The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill documents commands that download and execute scripts directly from remote sources. Specifically, it includes a pattern that pipes a remote script from GitHub into the shell (curl ... | sh) and instructions for downloading and running payloads from arbitrary, attacker-controlled IP addresses.
[DATA_EXFILTRATION]: Provides explicit examples for archiving sensitive system directories, such as /etc and /home, and transmitting the resulting data to a remote listener using nc (netcat).
[PRIVILEGE_ESCALATION]: Includes extensive documentation on techniques to gain unauthorized root access, including abusing sudo permissions (e.g., sudo find ... -exec /bin/bash), exploiting SUID/SGID binaries, targeting kernel vulnerabilities, and leveraging Linux capabilities.
[PERSISTENCE]: Outlines multiple methods for maintaining access to a compromised system, such as creating backdoor user accounts, injecting SSH public keys into the root directory, and setting up malicious cron jobs to execute reverse shells periodically.
[COMMAND_EXECUTION]: Contains a wide array of dangerous shell one-liners for generating reverse shells (via bash, python, and netcat), stabilizing unauthorized TTY shells, and deleting system logs to conceal attacker activity.
[EXTERNAL_DOWNLOADS]: Instructs the user to clone an external repository (HIMANSHUSHARMA20/Linux-for-a-Pentester) and download tools from unverified third-party GitHub releases.

linux-pentester-notes