The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Python subprocess module to interface with system utilities such as tshark, capinfos, ipconfig, ping, and tracert. These calls are safely constructed using lists of arguments without enabling shell execution, which prevents command injection from user-supplied targets or interface names.
[EXTERNAL_DOWNLOADS]: The documentation refers to official installation sources for Wireshark and Npcap. These are established and trusted services for network monitoring software.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from network traffic, including packet contents, service banners during port scans, and external pcap files. The skill mitigates risks by utilizing structured JSON output, truncating banner strings to 200 characters, and using robust text decoding methods to prevent agent misinterpretation of malicious text embedded in network data.

net