claude2figma
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill documentation provides instructions for users to install sub-skills (
figma-preflight,component-rules,figma-style-binding,reference-interpreter) from a remote repository (senlindesign/claude2figma) using thenpx skills addcommand or the Claude Code plugin marketplace. This is a standard installation procedure for catalogue entries. - [INDIRECT_PROMPT_INJECTION]: The
reference-interpretercomponent is designed to ingest external data including screenshots, URLs, and descriptions to generate design briefs, representing a vulnerability surface where untrusted content could influence agent behavior. - Ingestion points: SKILL.md specifies that the
reference-interpreterskill reads screenshots, URLs, and text descriptions from the user or external sources. - Boundary markers: None mentioned in the provided documentation; instructions do not explicitly specify markers for separating external content from system instructions.
- Capability inventory: The tools interact with the Figma MCP server for searching design systems, creating component instances, and binding styles. No dangerous capabilities such as arbitrary command execution or unauthorized file system access are described.
- Sanitization: Not present in the provided documentation.
Audit Metadata