What is the difference between oh-my-opencode and oh-my-openagent?
oh-my-opencode has been renamed to oh-my-openagent. While the project is transitioning, both names are recognized. The compatibility layer now prefers 'oh-my-openagent' in opencode.json, but legacy 'oh-my-opencode' entries still load with a warning, and plugin config files commonly use the old name.
How does oh-my-openagent handle different AI models?
oh-my-openagent orchestrates various AI models like Claude, GPT, Kimi, and Gemini, assigning them to tasks based on their strengths. It uses a category-based delegation system, where agents specify the type of work, and the harness automatically routes it to the most suitable model, such as GPT-5.4 for 'ultrabrain' tasks.
What is the 'ultrawork' command and how does it function?
The 'ultrawork' (or 'ulw') command is a core feature that activates a full team of AI agents to complete a task from start to finish. It leverages all underlying features and optimizations, including Discipline Agents, IntentGate, and Ralph Loop, to ensure the task is done without requiring further user intervention.
How does oh-my-openagent ensure code changes are accurate and prevent stale-line errors?
It uses a Hash-Anchored Edit Tool, inspired by oh-my-pi. Every line read by an agent is tagged with a content hash. When an agent proposes an edit, it references these tags. If the file has changed, the hash won't match, and the edit is rejected, preventing corruption and stale-line errors.
Can I integrate my existing Claude Code setup with oh-my-openagent?
Yes, oh-my-openagent is designed for full compatibility with Claude Code. Your existing hooks, commands, skills, MCPs, and plugins from your Claude Code setup will work seamlessly within oh-my-openagent without requiring any changes.