scillm

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly accepts and forwards arbitrary remote content (e.g., SKILL.md and SCILLM_PAVED_PATH_CONTRACT.md show analyze_image/analyze_image_json and a curl multimodal example using external image URLs like https://picsum.photos), meaning the agent ingests untrusted public third‑party content (remote images/URLs) that the LLM will read and could materially influence subsequent outputs or actions.