google-search

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly instructs using the WebSearch tool to fetch and read content from open/public third-party sites (e.g., reddit.com, news.ycombinator.com, stackoverflow.com and other public web pages), which are untrusted user-generated sources that the agent is expected to interpret during its workflow.