The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requires and extensively utilizes sudo for high-privilege operations, including modifying kernel parameters via sysctl and manipulating network routing rules with iptables.
[COMMAND_EXECUTION]: The workflow involves the execution of powerful network attack tools such as bettercap, mitmproxy, and ettercap to perform ARP, DHCP, and DNS spoofing.
[COMMAND_EXECUTION]: The agent.py script uses the scapy library to craft and send raw ARP packets, which is a core technique for network traffic redirection.
[DATA_EXFILTRATION]: The skill includes instructions to intercept and log sensitive information, such as authentication cookies and POST request data (credentials/forms), to local log files (captured_forms.log, captured_cookies.log).
[COMMAND_EXECUTION]: The skill demonstrates local script generation and execution by writing Python logic to extract_creds.py which is then executed as a plugin for mitmproxy.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it ingests untrusted network data (sniffing) which is processed and logged without explicit sanitization or validation.

conducting-man-in-the-middle-attack-simulation