AI Prompts for LerianStudio Ecosystem

A comprehensive collection of AI prompts designed for complete software development lifecycle management within the LerianStudio ecosystem.

🎯 Overview

This repository provides integrated systems that work together to support the entire software development lifecycle:

System	Purpose	Workflow Type	Quick Start
Memory System	Cross-session learning & pattern recognition	Sequential (5 phases)	claude 0-memory-system/m0-memory-orchestrator.mdc
Product Development	Comprehensive planning from idea to implementation	Dynamic (4 phases, 2 checkpoints)	claude 1-pre-dev-product/0-pre-dev-orchestrator.mdc
Feature Development	Adaptive feature addition with complexity-based flow	Adaptive (4 phases, complexity-based)	claude 2-pre-dev-feature/0-feature-orchestrator.mdc
Frontend Development	Complete frontend development with flexible design inputs	Flexible (5 phases)	claude 3-frontend/0-frontend-orchestrator.mdc
Code Review	Streamlined 5-phase analysis with validation gates	Systematic (5 phases, 5 gates)	claude 4-code-review/0-review-orchestrator.mdc
Documentation Generation	Comprehensive documentation generation and distribution	Comprehensive (5 phases)	claude 5-generate-docs/0-docs-orchestrator.mdc

📚 Documentation

• CLAUDE.md - Complete orchestrator architecture and execution patterns for AI assistants
• Memory Management README - Detailed memory system documentation
• Pre-Development README - Interactive planning workflow guide
• Frontend Development README - Complete frontend development workflow
• Code Review README - Systematic analysis documentation
• Documentation Generation README - Comprehensive documentation workflow guide

🆕 Dynamic Workflow Innovation

Confidence-Based Interactions

The Pre-Development Product workflow now uses AI confidence scoring to minimize user interaction:

High Confidence (80-100%):
  - AI proceeds autonomously
  - Notifies user of decisions
  - User can intervene if needed

Medium Confidence (50-79%):
  - AI presents 2-3 options
  - User selects preferred approach
  - AI refines based on selection

Low Confidence (<50%):
  - AI requests specific guidance
  - Presents gaps in understanding
  - User provides targeted input

Tool Integration (Enhanced)

• Memory MCP: Pattern recognition, knowledge persistence, cross-project learning
• Sequential Thinking MCP: Complex problem decomposition with iterative refinement
• Zen MCP Suite: Comprehensive analysis tools:
  • codereview: Deep code quality analysis with severity prioritization
  • analyze: Architecture and pattern analysis across directories
  • thinkdeep: Complex problem exploration and solution validation
  • debug: Root cause analysis with 1M token capacity
  • chat: Collaborative thinking and validation
• Task Tool: Parallel search operations for maximum efficiency in pattern discovery

🚀 Quick Start Workflows

1. New Feature Development

# Initialize memory context
claude 0-memory-system/m0-memory-orchestrator.mdc

# Plan the feature interactively
claude 1-pre-dev-product/0-pre-dev-orchestrator.mdc

# Analyze integration points with foundation analysis
claude 4-code-review/1-foundation-analysis.mdc

2. Frontend Development

# Design input analysis (any format: sketches, Figma, specs)
claude 3-frontend/1-design-input-analysis.mdc

# Complete frontend implementation
claude 3-frontend/0-frontend-orchestrator.mdc

# Validate with streamlined code review
claude 4-code-review/0-review-orchestrator.mdc

3. Existing Code Analysis

# Full streamlined code review
claude 4-code-review/0-review-orchestrator.mdc

# Store findings
claude 0-memory-system/m4-memory-workflow.md

4. Documentation Generation

# Comprehensive documentation suite
claude 5-generate-docs/0-docs-orchestrator.mdc

# Store documentation patterns
claude 0-memory-system/m4-memory-workflow.md

5. Quick Security Check

# Security-focused analysis (phases 1, 2, and 5 only)
claude 4-code-review/1-foundation-analysis.mdc
claude 4-code-review/2-security-compliance.mdc
claude 4-code-review/5-production-readiness.mdc

🔄 Development Lifecycle Integration

Complete Development Cycle

graph LR
    A[Memory Context] --> B[Pre-Development Planning]
    B --> C[Frontend Development]
    C --> D[Implementation]
    D --> E[Code Review]
    E --> F[Documentation Generation]
    F --> G[Memory Storage]
    G --> A

Loading

1. Initialize: Start with memory context to leverage existing patterns
2. Plan: Use pre-development for interactive requirements and design
3. Frontend: Design and implement frontend with systematic precision
4. Implement: Follow generated tasks and sub-tasks
5. Review: Comprehensive code analysis and validation
6. Document: Generate comprehensive documentation for all audiences
7. Learn: Store insights back to memory for future projects

🎯 Key Features

🔗 Orchestrator Architecture

• Dynamic Phase Execution: Pre-Development now uses confidence-based execution with only 2 mandatory checkpoints
• Memory MCP Integration: Context retrieval, decision storage, pattern analysis across sessions
• Multi-Tool Integration: Sequential Thinking, Zen MCP, and Task tool for enhanced capabilities
• Cross-System Dependencies: Workflows feed into each other with validation gates

📋 User Interaction Patterns

• Adaptive Complexity Assessment: Feature Development now includes Phase 0 complexity analysis
• Confidence-Based Interactions: AI decides when user input is needed based on confidence scores
• Structured Feedback Loops: Draft → User feedback → AI incorporation → Approval
• Minimal Interruption: From 6+ checkpoints down to just 2 in Pre-Development workflow
• Flexible Design Inputs: iPad sketches, Figma designs, written specs, reference apps
• Risk-Based Planning: Complex features get automatic risk assessment and mitigation strategies

🚀 Workflow Features

• Deliverable-Focused Planning: Shifted from atomic tasks to complete deliverables (2-pre-dev-feature)
• Autonomous Refinement: AI works independently when confidence is high
• Parallel Processing: Multiple operations execute simultaneously for efficiency
• Adaptive Task Sizing: 2-8 hours based on complexity assessment
• Smart Validation: Consistency checks with auto-correction capabilities
• Priority-Based Todos: 🔴 CRITICAL → 🟡 HIGH → 🟢 MEDIUM → 🔵 LOW organization
• Pattern Reuse: Automatic detection and application of similar solutions

📦 Output Organization

Pre-Development Outputs

docs/pre-development/
├── prd-[feature].md              # Product requirements
├── trd-[feature].md              # Technical specifications
├── validation-report-[feature].md # Consistency validation
└── tasks/
    ├── tasks-[feature].md        # Atomic phases
    └── MT-[XX]-[name]/           # Sub-task details

Code Review Outputs

docs/code-review/
├── review-summary.md             # Executive summary
├── 1-foundation-analysis.md      # Technical foundation
├── 2-security-compliance.md      # Security findings
├── 3-quality-operations.md       # Quality assessment
├── 4-business-documentation.md   # Business alignment
├── 5-production-readiness.md     # Final validation
└── code-review-todos.md          # Consolidated action items

Documentation Generation Outputs

docs/documentation/
├── documentation-audit.md        # Discovery phase audit
├── documentation-plan.md         # Strategic planning
├── validation-report.md          # Quality validation
├── distribution-strategy.md      # Multi-channel distribution
└── content/
    ├── business/                 # Product team docs
    ├── technical/                # Developer docs
    ├── integration/              # API consumer docs
    └── operations/               # DevOps docs

🏆 Orchestrator Best Practices

Execution Order Guidelines

1. Memory First: Always start with memory_search and memory_get_context before any workflow
2. Follow Phase Dependencies:
   • Pre-Development: Discovery (conditional) → Strategic Decision✓ → Autonomous Refinement → Final Validation✓
   • Feature Development: Phase 0 Complexity Assessment → Adaptive Brief (1-7 questions) → Technical Approach✓ → Deliverable Planning → Test Strategy
   • Code Review: Foundation → Gate 1✓ → Security → Gate 2✓ → Quality → Gate 3✓ → Business → Gate 4✓ → Production → Gate 5✓
   • Frontend: Design✓ → Tech✓ → Architecture → Tasks → Validation
   • Documentation: Discovery✓ → Planning✓ → Generation✓ → Validation✓ → Distribution
3. Leverage AI Autonomy: Let confidence scores guide interaction needs
4. Use Tools Liberally:
   • Zen MCP: Use specific tools (codereview, analyze, thinkdeep, debug, chat) based on context
   • Task Tool: Always prefer for parallel searches and pattern discovery
5. Store All Decisions: Use memory_store_decision for architectural choices and memory_store_chunk for insights

Integration Patterns

• New Product: Memory → Pre-Development Product → Frontend → Code Review → Documentation → Memory
• Feature Enhancement: Memory → Pre-Development Feature → Implementation → Code Review → Documentation → Memory
• Existing Analysis: Memory → Code Review → Memory
• Documentation Focus: Memory → Documentation Generation → Distribution → Memory
• Security Focus: Memory → Code Review (phases 1,2,5) → Memory

🏗️ Integration with LerianStudio

This repository supports the broader LerianStudio ecosystem:

• Midaz Financial Ledger: Architecture and security analysis
• Plugin Ecosystem: Component analysis and integration patterns
• SDK Development: API design and documentation workflows
• Infrastructure: Observability and deployment analysis

Repository Context

# Use appropriate repository context
repository="github.com/lerianstudio/midaz"
repository="github.com/lerianstudio/midaz-private"
repository="github.com/lerianstudio/monorepo"

📈 Key Benefits

• Consistency: Standardized workflows across all projects
• Quality: Comprehensive analysis at every stage with Zen MCP integration
• Learning: Continuous improvement through memory and pattern recognition
• Efficiency: Reuse patterns and decisions, deliverable-focused planning
• Collaboration: Clear user interaction points with adaptive complexity
• Tool Synergy: Zen MCP + Task Tool + Memory MCP work together for comprehensive analysis

🚀 Efficiency Improvements

• 70% Reduction in Interactions: Pre-Development workflow reduced from 6+ to 2 mandatory checkpoints
• Adaptive Complexity: Feature Development adjusts depth based on feature complexity (1-7 questions)
• Parallel Processing: Multiple analyses run simultaneously with Task Tool
• Pattern Reuse: Memory MCP enables >60% pattern reuse from similar projects
• Autonomous Refinement: AI works independently when confidence is high (>80%)
• Smart Validation: Auto-correction of minor issues without user intervention
• Tool-Powered Analysis: Zen MCP tools accelerate deep analysis by 5-10x

🔧 Getting Started

For AI Assistants

1. Read CLAUDE.md First: Complete orchestrator architecture and execution patterns
2. Choose Your Workflow:
   • New project? Start with Pre-Development Product (4-phase dynamic, only 2 checkpoints!)
   • Existing code? Begin with Code Review (5-phase streamlined with validation gates)
   • Frontend focus? Use Frontend Development (5-phase flexible)
   • Feature addition? Use Pre-Development Feature (4-phase adaptive with complexity assessment)
3. Leverage Tools Strategically:
   • Zen MCP: Choose the right tool for the task:
     • codereview for quality assessment
     • analyze for architecture review
     • thinkdeep for complex reasoning
     • debug for troubleshooting (1M token capacity)
     • chat for collaborative validation
   • Task Tool: Always use for parallel searches and pattern discovery
   • Memory MCP: Start and end every workflow with memory operations
   • Let confidence scores and complexity assessments guide your interactions

For Developers

1. Start with Orchestrators: Each system has a 0-*-orchestrator.mdc entry point
2. Follow Phase Dependencies: Respect mandatory checkpoints and user feedback loops
3. Leverage Memory: Check context before starting, store learnings after completion
4. Use Integration Patterns: Connect workflows for complete development lifecycle

Command Examples

# Complete new feature workflow with adaptive complexity
claude 0-memory-system/m0-memory-orchestrator.mdc
claude 2-pre-dev-feature/0-feature-orchestrator.mdc  # Now with Phase 0 complexity assessment
claude 3-frontend/0-frontend-orchestrator.mdc
claude 4-code-review/0-review-orchestrator.mdc  # Streamlined 5-phase with validation gates
claude 5-generate-docs/0-docs-orchestrator.mdc

# Quick feature addition with risk assessment
claude 2-pre-dev-feature/0-feature-orchestrator.mdc  # Adapts to feature complexity
# Phase 0: Automatic complexity assessment
# Phases 1-4: Adjusted based on complexity (Simple/Medium/Complex)

📄 License

This project is part of the LerianStudio ecosystem. See the main repository for licensing information.

---

Part of the LerianStudio ecosystem - Building the future of financial technology with AI-assisted development.