sc-gemini-imagegen

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly supports and documents "Google Search Grounding" (SKILL.md
• "Advanced Features" example) and the library (gemini_images._build_config and generate) can enable a google_search tool that fetches real-time public web content which the model will read and use to influence image generation, so it ingests untrusted third-party web content as part of its workflow.