novita-ai

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md and referenced docs (references/image-api.md, references/llm-api.md, references/audio-api.md) explicitly show endpoints that accept arbitrary external URLs or input image/audio URLs (e.g., FLUX Kontext "images" input URLs, multimodal message image_url, GLM ASR "file" URL), so the agent can ingest untrusted third‑party content that may influence LLM inputs and subsequent actions.