abuselpdb

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

• Third-party content exposure detected (high risk: 0.70). The skill explicitly directs the agent to fetch and act on data from the public AbuseIPDB API (see "Popular actions" like "Get Reports for IP" and the "Proxy requests" section showing membrane request CONNECTION_ID /path/to/endpoint), which returns user-generated/untrusted reports that the agent is expected to read and which can materially affect decisions (e.g., blocking IPs).