Coding & Developmentπ€claudeβ¨οΈcursorπuniversal
ICMwRIPER-5 (Iterative Context Management with RIPER-5) Method
An AI-assisted development method combining Iterative Context Management with [RIPER-5](https://forum.cursor.com/t/i-created-an-amazing-mode-called-riper-5-mode-fixes-claude-3-7-drastically/65516) rules (Research, Innovate, Plan, Execute, Review) to produce better-quality code that meets requirements.
Quick Start
Install this Skill
npx skills add wubin28/ICMwRIPER-5# ICMwRIPER-5 (Iterative Context Management with RIPER-5) Method
A cross-platform AI-assisted software development methodology that iteratively applies 1 human-driven Context Management step and 5 [RIPER-5](https://forum.cursor.com/t/i-created-an-amazing-mode-called-riper-5-mode-fixes-claude-3-7-drastically/65516) rules (Research, Innovate, Plan, Execute, Review) to guide AI code generation, producing higher-quality code that better meets your requirements.
**Supported Platforms**: Ubuntu (WSL2), macOS, and Windows PowerShell
## Cheatsheet for WSL2-Ubuntu/macOS/Windows-PowerShell
```bash
# create-html-data-dashboard
icm4u/icm4m/icm4p.ps1 create-html-data-dashboard <project-name>
# create-nextjs-web-app
icm4u/icm4m/icm4p.ps1 create-nextjs-web-app <project-name>
# snb
icm4u/icm4m/icm4p.ps1 snb <source-story-file>
# bubble-only
icm4u/icm4m/icm4p.ps1 bo
# dialog-log
icm4u/icm4m/icm4p.ps1 d
```
## The Problem
Directly presenting final requirements to AI often leads to code generation that deviates from intended outcomes because the AI lacks human intervention during the development process. Without structured guidance, AI assistants may:
- Make assumptions that don't align with your actual needs
- Skip important planning steps and jump directly to implementation
- Generate code that technically works but doesn't solve the right problem
- Introduce unnecessary complexity or miss edge cases
- Drift away from the original vision as the codebase evolves
## The Solution: ICMwRIPER-5 Method
ICMwRIPER-5 provides a structured, iterative approach that keeps both human and AI aligned throughout the development lifecycle. By separating concerns into distinct phases and maintaining clear context management, this method ensures that AI-generated code stays on track and meets actual requirements.
## Platform Support
ICMwRIPER-5 provides native command-line tools for Ubuntu, macOS, and Windows PowerShell environments:
### Ubuntu (WSL2)
- **Environment**: Windows 11 with WSL2 (Ubuntu 24.04 or later)
- **Command**: `icm4u`
- **Installation**: Global installation to `/usr/local/bin/`
- **Prerequisites**: Git, curl
### macOS
- **Environment**: macOS Sequoia 15.6.1 or later
- **Command**: `./icm4m` (local script)
- **Installation**: Local script usage
- **Prerequisites**: iTerm2, zsh, Git, curl
### Windows PowerShell
- **Environment**: Windows 11 with PowerShell 7.5.3 or later, or Windows 11 default PowerShell 5.1
- **Command**: `icm4p.ps1` (after global install) or `pwsh -File icm4p.ps1` (local script)
- **Installation**: Optional global installation to user PATH or local script usage
- **Prerequisites**: PowerShell 7.5.3+ or PowerShell 5.1 (default on Windows 11), Git
- **Note for PowerShell 5.1 users**: Before running `icm4p.ps1`, execute `Set-ExecutionPolicy RemoteSigned -Scope CurrentUser` to enable script execution. Otherwise you may see an error: "icm4p.ps1 cannot be loaded because running scripts is disabled on this system."
All three implementations provide identical functionality with platform-specific optimizations for the best native experience.
## How It Works
```mermaid
%%{init: {'theme':'base', 'themeVariables': { 'primaryColor':'#ffffff'}}}%%
flowchart TD
Start([Start New Iteration]) --> ICM["<div style='text-align: left'>Phase 1: Iterative Context Management<br/>by human<br/><br/>Review & update:<br/>β’ icm-story-yyyy-mm-dd--hh-mm.md<br/>β’ icm-bubble-yyyy-mm-dd--hh-mm.md<br/>β’ Reference latest context file<br/>β’ Reference icmwriper-5.md</div>"]
ICM --> SendPrompt["<div style='text-align: left'>Send prompts from bubble file to AI</div>"]
SendPrompt --> Research["<div style='text-align: left'>Phase 2: RESEARCH<br/>by AI & human<br/><br/>MODE: RESEARCH<br/><br/>β’ Information gathering & file reading only<br/>β’ AI clarifies requirements with human<br/>β’ No code changes or implementation decisions</div>"]
Research --> Innovate["<div style='text-align: left'>Phase 3: INNOVATE<br/>by AI & human<br/><br/>MODE: INNOVATE<br/><br/>β’ Brainstorm approaches without implementation<br/>β’ Propose multiple solution strategies<br/>β’ Discuss trade-offs</div>"]
Innovate --> Plan["<div style='text-align: left'>Phase 4: PLAN<br/>by AI & human<br/><br/>MODE: PLAN<br/><br/>β’ Create technical specifications<br/>β’ Save to: todo-yyyy-mm-dd--hh-mm.md<br/>β’ Plan must be detailed enough to execute mechanically</div>"]
Plan --> Execute["<div style='text-align: left'>Phase 5: EXECUTE<br/>by AI & human<br/><br/>MODE: EXECUTE<br/><br/>β’ Follow the plan exactly<br/>β’ No creative decisions<br/>β’ Implement only what was specified</div>"]
Execute --> Review["<div style='text-align: left'>Phase 6: REVIEW<br/>by AI & human<br/><br/>MODE: REVIEW<br/><br/>β’ Validate implementation against plan<br/>β’ Check: completeness, correctness, alignment</div>"]
Review --> GenContext["<div style='text-align: left'>Phase 7: GENERATE NEW CONTEXT FILE<br/>by human<br/><br/>Ask AI to generate:<br/>β’ icm-context-y
... (truncated)Repository
6stars
2forks
Details
Related Skills
More coding & development skills you might like
Reticle
Coding & DevelopmentAI agents can generate code, but still struggle to understand what they build. Reticle brings Jev-style machine-native runtime perception to web & desktop applications.
π€claudeπuniversal
1.2k
190
View
agentscope-java
Coding & DevelopmentExpert Java developer skill for AgentScope Java framework - a reactive, message-driven multi-agent system built on Project Reactor. Use when working with reactive programming, LLM integration, agent orchestration, multi-agent systems, or when the user mentions AgentScope, ReActAgent, Mono/Flux, Project Reactor, or Java agent development. Specializes in non-blocking code, tool integration, hooks, pipelines, and production-ready agent applications.
πuniversal
6.0k
1.5k
View
ospec
Coding & DevelopmentDocument-driven OSpec workflow for initialization, change/goal routing, validation, archiving, and durable project knowledge.
β¨geminiπ€claudeπ¬gptπuniversal
524
34
View
misakanet-failure-memory
Coding & Developmentπ A zero-dependency, git-backed micro-lesson library for AI Agents to asynchronously share and search verified debugging experience. Python stdlib only. | https://misakanet.org
π€claudeβ¨οΈcursorπuniversal
424
158
View