions in layer names or text intended to influence the code output.
Ingestion points: Design data is fetched using get_design_context and get_metadata as described in Step 2 of SKILL.md.
Boundary markers: The instructions do not define clear delimiters or warnings to ignore potentially malicious content embedded within the Figma data.
Capability inventory: The skill's primary purpose is to "Translate Figma nodes into production-ready code," which involves significant code generation and local file implementation capabilities.
Sanitization: There is no documentation of sanitization or validation logic for the data received from the Figma API before it is processed by the AI.
[COMMAND_EXECUTION]: The skill provides instructions for the agent or user to execute commands that modify the local configuration and add external tools.
Evidence: SKILL.md Step 0 includes commands such as codex mcp add figma and instructions to modify config.toml to enable remote MCP clients, which alters the security posture of the execution environment.

figma-implement-design