opencode-omo

What This Skill Does

The opencode-omo skill is an advanced orchestration layer designed to transform OpenCode into a powerhouse for autonomous software development. It functions as a structured operating guide and interface for the Oh-My-OpenCode ecosystem, utilizing a specialized trio of agents: Prometheus for strategic architectural planning, Atlas for precise execution, and Sisyphus for iterative refinement. By leveraging these agents, users can transition from simple coding tasks to complex, multi-file software projects with consistent quality control and repeatable workflows.

Installation

To integrate this skill into your OpenClaw environment, execute the following command in your terminal:

clawhub install openclaw/skills/skills/iisweetheartii/opencode-omo

Once installed, you should verify that your local OpenCode environment recognizes the plugins. You can check this by running cat ~/.config/opencode/opencode.json | grep "oh-my-opencode" or by executing the provided diagnostic tool ./scripts/check-omo.sh to ensure all dependencies are properly configured.

Use Cases

This skill is ideal for developers who need to bridge the gap between intent and implementation. It is particularly effective for:

• Rapid feature prototyping: Use ulw (ultrawork) to quickly generate boilerplate or unit tests for new modules.
• Complex architectural refactoring: Utilize the @plan (Prometheus) feature to map out dependencies before allowing the execution agent (Atlas) to modify the codebase.
• Maintaining consistent coding standards: By using the bundled helper scripts, teams ensure that every automated commit or pull request follows the same organizational rigor.

Example Prompts

1. "@plan implement a secure authentication middleware for our Express backend that handles JWT validation and error logging."
2. "/start-work using the existing plan to migrate the legacy React components to TypeScript."
3. "opencode run --agent sisyphus 'fix the race condition in the WebSocket handler and add unit tests to verify the fix.'"

Tips & Limitations

To maximize effectiveness, always prioritize the planning phase for non-trivial tasks. Do not attempt to force direct edits outside of the OpenCode ecosystem, as this can break the orchestration flow. If the agent stalls or encounters an ambiguity, use the iterative nature of Sisyphus to narrow your request scope. Always treat the agent's planning output as a draft that requires human oversight before triggering the /start-work command. Remember that while this tool automates significant portions of the workflow, it remains a collaborative developer agent that performs best when the human provides clear, scoped objectives.