What problem does it solve? Operating WezTerm panes and tabs manually while running an AI agent is error-prone: commands can land in the wrong pane, destructive actions can hit unintended targets, and command results go unverified. This Skill provides a safe, ID-explicit workflow for controlling WezTerm through wezterm cli. ## Core Features & Use Cases - Pane and tab management: Split, spawn, move, resize, zoom, focus, and close panes and tabs using explicit --pane-id and --tab-id targeting. - Remote command execution with verification: Send commands to other panes with send-text --no-paste, then confirm results via get-text or marker-based polling for long-running tasks. - Safety guardrails: Always inspects the current layout with wezterm cli list --format json first, restricts kill-pane to self-created panes, and requires confirmation before destructive commands. - Use Case: Ask the agent to split a pane, run a build in it, and report the exit status—it will resolve the target pane ID, send the command with a completion marker, poll the output, and report the result. ## Quick Start Ask the agent to split the current WezTerm pane to the right and run your test suite in the new pane, then report the results.