The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user or agent to install the omniroute package from the npm registry. This package is not from a recognized trusted organization.
[COMMAND_EXECUTION]: The skill provides instructions for executing various CLI commands to manage network tunnels, including creating, listing, and rotating connections.
[PROMPT_INJECTION]: The skill contains an attack surface for indirect prompt injection via the processing of external data.
Ingestion points: Subcommands such as tunnel logs, tunnel status, and tunnel info pull data from external network providers and log streams into the agent's context.
Boundary markers: There are no markers or safety instructions provided to help the agent distinguish tool output from potential malicious instructions embedded in the logs.
Capability inventory: The skill has the capability to create and stop network tunnels, which could be misused if the agent obeys instructions found in untrusted data.
Sanitization: No validation or sanitization of the data returned by the subcommands is described.

cli-tunnel