greptile

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

• Third-party content exposure detected (high risk: 0.70). This skill integrates with Greptile via the Membrane CLI and instructs the agent to run searches/actions (e.g., "membrane action list --connectionId=CONNECTION_ID --intent 'QUERY'" and "membrane action run ...") against Greptile, a search/analytics platform for source code, which will return untrusted, third‑party repository content that the agent is expected to read and act on as part of its workflow.