Wireshark Network Traffic Analysis

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

• Third-party content exposure detected (high risk: 0.70). This skill’s required workflow ingests PCAP/PCAPNG contents (Phase 1 “Open PCAP File” and subsequent stream/statistics/security analysis), which may include outsider-authored payloads from arbitrary network traffic the operator didn’t author, and those packet payloads can be rendered as readable text during “Follow > TCP Stream” / “View reconstructed conversation”.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (medium risk: 0.60). The skill explicitly requires administrator/root privileges for live packet capture (and references obtaining server private keys), which could push an agent to request or use elevated access to the host and thereby compromise machine state.