The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructs the agent to download and execute shell scripts from external sources, including piping a script from a GitHub repository directly into the shell (curl -L ... | sh). It also describes downloading files from an external IP address for execution on the local system.
[PRIVILEGE_ESCALATION]: The core functionality is dedicated to elevating privileges from a low-level user to root. It provides specific methods for exploiting kernel vulnerabilities, sudo misconfigurations (such as LD_PRELOAD), SUID binaries, system capabilities, and cron jobs.
[DATA_EXFILTRATION]: The skill includes multiple reverse shell one-liners for Bash, Python, Netcat, and Perl intended to establish outbound connections to an external IP. It also provides instructions to read sensitive system files like /etc/shadow.
[DYNAMIC_EXECUTION]: The workflow includes steps to compile C source code at runtime using gcc and execute the resulting binaries to perform kernel exploits or create SUID-enabled shells.
[COMMAND_EXECUTION]: The skill provides numerous commands that utilize standard Linux utilities like find, vim, awk, and less in ways designed to bypass security restrictions and spawn administrative shells.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process arbitrary output from the target system's shell, such as process lists and file contents. There are no boundary markers or sanitization steps defined to prevent the agent from potentially interpreting instructions embedded within that system output.

Linux Privilege Escalation