gemini-omni-flash-api

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

• Third-party content exposure detected (high risk: 0.75). The required runtime path takes outsider-authored prompt free text (e.g., from --batch JSON / --prompts-file or the CLI prompt argument) and inserts it directly into generate_video.py as input_parts ("type": "text", "text": prompt) which is then sent to client.interactions.create, so any non-user text can become LLM-context input (indirect prompt injection risk).