face-swap

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

• Third-party content exposure detected (medium risk: 0.30). The face-swap skill’s runtime flow classifies the user intent and then invokes runcomfy run <model_id> with --input containing user-supplied image_url/identity_url/audio/video URLs; the CLI later fetches those untrusted reference assets (potentially including their embedded text/media metadata) to generate output.