image-mining

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly supports ingesting images from arbitrary web URLs ("phone_mining: capture: sources: ... url: 'Import from web'") and the pipeline/LLM vision is expected to read and interpret those images as part of its mining workflow, exposing the agent to untrusted third‑party (user/web) content.