The Agent Skills Directory

[COMMAND_EXECUTION]: The skill's primary function is to remote-control terminal sessions by sending keystrokes and executing commands via tmux, including starting new sessions and sending input to panes.
[COMMAND_EXECUTION]: The documentation includes examples of using coding agents with --yolo or --full-auto flags, which promotes autonomous execution without user confirmation.
[DATA_EXPOSURE]: The skill uses tmux capture-pane to read terminal output and history, which can lead to the exposure of sensitive data displayed in the terminal by interactive processes.
[PROMPT_INJECTION]: By reading and reacting to output from terminal panes (e.g., using wait-for-text.sh), the skill introduces a surface for indirect prompt injection where malicious output from a compromised process could influence subsequent agent actions.

tmux