Build your first Claude Code app

free beta with Pawgrammer

claude-code-riper-5

Structured 5-phase development workflow for Claude Code using custom slash commands and subagents. Separates research, planning, and execution phases to reduce context usage and prevent premature implementation.

development

RIPER Workflow for Claude Code

Research ā€¢ Innovate ā€¢ Plan ā€¢ Execute ā€¢ Review

A structured, context-efficient development workflow for Claude Code using custom slash commands and subagents.

šŸŽÆ 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

šŸš€ Quick Start

Installation

  1. Copy the .claude directory to your project root:
BASH
cp -r .claude /path/to/your/project/

  1. Update .claude/project-info.md with your project details

  2. Customize .claude/riper-config.json with your tech stack and paths

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

    BASH
    echo ".claude/memory-bank/" >> .gitignore

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

    BASH
    echo ".claude/memory-bank/*/plans/" >> .gitignore
    BASH
    echo ".claude/memory-bank/*/reviews/" >> .gitignore

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

    BASH
    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

šŸ“š Commands Reference

RIPER Workflow Commands

CommandDescriptionMode Restrictions
/riper:strictEnable strict protocol enforcementNone
/riper:researchEnter research mode (read-only)Read only
/riper:innovateBrainstorm approaches (optional)Read only
/riper:planCreate technical specificationsRead + Write to memory bank
/riper:executeImplement approved planFull access
/riper:execute <step>Execute specific plan stepFull access
/riper:reviewValidate against planRead + Test execution
/riper:workflowFull guided workflowVaries by phase

Memory Bank Commands

CommandDescription
/memory:save <context>Save important context
/memory:recall <topic>Retrieve saved context
/memory:listList all memories

šŸ—ļø Architecture

Agents

RIPER uses 3 consolidated agents for efficiency:

  1. research-innovate - Handles research and brainstorming phases
  2. plan-execute - Manages planning and implementation
  3. review - Validates implementation against specifications

Memory Bank Structure

.claude/memory-bank/
ā”œā”€ā”€ main/                  # Main branch memories
ā”‚   ā”œā”€ā”€ plans/            # Technical specifications
ā”‚   ā”œā”€ā”€ reviews/          # Code review reports
ā”‚   ā””ā”€ā”€ sessions/         # Session contexts
ā””ā”€ā”€ [feature-branch]/     # Feature branch memories
    ā””ā”€ā”€ ...

Mode Capabilities

ModeReadWriteExecutePlanValidate
RESEARCHāœ…āŒāŒāŒāŒ
INNOVATEāœ…āŒāŒāŒāŒ
PLANāœ…šŸ“„*āŒāœ…āŒ
EXECUTEāœ…āœ…āœ…āŒāŒ
REVIEWāœ…šŸ“„*āœ…**āŒāœ…

* Only to memory bank
** Only test execution

šŸ”§ Configuration

project-info.md

Update with your project details:

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

riper-config.json

Customize workflow settings:

  • Important paths
  • Tech stack patterns
  • Memory bank configuration
  • Agent models

settings.json

Claude Code project settings:

  • Project name
  • Description
  • Custom instructions

šŸ’” Tips & Best Practices

When to Use RIPER

āœ… Good for:

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

āŒ 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

šŸ¤ Contributing

Improvements welcome! Key areas:

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

šŸ“œ Credits

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

šŸ“„ License

MIT - Use freely in your projects

Example Session

Start strict mode (always):

BASH
/riper:strict

Research existing code:

BASH
/riper:research analyze current API structure

Create a plan:

BASH
/riper:plan add rate limiting to API endpoints

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

BASH
/riper:execute

Or execute specific steps

BASH
/riper:execute 1.2
BASH
/riper:execute 2

Validate implementation:

BASH
/riper:review

Troubleshooting

Q: Commands not working?
A: Ensure .claude directory is in project root

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

Q: Agent errors?
A: Verify riper-config.json is valid JSON

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

Recognition

Mentioned in Awesome Claude Code