oc-agent

Create and refine AI agents for Claude Code and OpenCode platforms.

Updated Jun 20, 2025
One-click install
npx skills add https://github.com/rfhold/dot --skill oc-agent
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: oc-agent
Source: https://github.com/rfhold/dot/tree/main/.config/opencode/skills/oc-agent
Command: npx skills add https://github.com/rfhold/dot --skill oc-agent

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill streamlines the creation and improvement of AI agents, ensuring they adhere to best practices for both Claude Code and OpenCode platforms, making agent design more efficient and effective.

Core Features & Use Cases

  • Agent Creation: Design new agents from scratch with proper YAML formatting.
  • Agent Refinement: Improve existing agents by applying prompt engineering principles.
  • Cross-Platform Compatibility: Supports agent formats for both Claude Code and OpenCode.
  • Use Case: You need to create a new agent for OpenCode that can analyze code quality. This Skill will guide you through defining its description, permissions, and instructions according to OpenCode's specifications.

Quick Start

Use the oc-agent skill to create a new agent for Claude Code named 'code-refactor'.

Frequently Asked Questions about oc-agent

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create an AI agent for Claude Code or OpenCode?

To create an AI agent for Claude Code or OpenCode, you define the agent's name, description, tool usage, model selection, and permission modes according to the specific platform's YAML formatting specifications.

What prompt engineering best practices apply to agent architecture design?

Agent architecture design applies prompt engineering best practices by structuring agent definitions with clear descriptions, appropriate tool usage, selected models, and configured permission modes to ensure efficient and effective agent behavior.

Can I refine existing AI agents to improve their performance on OpenCode?

Yes, you can refine existing AI agents on OpenCode by applying prompt engineering principles to improve their instructions, tool selection, and overall architecture without redesigning them from scratch.

Does this agent design approach support cross-platform compatibility between Claude Code and OpenCode?

Yes, this agent design approach supports cross-platform compatibility by guiding users through the specific formats, permission modes, and tool usage specifications required by both Claude Code and OpenCode platforms.

What's the best way to configure permissions and tool usage for a new subagent?

The best way to configure permissions and tool usage for a new subagent is to follow platform-specific formatting guidelines, defining exact tool access and permission modes during the agent architecture design process.

Why do my custom agents fail to follow instructions properly in Claude Code?

Custom agents may fail to follow instructions in Claude Code if their architecture lacks proper prompt engineering structure, requiring refinement of descriptions, model selection, or permission modes to align with platform specifications.