What is the primary purpose of the OpenCode Workflow skill?

It allows Claude to act as a manager that delegates heavy lifting—like code generation, research, and debugging—to OpenCode (GitHub Copilot) agents, saving Claude's context and tokens for orchestration.

How do I choose between opencode-orchestrator and opencode-ask?

Use opencode-orchestrator for iterative, multi-step tasks that need an autonomous loop. Use opencode-ask for quick, one-shot queries or simple code generation.

Can I use this for security reviews?

Yes, by using the --agent-prompt 'security-reviewer' flag, you can delegate deep security audits of your codebase to specialized models.

What are the prerequisites for using this skill?

You need the OpenCode binary installed, an authenticated GitHub Copilot account, and the opencode-ask wrapper configured in your system's PATH.

Which AI models are supported by this workflow?

Through the GitHub Copilot provider, it supports Claude 3.5 Sonnet, Claude 3 Opus, various GPT models (including GPT-5 equivalents), and Gemini 2.5 Pro for large context tasks.

OpenCode Delegation Workflow

Name: OpenCode Delegation Workflow
Author: Moliboy5000

byMoliboy5000

0•

生産性とワークフロー

Orchestrates complex task delegation between Claude and GitHub Copilot to optimize token usage and execution speed.

The OpenCode Workflow skill establishes Claude as a high-level orchestrator that intelligently delegates resource-intensive tasks to OpenCode (GitHub Copilot). It provides a structured framework for routing rules, agent-to-role mapping, and multi-model execution, allowing Claude to manage planning and file manipulation while offloading heavy implementation, research, and security audits to specialized sub-agents. This workflow supports a variety of models including Claude Sonnet, Opus, GPT-5, and Gemini, ensuring the best tool is used for every specific coding challenge.

主な機能

01Parallel workload capability for running background research or code generation

02Multi-model routing support for Claude, GPT, and Gemini models via GitHub Copilot

03Role-based agent mapping for specialized tasks like security reviews, debugging, and architecture

04Autonomous multi-step task delegation via the opencode-orchestrator agent

05One-shot query execution using the opencode-ask command for rapid answers

060 GitHub stars

ユースケース

01Delegating token-intensive research or multi-file analysis to Gemini's large context window

02Executing long-running implementation tasks in the background while Claude maintains high-level project orchestration

03Running automated security audits and code quality reviews using specialized agent roles

主な機能

01Parallel workload capability for running background research or code generation

02Multi-model routing support for Claude, GPT, and Gemini models via GitHub Copilot

03Role-based agent mapping for specialized tasks like security reviews, debugging, and architecture

04Autonomous multi-step task delegation via the opencode-orchestrator agent

05One-shot query execution using the opencode-ask command for rapid answers

060 GitHub stars

ユースケース

01Delegating token-intensive research or multi-file analysis to Gemini's large context window

02Executing long-running implementation tasks in the background while Claude maintains high-level project orchestration

03Running automated security audits and code quality reviews using specialized agent roles