openai-api

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

• Third-party content exposure detected (high risk: 0.70). The skill includes examples that send external URLs and web content to the model (e.g., the "Vision (GPT-4o)" -> "Image via URL" examples and the Function Calling section's "search_web" tool), which means the agent can ingest and interpret arbitrary third-party/user-provided content from the open web as part of its workflow.