The Agent Skills Directory

[SAFE]: The skill serves as a high-level orchestration layer. It implements a strict parsing logic for subcommands (audit, config, init, route) and defaults to a safe usage summary if an unrecognized argument is provided.
[SAFE]: The skill configuration includes disable-model-invocation: true, ensuring that it only runs when explicitly requested by the user, which mitigates risks of accidental or automated execution.
[SAFE]: No evidence of data exfiltration, obfuscation, privilege escalation, or remote code execution was found in the skill definitions or metadata.

agent-dispatch