What problem does it solve? Testing a terminal UI by hand is slow and error-prone, and naive automation produces false passes — sleeping instead of waiting, scrolling content that fits the window, or clicking where no link exists. This Skill teaches how to actually operate the GAIA TUI end to end through its loopback control API so test results reflect real behavior. ## Core Features & Use Cases - Full remote control: Send keys, text, and mouse events (click, scroll, double-click) to a live TUI session via authenticated HTTP endpoints, and resize the terminal before capturing. - Visual proof: Capture the current frame as SVG screenshots or replay the session as a looping animated SVG recording, with frame-level history for clipping specific actions. - State-based waiting and diagnostics: Block server-side on screen content or state conditions instead of sleeping, and read state.chat diagnostics (scroll position, mouse ownership, viewport size) to distinguish real defects from mis-aimed tests. - Use Case: Validate that a new chat feature renders correctly by launching an isolated TUI with its own daemon and throwaway memory database, sending a query, waiting for the response, and capturing an SVG recording of the session. ## Quick Start Launch the GAIA TUI with a control port and isolated GAIA_TUI_HOME, GAIA_DAEMON_HOME, and GAIA_MEMORY_DB paths, then use the control API to resize, send a query, wait for the response, and capture an SVG screenshot.
