fal-text-to-image

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

• Third-party content exposure detected (high risk: 0.90). The skill accepts arbitrary external image URLs/paths as inputs (see the CLI arguments like input_image / -i reference image in SKILL.md and the prepare_image_input functions in fal-image-remix, fal-image-edit, and fal-text-to-image that return or upload http(s) URLs), so untrusted third-party content (public/user-hosted images) is ingested and fed into model runs where that content can influence generation and model selection.