How does OmO decide which agents to use?

It uses routing signals based on task clarity and risk. For example, it adds the 'explore' agent if code location is unclear and the 'oracle' agent for high-risk changes involving concurrency or public APIs.

Does OmO write code directly?

No, OmO is strictly an orchestrator; it delegates all code modifications to specialized implementation agents like 'develop' or 'frontend-ui-ux-engineer' to ensure quality.

What is the primary role of the OmO skill?

OmO acts as an orchestrator (persona: Sisyphus) that analyzes your request and routes it to specialized agents for exploration, planning, and implementation.

Can I use OmO for documentation tasks?

Yes, OmO includes a dedicated 'document-writer' agent and can orchestrate the necessary exploration to ensure your documentation accurately reflects the codebase.

OmO Multi-Agent Orchestrator

Name: OmO Multi-Agent Orchestrator
Author: cexll

bycexll

•

1,960

•

生産性とワークフロー

Orchestrates specialized AI agents to handle complex code analysis, bug investigation, and implementation workflows using a routing-first approach.

OmO acts as a high-level manager named Sisyphus that delegates technical tasks to a suite of specialized agents including Explore, Librarian, Oracle, and Develop. Instead of following a rigid pipeline, it evaluates task risk and clarity to route requests through the most efficient path—whether that's a simple one-step fix or a multi-stage investigation and review process. By focusing on context propagation and avoiding direct code modification, OmO ensures that complex repository changes are planned, validated, and executed by the most appropriate specialized tools, minimizing errors and maximizing developer productivity.

主な機能

011,960 GitHub stars

02Specialized support for external documentation lookups via the Librarian agent

03Dynamic agent routing based on task complexity and risk signals

04Context-aware delegation that passes history and requirements between agents

05Risk-based architectural review using the Oracle agent for high-impact refactors

06Automated handoff to UI/UX or documentation specialists for specific domains

ユースケース

01Executing high-risk, multi-file refactors that require a planning and review stage before implementation

02Implementing new features that require cross-referencing internal patterns with external API documentation

03Investigating and fixing bugs where the root cause or file location is unknown

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add cexll/myclaude omo

For use in Claude.ai and ChatGPT