prompt
FreeTeams-First Multi-Agent Orchestration for Claude Code
FreeFree tier
Inputs: text, codeOutputs: text, code
About prompt
A specialized system prompt for Claude Code that implements a 'Teams-First Multi-Agent Orchestrator' (OMC), enabling coordinated multi-agent workflows. It defines operating principles, delegation rules, model routing (haiku/sonnet/opus), and a catalog of 18 specialized agents (e.g., explorer, analyst, planner, architect, debugger, executor) to handle complex software development tasks such as multi-file changes, refactors, debugging, planning, research, and verification. Integrates external AI tools (Codex, Gemini), OMC state management, team messaging, notepad, project memory, and code intelligence (LSP, AST).
Key Features
Coordinates 18 specialized agents (explorer, analyst, planner, architect, debugger, executor, etc.)
Operating principles: delegate specialized work, prefer evidence, choose lightest-weight path
Delegation rules for complex vs. trivial tasks
Model routing: haiku for quick lookups, sonnet for standard work, opus for deep analysis
External AI integration: /team N:executor task, omc team N:codex|gemini, omc ask claude|codex|gemini
OMC state management: read, write, clear, list active, get status
Team management: create, delete, send messages, task lifecycle
Notepad system for priority, working, and manual notes
Project memory with notes and directives
Code intelligence via LSP (hover, goto definition, find references, diagnostics) and AST (grep, replace)
Pros & Cons
Pros
- Automates complex multi-step development tasks through coordinated agent swarm
- Clear delegation and model routing optimized for different task complexities
- Integrates external AI models (Codex, Gemini) for expanded capabilities
- Provides state management and team collaboration features
- Includes code intelligence tools (LSP, AST) for precise code manipulation
- Open-source and free to use with Claude Code
Cons
- Only works within the Claude Code environment
- Requires understanding of agent hierarchy and delegation rules for effective use
- May introduce overhead for simple tasks that could be done directly
- Dependent on external AI API availability for some features
- Limited to text/code inputs and outputs
Best For
Multi-file code changes and refactorsDebugging complex issues across codebaseCode review with specialized reviewer agentsArchitecture planning and deep analysisSDK integration with official documentation consultingAutomated software development workflows (CI/CD-like)Research and verification tasksProject memory and notepad-based knowledge management
FAQ
What is oh-my-claudecode?
oh-my-claudecode is a multi-agent orchestration layer for Claude Code that coordinates specialized agents, tools, and skills to complete work accurately and efficiently. It is open-source with over 35k stars on GitHub.
How does delegation work?
Delegation rules specify when to delegate: for multi-file changes, refactors, debugging, reviews, planning, research, verification. Work directly on trivial ops, small clarifications, single commands. Specialized agents like executor (for code) and document-specialist (for SDK docs) are used.
What agents are available?
The agent catalog includes 18 agents: explore (haiku), analyst (opus), planner (opus), architect (opus), debugger (sonnet), executor (sonnet), verifier (sonnet), tracer (sonnet), security-reviewer (sonnet), code-reviewer (opus), test-engineer (sonnet), designer (sonnet), writer (haiku), qa-tester (sonnet), scientist (sonnet), document-specialist (sonnet), git-master (sonnet), code-simplifier (opus), critic (opus).
What tools are available?
Tools include external AI integration via omc team commands, OMC state management, team creation and messaging, notepad read/write, project memory, and code intelligence (LSP hover, goto definition, find references, AST grep/replace, Python REPL).