The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to fetch and process design data from the Figma API using get_figma_data. This data, which includes node names, component descriptions, and layout metadata, is authored by external users and could contain malicious instructions designed to hijack the agent's logic during design-to-code tasks.
Ingestion points: Design data and metadata retrieved via get_figma_data as described in SKILL.md.
Boundary markers: None present. The instructions do not define delimiters or provide warnings to the agent about ignoring potential instructions within the design data.
Capability inventory: The skill uses get_figma_data to read external structures and download_figma_images to write files to a local path (referenced in SKILL.md).
Sanitization: None present. The content is processed directly to generate code or extract tokens without filtering.
[EXTERNAL_DOWNLOADS]: The skill references external GitHub repositories for the Figma MCP server implementation and community-maintained cached versions.
Evidence: Links to https://github.com/GLips/Figma-Context-MCP in SKILL.md and https://github.com/pactortester/figma-context-mcp-cached in references/rate-limit-recovery.md.

figma-context-mcp