mgrep

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly instructs the agent to fetch and summarize web search results using "mgrep --web --answer", pulling content from the public web (mixedbread/web store), which is untrusted third-party/user-generated content that the agent will read and interpret.