universal-learner

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly ingests arbitrary prompts provided by users (e.g., "学习这个Prompt: [粘贴完整Prompt]" and batch import of prompts) and stores full original content (visual_reference field), so it will read and process untrusted, user-supplied third‑party content that could carry indirect prompt-injection instructions.