What problem does it solve?
This Skill enables programmatic interaction with interactive command-line interfaces (CLIs) like Python REPLs or debuggers, which are typically difficult to automate. It solves the challenge of sending input and capturing output from these tools, allowing AI agents to drive complex terminal-based workflows.
Core Features & Use Cases
- Remote Session Control: Start, manage, and kill
tmux sessions on an isolated socket, keeping agent activities separate from personal tmux configurations.
- Input Sending & Output Scraping: Send literal keystrokes to panes and capture their output, including historical context, for precise control and monitoring.
- Prompt Synchronization: Use helper scripts to wait for specific text patterns (e.g., Python prompts) before sending commands, ensuring reliable interaction with interactive tools.
- Use Case: An agent needs to debug a C program using
gdb. This Skill allows the agent to start gdb in a tmux pane, send commands like break main and run, and capture the debugger's output to analyze program state.
Quick Start
To start a new isolated tmux session for a Python REPL, use the command 'SOCKET_DIR=${TMPDIR:-/tmp}/claude-tmux-sockets; mkdir -p "$SOCKET_DIR"; SOCKET="$SOCKET_DIR/claude.sock"; SESSION=claude-python; tmux -S "$SOCKET" new -d -s "$SESSION" -n shell; tmux -S "$SOCKET" send-keys -t "$SESSION":0.0 -- 'python3 -q' Enter'.