protocol-reverse-engineering
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted network traffic data which could contain malicious instructions. \n
- Ingestion points: Untrusted data enters the agent's context through network captures (PCAP) and live traffic analysis using
tshark,tcpdump, andscapyas shown inresources/implementation-playbook.md. \n - Boundary markers: Absent. The instructions do not define delimiters or provide warnings to the agent to disregard instructions found within captured packet data. \n
- Capability inventory: The skill provides methods for shell command execution (via
tcpdump,tshark,mitmproxy) and script execution (via Python withscapyandboofuzz). \n - Sanitization: There are no instructions for sanitizing or escaping data extracted from network packets before it is processed by the agent. \n- [COMMAND_EXECUTION]: The skill documentation includes numerous examples of shell commands for capturing and analyzing network traffic on local interfaces using
wireshark,tcpdump, andtshark. These commands interact directly with the system's network stack for traffic monitoring.
Audit Metadata