macos-golden-gate-design

Pass

Audited by Gen Agent Trust Hub on Sep 21, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, creating a potential surface for indirect prompt injection. This is addressed through explicit defensive instructions.
  • Ingestion points: The skill ingests data from Figma URLs, external web pages (via Firecrawl), and user-supplied screenshots or repositories as specified in the workflow section of SKILL.md.
  • Boundary markers: SKILL.md includes a specific "Source precedence" section that instructs the agent to treat all external material as untrusted and to ignore any instructions embedded within that content.
  • Capability inventory: The skill utilizes tools for Figma inspection (figma:figma-use), web browsing (ego-browser), web searching (firecrawl), and code implementation for macOS applications.
  • Sanitization: The agent is instructed to extract only design facts from sources and to report conflicts if source content attempts to override user instructions or safety rules.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 21, 2026, 06:51 PM
Security Audit — agent-trust-hub — macos-golden-gate-design