The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The skill performs extensive shell operations including process management (pkill, pgrep), service management (launchctl), and network diagnostics (ssh, curl, nc, ping, lsof). These are necessary for the primary purpose of debugging SSH tunnels.
[DATA_EXFILTRATION] (SAFE): While the skill uses network-capable tools, there is no evidence of sensitive data being exfiltrated to unauthorized domains. Connectivity tests are directed at the user's specified hosts or localhost.
[PRIVILEGE_ESCALATION] (SAFE): The skill interacts with user-level LaunchAgents and process lists. It suggests using the root user for remote SSH connections as a troubleshooting step, but does not attempt local privilege escalation (e.g., sudo).
[INDIRECT_PROMPT_INJECTION] (LOW): The skill processes untrusted data from external sources.
Ingestion points: The agent reads output from ssh-tunnels logs and remote command execution results via ssh (e.g., netstat output from the remote host).
Boundary markers: No explicit delimiters or instructions are provided to the agent to treat this external data as untrusted text.
Capability inventory: The skill can execute arbitrary bash commands, manage system services via launchctl, and kill processes via pkill.
Sanitization: No sanitization or validation of the ingested log content or remote command output is performed before it is processed by the agent's logic.

ssh-tunnel-debug