agent-comms

The documented skill provides useful debugging coordination features but implements high-risk capabilities: arbitrary input injection into other tmux sessions (command execution) and reading full terminal buffers (exposes secrets). The documentation omits critical safeguards (authentication, consent, filtering, logging) and the actual script source is not provided, preventing full trust. Recommendation: do not deploy without reviewing the script source, adding access controls and confirmations, and restricting usage to trusted environments. Use as-is only with caution and strong operational controls.