roboflow-api-reference

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

• Third-party content exposure detected (high risk: 0.70). This skill's inference and REST examples (inference.md and rest-api.md) explicitly accept and POST arbitrary image URLs (the "image" query/body parameter and workflow inputs) so the agent will fetch/ingest untrusted, user-hosted content from the open web which can materially influence model outputs and downstream decisions.