RIPER Workflow for OpenCode

Research • Innovate • Plan • Execute • Review

A structured, context-efficient development workflow for OpenCode using custom commands/ and agents.

🎯 Overview

RIPER creates a controlled, guided flow that enforces separation between research, planning, and execution phases. This helps:

• Reduce context usage through specialized agents
• Prevent premature implementation before understanding
• Maintain clear documentation of decisions
• Enable reproducible development processes

🚀 Quick Start

Installation

1. Copy the .opencode directory to your project root:

   cp -r .opencode /path/to/your/project/

2. Update .opencode/project-info.md with your project details

3. Optional: Add .opencode/memory-bank/ to .gitignore to keep plans and reviews private:

   echo ".opencode/memory-bank/" >> .gitignore

   Or selectively ignore just plans and reviews while keeping memory bank structure:

   echo ".opencode/memory-bank/*/plans/" >> .gitignore

   echo ".opencode/memory-bank/*/reviews/" >> .gitignore

   Note: If you accidentally commit these, remove them before merging:

   git rebase -i HEAD~n  # where n is number of commits to review

Basic Usage

Start a guided RIPER session:

/riper-strict

Then use the workflow commands/:

1. Research - Investigate the codebase

   /riper-research analyze authentication system
   

2. Plan - Create technical specifications

   /riper-plan add OAuth2 support
   

3. Execute - Implement the approved plan

   /riper-execute
   

   Or execute a specific substep:

   /riper-execute 2.3
   

4. Review - Validate implementation

   /riper-review
   

📚 commands/ Reference

RIPER Workflow commands/

Command	Description	Mode Restrictions
/riper-strict	Enable strict protocol enforcement	None
/riper-research	Enter research mode (read-only)	Read only
/riper-innovate	Brainstorm approaches (optional)	Read only
/riper-plan	Create technical specifications	Read + Write to memory bank
/riper-execute	Implement approved plan	Full access
/riper-execute <step>	Execute specific plan step	Full access
/riper-review	Validate against plan	Read + Test execution
/riper-workflow	Full guided workflow	Varies by phase

Memory Bank commands/

Command	Description
/memory-save <context>	Save important context
/memory-recall <topic>	Retrieve saved context
/memory-list	List all memories

🏗️ Architecture

Agents

RIPER uses 3 consolidated agents with optimized model assignments:

Agent	Model	Temperature	Purpose
research-innovate	opencode/gpt-5-nano	0.7	Fast analysis and creative brainstorming
plan-execute	opencode/glm-5-free	0.2	Balanced planning and implementation
review	opencode/minimax-m2.5-free	0.1	Quality validation and review

Agent Permissions

• research-innovate: Read-only, git commands/ allowed (git log, diff, status, branch)
• plan-execute: Full access with bash operations requiring approval
• review: Read-only, test commands/ allowed (pnpm test, lint, type-check), git diff/log

Memory Bank Structure

.opencode/memory-bank/
├── main/                  # Main branch memories
│   ├── plans/            # Technical specifications
│   ├── reviews/          # Code review reports
│   └── sessions/         # Session contexts
└── [feature-branch]/     # Feature branch memories
    └── ...

Mode Capabilities

Mode	Read	Write	Execute	Plan	Validate
RESEARCH	✅	❌	❌	❌	❌
INNOVATE	✅	❌	❌	❌	❌
PLAN	✅	📄*	❌	✅	❌
EXECUTE	✅	✅	✅	❌	❌
REVIEW	✅	📄*	✅**	❌	✅

* Only to memory bank ** Only test execution

🔧 Configuration

project-info.md

Update with your project details:

• Project name and description
• Common commands/
• Directory structure
• Tech stack
• Development guidelines

opencode.jsonc

OpenCode project configuration:

• Provider settings
• MCP tools
• Custom tools configuration

💡 Tips & Best Practices

When to Use RIPER

✅ Good for:

• Complex multi-step features
• Refactoring critical code
• Implementing architectural changes
• Tasks requiring careful planning

❌ Skip for:

• Simple bug fixes
• Minor text changes
• Straightforward updates
• Exploratory coding

Workflow Tips

1. Always start with research - Understanding before coding prevents wasted effort
2. Use memory bank - Save important context for future sessions
3. Be specific in plans - Detailed specs lead to accurate implementation
4. Review ruthlessly - Catch deviations early
5. Iterate on substeps - Use /riper-execute 2.3 to fix specific parts

Context Efficiency

• Research phase uses cheaper, focused tools
• Plans are saved to memory, reducing repetition
• Execute phase has full context from approved plan
• Review validates without reimplementing

🤝 Contributing

Improvements welcome! Key areas:

• Additional workflow commands/
• Better memory management
• Enhanced agent capabilities
• Integration examples

📜 Credits

This implementation is based on the RIPER-5 workflow created by an anonymous contributor (robotlovehuman) on the Cursor Forums.

Adapted for OpenCode from tony/claude-code-riper-5.

📄 License

MIT - Use freely in your projects

---

Example Session

Start strict mode (always):

/riper-strict

Research existing code:

/riper-research analyze current API structure

Create a plan:

/riper-plan add rate limiting to API endpoints

Review the generated plan in .opencode/memory-bank/main/plans/:

/riper-execute

Or execute specific steps

/riper-execute 1.2

/riper-execute 2

Validate implementation:

/riper-review

Troubleshooting

Q: commands/ not working? A: Ensure .opencode directory is in project root

Q: Memory bank not saving? A: Check write permissions on .opencode/memory-bank/

Q: Agent errors? A: Verify opencode.jsonc is valid JSON

Q: How to customize for my project? A: Start with project-info.md

Recognition

Adapted from the excellent claude-code-riper-5 project for Claude Code.