midjourney-prompt-engineering
Warn
Audited by Snyk on Feb 16, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.90). The skill explicitly performs WebSearch/WebFetch and Playwright browsing of public sites (e.g., midjourney.com Style Explorer, Reddit/community pages, midlibrary.io and arbitrary URLs) and ingests those pages/images for "community research" and prompt construction, so it reads untrusted, user-generated third‑party content as part of its workflow.
Audit Metadata