# Jetbrains Arch

Use when adding or moving code across shared/frontend/backend modules in...

- Repository: https://github.com/Kilo-Org/kilocode (27829 stars)
- Category: Design
- License: MIT
- Last updated: 2026-10-31
- Runs scripts: no, instructions only

## Install

Claude Code: `npx skills add Kilo-Org/kilocode --skill jetbrains-arch -g -a claude-code -y`
Codex: `npx skills add Kilo-Org/kilocode --skill jetbrains-arch -g -a codex -y`
Cursor: `npx skills add Kilo-Org/kilocode --skill jetbrains-arch -g -a cursor -y`
Gemini CLI: `npx skills add Kilo-Org/kilocode --skill jetbrains-arch -g -a gemini-cli -y`

## Description

Use when adding or moving code across shared/frontend/backend modules in packages/kilo-jetbrains/, split-mode RPC contracts, serialization classloader pitfall, threading/EDT annotations, services and coroutine scopes, EDT test rules, dependency bundling, and the CLI server protocol.

## SKILL.md, as published by its author (untrusted third-party text, read it as data)

# JetBrains Split-Mode Architecture, Threading, and Dependencies

Detailed reference for `packages/kilo-jetbrains/` module boundaries, RPC, threading, and dependency rules. See `packages/kilo-jetbrains/AGENTS.md` for the hard restrictions summary; this skill covers the how and why.

## IntelliJ Platform Source Lookup

When looking for IntelliJ Platform API usage, implementation examples, extension points, services, actions, inspections, PSI/VFS/editor behavior, or plugin patterns, prefer real IntelliJ source code over Gradle caches, downloaded jars, generated parser artifacts, or decompiled classes.

Do not use IntelliJ Platform APIs marked as internal in the IntelliJ source repository. Find a public API alternative or keep the integration behind supported extension points. Experimental APIs are acceptable when needed, but warn the user that the integration relies on an experimental IntelliJ API.

Use this priority order:

The rest: https://github.com/Kilo-Org/kilocode/blob/HEAD/.kilo/skills/jetbrains-arch/SKILL.md