anycap-media-production

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly permits using external URLs as reference images (see "Multiple Reference Images" example with --param images='["./local-ref.png","https://example.com/style-ref.jpg"]') and the annotation workflow ingests user-provided annotation text/recordings, so untrusted third‑party content from arbitrary web URLs or collaborators is fetched and directly used to build prompts and drive generation.