Dynamic Pricing Intelligence Agent

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

• Third-party content exposure detected (high risk: 0.90). This skill explicitly calls APIClaw endpoints (e.g., realtime/product, products/competitors, products/search, reviews/analysis) to fetch public Amazon product data and user-generated reviews, and the SKILL.md and scripts show the agent must read and use that untrusted third‑party content to compute RAISE/HOLD/LOWER signals and take pricing actions, so third‑party instructions/data can materially influence behavior.