The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of networking and system administration commands to establish tunnels and proxies. This includes ssh, socat, netsh, and the ip command suite.
[EXTERNAL_DOWNLOADS]: Instructions are provided for downloading pivoting tools like Ligolo-ng and Chisel to compromised hosts using standard utilities such as wget, curl, certutil, and Invoke-WebRequest.
[REMOTE_CODE_EXECUTION]: The skill describes procedures for executing downloaded binaries on pivot hosts to establish remote access and networking capabilities. This is the primary intended function of the skill.
[DATA_EXFILTRATION]: The skill documents techniques for routing network traffic through various tunnels (SSH, DNS, ICMP, HTTP), which represents a technical capability for data exfiltration during security assessments.
[PROMPT_INJECTION]: There is a potential for indirect prompt injection as the skill instructs the agent to process output from external network tools and pivot hosts while maintaining capabilities like command execution.
Ingestion points: Output from tools like ssh, nmap, chisel, and ligolo-ng (SKILL.md).
Boundary markers: No explicit delimiters or warnings are used for tool output processing.
Capability inventory: Shell command execution (ssh, socat, etc.) and file writing to the engagement/evidence/ directory.
Sanitization: No input validation or sanitization of external data is specified.

pivoting-tunneling