venice-ai-api

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly supports fetching and ingesting open-web content—e.g., references/chat-completions.md documents enable_web_search/enable_web_scraping (with an example that the model will "Summarize this page: https://example.com/article") and multimodal inputs accept http/https image/video URLs—so the agent can read arbitrary, untrusted third‑party pages as part of its workflow.