The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes tmux and subprocess.Popen to launch and interact with local command-line applications. These operations are intended for local harness building and profiling as described in the skill purpose.
[DYNAMIC_EXECUTION]: The skill generates and executes temporary shell scripts and Python scripts (using the pty and subprocess modules) to facilitate automated interactions with interactive terminal sessions.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The skill reads terminal output through tmux capture-pane and by reading from PTY master file descriptors in SKILL.md.
Boundary markers: The instructions suggest waiting for specific text patterns or prompts to delimit interactions, but these are not robust security boundaries.
Capability inventory: The skill has the ability to execute commands, send keyboard input to sessions, and interact with the local filesystem and loopback network interfaces (via Node/Bun inspectors).
Sanitization: The provided script templates do not demonstrate sanitization or escaping of the terminal output before it is processed or used to trigger subsequent actions.

control-cli