protocol-reverse-engineering
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection through the analysis of untrusted network data.
- Ingestion points: Raw network packet payloads are ingested and printed to the session via Scapy scripts (
rdpcap,pkt[Raw].load) and tshark extraction commands documented inresources/implementation-playbook.md. - Boundary markers: The instructions do not define boundary markers or delimiters to differentiate between protocol metadata and potentially malicious payload content.
- Capability inventory: The skill enables the agent to send arbitrary network packets (
scapy.send()), perform protocol fuzzing (boofuzz), and execute local capture tools. - Sanitization: There is no evidence of sanitization or filtering to prevent embedded instructions in network traffic from influencing agent behavior.
- [COMMAND_EXECUTION]: The skill provides numerous pre-configured command-line templates for network capture and manipulation tools, including
tcpdump,wireshark,tshark, andmitmproxy. While these are standard tools for the stated purpose, they represent a capability for local system command execution.
Audit Metadata