btc-bottom-model

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

• Third-party content exposure detected (high risk: 1.00). This skill explicitly instructs the agent to use web_search and ingest live data from public third‑party sources (Execution Steps 1–3 and the Analysis Framework), naming social platforms like Twitter/Reddit/Telegram and public on‑chain sites (Glassnode, CryptoQuant, Alternative.me), and requires the agent to read and act on that content to form trading recommendations, which exposes it to untrusted user-generated content that could enable indirect prompt injection.