The Agent Skills Directory

[COMMAND_EXECUTION]: Provides functional reverse shell templates for establishing remote access to a system, including standard bash redirection and Python pty module spawns.
[PRIVILEGE_ESCALATION]: Contains extensive instructions for elevating user privileges on Linux and Windows systems by exploiting SUID binaries, sudo misconfigurations, Docker group memberships, and utilizing exploitation tools like PrintSpoofer and GodPotato.
[PERSISTENCE]: Details multiple methods for maintaining unauthorized long-term access to compromised systems, such as modifying system initialization scripts, scheduled tasks, registry run keys, and SSH authorized keys.
[CREDENTIALS_UNSAFE]: Instructs the agent on how to locate and harvest sensitive authentication data from shell history files, configuration files (e.g., .git-credentials, .my.cnf), and system memory dumps (e.g., LSASS via procdump).
[DATA_EXFILTRATION]: Describes techniques for establishing command-and-control (C2) channels and data exfiltration tunnels using HTTPS, DNS, Domain Fronting, and legitimate third-party services like Slack, Telegram, or GitHub.
[DYNAMIC_EXECUTION]: Mentions techniques for executing code directly in memory to avoid disk-based detection, including the use of memfd, DDexec, and Donut for Windows and Linux environments.

post-exploitation