Skip to main content

Oh My OpenCode

tip

59,187 GitHub stars — The most comprehensive agent harness for OpenCode. Previously named oh-my-opencode, now oh-my-openagent.

Overview​

Oh My OpenCode bundles everything you need for advanced OpenCode workflows into a single package: background agents, pre-built LSP/AST/MCP tools, curated agents, Claude Code compatibility, and an orchestration layer. It is the closest thing to an OpenCode distribution with batteries included.

What's Included​

ComponentDescription
Background AgentsRun agents asynchronously in the background
LSP ToolsPre-configured Language Server Protocol integrations
AST ToolsAbstract syntax tree analysis and manipulation
MCP ToolsModel Context Protocol server integrations
Curated AgentsReady-to-use specialized agents
OrchestrationMulti-agent coordination and task distribution
Claude Code CompatCompatibility layer for Claude Code skills and workflows

When to Use​

ScenarioBenefit
Power usersGet advanced features without manual configuration
Multi-agent setupsBuilt-in orchestration for complex workflows
IDE-like featuresLSP and AST tools for code intelligence
Migration from Claude CodeCompatibility layer eases transition
All-in-one configurationSingle install instead of wiring multiple plugins

Installation​

Add to your opencode.json:

{
  "plugin": ["oh-my-opencode"]
}

Or clone the repository for local development:

git clone https://github.com/code-yeongyu/oh-my-openagent.git
ln -s $(pwd)/oh-my-openagent ~/.config/opencode/plugins/oh-my-openagent

Configuration​

Basic Setup​

{
  "plugin": ["oh-my-opencode"],
  "agent": {
    "background-worker": {
      "description": "Handles background tasks delegated by the main agent",
      "mode": "subagent",
      "hidden": true
    }
  }
}

Enable Specific Features​

{
  "experimental": {
    "batch_tool": true
  }
}

Usage​

Background Agent Delegation​

The primary agent can delegate long-running tasks to background agents:

@background-worker analyze the performance impact of this refactor across all modules

LSP Code Intelligence​

With LSP tools enabled, agents can:

  • Navigate to symbol definitions.
  • Find all references across the project.
  • Get hover documentation for any symbol.
  • List symbols in a file or workspace.

AST Analysis​

Agents can parse and analyze code structure:

  • Detect unused exports and imports.
  • Validate naming conventions programmatically.
  • Generate metrics like cyclomatic complexity.

Best Practices​

Selective Feature Enablement​

  • Start with the core plugin and enable advanced features gradually.
  • Monitor context usage since more tools means more tokens per request.

Background Agent Design​

  • Give background agents clear, bounded tasks.
  • Set appropriate step limits to prevent runaway execution.
  • Use permissions to restrict what background agents can modify.

GitHub​