The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides templates for the agent to execute shell commands using tmux (e.g., tmux new-session, send-keys, capture-pane) to interact with and inspect local processes.
[DYNAMIC_EXECUTION]: The skill instructs the agent to generate and run temporary Python and Bash scripts at runtime. The Python template uses pty.openpty() and subprocess.Popen() to create a pseudo-terminal for driving interactive applications.
[INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection as it captures output from external processes (tmux capture-pane, os.read(master_fd, 4096)) and processes it to determine the next action.
Ingestion points: Terminal output is captured via tmux and Python PTY buffers and printed/evaluated by the agent.
Boundary markers: None present in the provided templates to distinguish between process output and agent instructions.
Capability inventory: The agent is granted capabilities to execute shell commands, spawn subprocesses, and manage local terminal sessions.
Sanitization: No sanitization or filtering of the captured terminal output is implemented before it is processed by the agent.

control-cli