OpenCode Orchestrator

🚀 Multi-Agent Orchestration Plugin for OpenCode

⚡ Quick Start

npm install -g opencode-orchestrator

Then in OpenCode:

/task "Implement a Diablo2 Online Game for Web"

⭐ Core Philosophy

┌───────────────────────────────────────────────────┐
│  🔍 EXPLORE → 📝 LEARN → 🔄 ADAPT → ⚡ ACT         │
│  Scan        Document    Adjust      Execute      │
└───────────────────────────────────────────────────┘

The Agents

Agent	Role	What It Does
🎯 Commander	Orchestrator	Leads the task, delegates work to other agents, ensures completion
📋 Planner	Researcher	Analyzes the project, creates the plan, documents findings
🔨 Worker	Implementer	Writes code, modifies files, follows the project's patterns
✅ Reviewer	Verifier	Tests changes, checks for errors, confirms quality

📖 Core Principles

Principle	What It Means
🔍 Never Assume	Always check the actual code and config files first
📝 Document Everything	Record patterns and findings for future tasks
📚 Verify with Evidence	Run tests, cite sources, complete with proof
🔄 Adapt to the Project	Match the project's existing style and patterns
🤝 Specialized Roles	Each agent has a clear, focused responsibility

🏛️ Workflow Architecture

        /task "Build REST API"
                │
    ╔═══════════════════════════════════╗
    ║  🎯 COMMANDER — "Start the task"  ║
    ╚═══════════════╤═══════════════════╝
                    │
    ┌───────────────┼───────────────┐
    ▼               ▼               ▼
┌───────┐      ┌───────┐      ┌───────┐
│PLANNER│      │WORKER │      │WORKER │   ← 🔥 50 PARALLEL
│plan.md│      │auth.ts│      │api.ts │      TASKS CONCURRENTLY
└───────┘      └───────┘      └───────┘
    │               │               │
    └───────────────┼───────────────┘
                    ▼
    ╔═══════════════════════════════════╗
    ║  ✅ REVIEWER — "Verify everything"║
    ╚═══════════════╤═══════════════════╝
                    │
           ┌────────┴────────┐
           │ TODO 100%?      │
           │ Issues = 0?     │
           └────────┬────────┘
             No ↙       ↘ Yes
          ♻️ LOOP      ✅ COMPLETE

Features

Feature	What It Does
🚀 60-80% Optimization	Smart polling & output caching for massive speed gains
⚡ 50 Parallel Tasks	Run up to 50 agents simultaneously
🔄 Non-Stop Recovery	WAL-based persistence (resumes tasks after crashes)
🔥 Multi-File Ops	Work on different files at the same time
🛡️ Self-Scaling	Dynamic concurrency limits based on success/failure
🩹 Memory Integrity	Strict resource cleanup prevents leaks in long sessions
🧬 Adaptive AI	Agents learn and adapt based on the project

🏔️ Extreme Resilience & Performance

Built for "Infinite Missions," the OpenCode Orchestrator is engineered to handle massive codebases and long-running tasks where absolute stability is non-negotiable.

🔄 Continuous Operation: Long-running missions survive process restarts. If the orchestrator or local machine crashes, work resumes exactly where it left off via WAL (Write-Ahead Log) persistence.
⚡ Maximized Resource Efficiency: Integrated output caching and smart polling reduce API overhead by 60-80%, allowing for more concurrent agents with less latency.
🛡️ Quality-Based Stability: Success/Failure tracking automatically scales currency limits. The system adapts to model reliability and rate limits in real-time, ensuring a stable mission even under chaos.
💎 Memory Integrity: Rigorous resource management guarantees zero memory leaks, even during sessions that last for days.

Error Handling

Error	What Happens
💥 Tool crash	Inject recovery prompt and retry
🚦 Rate limit hit	Wait and retry with exponential backoff
📦 Context overflow	Compact the context automatically
⏱️ Session timeout	Resume from the last checkpoint
🔨 Build failure	Fix the issue and retry

📚 Documentation

System Architecture — Full technical deep-dive

🎹 Developer's Words

Read the full note →

📄 License

MIT License. LICENSE