The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documentation includes several command-line examples for establishing network tunnels and port forwarding. These include:
Ligolo-ng agent setup and routing.
Chisel reverse server and client SOCKS proxy configurations.
Sshuttle VPN-like tunneling.
SSH local, dynamic, and remote port forwarding (-L, -D, -R).
Socat relay for TCP port forwarding.
Windows netsh portproxy commands.
Plink.exe SOCKS proxy execution.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon data retrieved from the target environment, which creates a surface for indirect prompt injection if the output of these commands is manipulated by an attacker.
Ingestion points: The agent is instructed to read output from network discovery commands such as ip a, ipconfig, and arp -a (found in SKILL.md).
Boundary markers: None present; the agent processes network output directly to determine the next phase of the pivot.
Capability inventory: The skill facilitates complex network routing and tool execution via the host's shell.
Sanitization: No explicit sanitization or validation of the network discovery output is described.

network-pivoting-tunneling