Agent skill
kitty-remote
Control kitty terminal via remote protocol and kitty-test-harness; launch panels/windows, send keys, capture screen, manage sockets, and debug remote-control failures when testing TUIs.
Stars
163
Forks
31
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/development/kitty-remote
Metadata
Additional technical details for this skill
- triggers
-
{ "type": "domain", "keywords": [ "kitty", "kitty test" ], "priority": "high", "enforcement": "suggest" }
SKILL.md
Kitty Remote Control
Drive kitty instances via the remote control protocol and kitty-test-harness.
Prerequisites
- Ensure kitty is installed and on PATH.
- Enable remote control (
allow_remote_control=yesin kitty.conf or via-o). - Run under an active Wayland or X11 session.
Quick start
- Launch kitty with a unix socket (
--listen-on unix:/abs/path.sockorKITTY_LISTEN_ON). - Wait for
@ lsto succeed before sending input. - Send input with
send-textand capture output withget-text --extent screen. - Close with
close-windowand remove the socket.
Harness usage
- Prefer
kitty +kitten panelon Wayland; fall back to windows on X11/WSL - Use absolute socket paths; delete stale sockets before launch
- Send
ESCto normalize modal state before:commands - Socket disappearance = app exit/crash; capture with
@ ls,get-text --ansi, logs - Source scripts/kitty-harness.sh for ready-to-use harness
References
- harness-notes.md - internals, failure signatures, debug flow
- escape-codes.md - byte-level key sequences and modifiers
Didn't find tool you were looking for?