gleap

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly instructs the agent to fetch Gleap data via Membrane actions and proxy requests (e.g., "membrane action run ..." and "membrane request CONNECTION_ID /path/to/endpoint") to read Reports/Comments (user-generated feedback), so the agent will consume untrusted third‑party content that could contain executable instructions.