openrouter-skill

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly accepts and invokes LLM analysis on arbitrary external URLs (e.g., "Vision Analysis" and "Multimodal Support" examples using image_url with chart_url, chart1_url/chart2_url and analyze_multiple_images), so the agent will fetch and interpret untrusted public third‑party content supplied via URLs.