KHA Entertainment's collection of AI coding tools, including the Grok Swarm multi-agent plugin powered by xAI Grok 4.20.
npx claudepluginhub khaentertainment/grok-multiagent-pluginMulti-agent intelligence powered by Grok 4.20 via OpenRouter. Use for codebase analysis, refactoring, code generation, and complex reasoning.
KHAEntertainmentDual-Platform: OpenClaw + Claude Code
Give any AI coding agent access to Grok 4.20's 4-agent swarm with ~2M token context.
You've been building with AI coding agents for a while now. They're great — they can write features, refactor modules, analyze codebases. But there's always been this ceiling. The models they run on are designed for single-turn conversations.
Enter Grok 4.20 Multi-Agent Beta.
It's different. Instead of one model responding, it's four agents coordinating in real-time. An orchestrator, specialists, critics — all working together to break down your request and reason through it from multiple angles. It can hold ~2M tokens of context — that's entire codebases in a single request.
The Problem:
Grok 4.20 is groundbreaking, but it doesn't play nicely with current coding tools. Claude Code doesn't have a Grok integration. OpenClaw's tooling system doesn't support multi-agent swarms. If you wanted to use Grok, you'd have to hack together custom scripts or modify your platform's core components. Not ideal.
The Solution:
This plugin bridges that gap. It makes Grok 4.20 available as a tool that any agent in Claude Code or OpenClaw can call. No core modifications, no hacking — just install and go.
Now when your agent needs deep codebase analysis, large-scale refactoring, or complex reasoning, it can delegate to Grok's swarm and get back the kind of coordinated, multi-perspective thinking that single models can't deliver.
When write_files=true, Grok parses code blocks for filename annotations and writes them directly to disk, returning only a compact summary instead of the full response.
Fenced code blocks with path in the language tag:
```typescript:src/auth/login.ts
export function login() { ... }
```
Fenced code blocks with // FILE: marker:
```typescript
// FILE: src/auth/login.ts
export function login() { ... }
```
const result = await tools.grok_swarm({
prompt: "Write a FastAPI auth module with JWT",
mode: "code",
write_files: true,
output_dir: "./src"
});
// Returns: "Wrote 3 files to ./src
// src/auth.py (1,234 bytes)
// src/jwt_utils.py (567 bytes)
// src/middleware.py (890 bytes)"
Grok can generate ~350K token responses. Without file writing, that floods your orchestrator's context window. With file writing, you get a brief summary and the files on disk.
Choose the method that fits your platform:
npm install -g @openclaw/grok-swarm
# Set up API key
./scripts/setup.sh
# Add the marketplace
/plugin marketplace add https://github.com/KHAEntertainment/grok-multiagent-plugin
# Install the plugin
/plugin install grok-swarm@khaentertainment
Then run /grok-swarm:setup inside Claude Code — an OAuth browser flow will
authorize your OpenRouter account without exposing your API key in-context.
clawhub install grok-swarm
git clone https://github.com/KHAEntertainment/grok-multiagent-plugin.git
cd grok-multiagent-plugin
# Auto-detect and install
./install.sh
# Or install for specific platform
./install.sh claude # Claude Code only
./install.sh openclaw # OpenClaw only
./install.sh both # Both platforms
For detailed instructions for each method, see INSTALL.md.
/grok-swarm:analyze Review the security of my auth module
/grok-swarm:refactor Convert these callbacks to async/await
/grok-swarm:code Write a FastAPI endpoint for user registration
/grok-swarm:reason Compare microservices vs monolith for this project
# After ClawHub or git install, add to openclaw.json:
openclaw gateway restart
Then use in your agent: