decoding-search-behavior

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly states it will analyze search engine results pages (SERPs) and "listen to online communities like Reddit, Quora, and industry-specific forums," which are untrusted public third-party sources the agent would read and interpret as part of its workflow.