figma-local-dev-loop

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Utilizes Bash commands for initializing Node.js projects, running build scripts with esbuild/concurrently, and fetching design snapshots via curl.
  • [EXTERNAL_DOWNLOADS]: Downloads development dependencies from the public npm registry and fetches design metadata from official Figma REST API endpoints (api.figma.com).
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of external data from Figma API responses. If the agent later processes this data (e.g., layer names or custom metadata), a maliciously crafted Figma file could serve as a vector for indirect prompt injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 03:45 AM
Security Audit — agent-trust-hub — figma-local-dev-loop