Skip to main content

🎯 MCP ADR Analysis Server Prompting Guide

Master the art of effective prompting for architectural analysis and decision-making

Version: 2.1.0 | Updated: September 2025 | Audience: AI Assistants, Developers, Architects

📖 Table of Contents​

  1. Quick Start Prompting
  2. Basic Prompting Patterns
  3. Advanced Prompting Techniques
  4. Tool-Specific Prompting
  5. Multi-Tool Workflows
  6. Troubleshooting Prompts
  7. Best Practices & Templates

🚀 Quick Start Prompting​

Essential Prompt Structure​

[CONTEXT] + [TOOL_REQUEST] + [PARAMETERS] + [EXPECTED_OUTPUT]

Example: Basic Project Analysis​

Context: "I'm analyzing a Node.js microservices project"
Tool: "Use analyze_project_ecosystem"
Parameters: "with comprehensive analysis and AI enhancement enabled"
Output: "Provide architectural insights and ADR suggestions"

Complete Prompt:

"I'm analyzing a Node.js microservices project. Use analyze_project_ecosystem with comprehensive analysis and AI enhancement enabled to provide architectural insights and ADR suggestions."

🎨 Basic Prompting Patterns​

Pattern 1: Single Tool Analysis​

**Template:**
"Analyze [PROJECT_TYPE] using [TOOL_NAME] with [SPECIFIC_PARAMETERS] to [DESIRED_OUTCOME]"

**Examples:**

- "Analyze this React application using analyze_project_ecosystem with enhanced mode enabled to identify missing ADRs"
- "Review existing ADRs using review_existing_adrs with tree-sitter analysis to assess compliance"
- "Check deployment readiness using deployment_readiness with strict validation for production environment"

Pattern 2: Context-Rich Prompting​

**Template:**
"Given [CONTEXT], I need to [GOAL]. Please use [TOOL] with these considerations: [CONSTRAINTS]"

**Example:**
"Given that we're migrating from monolith to microservices, I need to document architectural decisions. Please use suggest_adrs with focus on distributed systems patterns and consider our existing Spring Boot stack."

Pattern 3: Progressive Analysis​

**Template:**
"Start with [INITIAL_TOOL], then based on results, proceed with [FOLLOW_UP_TOOLS]"

**Example:**
"Start with analyze_project_ecosystem to understand the codebase, then use suggest_adrs to identify missing decisions, and finally generate_adr_todo to create implementation tasks."

🧠 Advanced Prompting Techniques​

Chain-of-Thought Prompting​

Enable the server's advanced AI capabilities:

**Enhanced Analysis Request:**
"Please analyze this project using methodological pragmatism:

1. First, use analyze_project_ecosystem with knowledge enhancement
2. Apply reflexion framework to validate findings
3. Generate confidence-scored recommendations
4. Provide explicit acknowledgment of limitations"

Memory-Centric Prompting​

Leverage the memory system:

**Memory Integration:**
"Load existing ADRs into memory using memory_loading, then analyze architectural gaps using the knowledge graph to suggest improvements with high confidence scoring."

Multi-Dimensional Analysis​

**Comprehensive Assessment:**
"Perform a security-focused architectural analysis:

- Use analyze_content_security for vulnerability detection
- Apply tree-sitter analysis for deep code inspection
- Generate content masking recommendations
- Validate against enterprise security standards"

🛠️ Tool-Specific Prompting​

analyze_project_ecosystem​

Optimal Prompting:

"Analyze [PROJECT_PATH] using analyze_project_ecosystem with:

- Enhanced mode: true
- Knowledge enhancement: true
- Technology focus: [TECH_STACK]
- Analysis depth: comprehensive
- Include environment analysis: true

Focus on [SPECIFIC_AREAS] and provide confidence-scored insights."

Parameters to Always Include:

  • enhancedMode: true - Enables APE + Knowledge Generation
  • knowledgeEnhancement: true - Activates advanced AI features
  • analysisDepth: "comprehensive" - Full analysis scope

suggest_adrs​

Effective Prompting:

"Suggest ADRs for [PROJECT_CONTEXT] using suggest_adrs with:

- Before/after code analysis enabled
- Tree-sitter analysis for accurate detection
- Focus on [ARCHITECTURAL_CONCERNS]
- Include implementation guidance"

content_masking_tool​

Security-Focused Prompting:

"Analyze security vulnerabilities using analyze_content_security with:

- Tree-sitter analysis enabled
- Custom pattern detection for [SPECIFIC_SECRETS]
- Generate masking strategies for [ENVIRONMENT]
- Validate masking effectiveness"

deployment_readiness​

Production-Ready Prompting:

"Check deployment readiness using deployment_readiness with:

- Strict validation mode
- Environment: production
- Include mock vs production analysis
- Generate deployment guidance with confidence scoring"

🔄 Multi-Tool Workflows​

Workflow 1: New Project Setup​

1. "Analyze project structure using analyze_project_ecosystem"
2. "Based on findings, suggest missing ADRs using suggest_adrs"
3. "Generate implementation tasks using generate_adr_todo"
4. "Validate deployment readiness using deployment_readiness"

Workflow 2: Security Assessment​

1. "Scan for security issues using analyze_content_security"
2. "Generate masking strategies using generate_content_masking"
3. "Validate masking effectiveness using validate_content_masking"
4. "Create security-focused ADRs using suggest_adrs with security focus"

Workflow 3: Legacy Modernization​

1. "Analyze existing architecture using analyze_project_ecosystem"
2. "Review current ADRs using review_existing_adrs"
3. "Identify modernization opportunities using suggest_adrs"
4. "Generate migration rules using generate_rules"
5. "Validate compliance using validate_rules"

🔧 Troubleshooting Prompts​

When Tools Return Errors​

**Generic Error Recovery:**
"The previous tool failed. Please use troubleshoot_guided_workflow to:

1. Diagnose the issue systematically
2. Provide step-by-step resolution
3. Suggest alternative approaches
4. Validate the fix"

When Results Are Incomplete​

**Enhanced Analysis Request:**
"The analysis seems incomplete. Please retry with:

- Enhanced mode enabled
- Increased analysis depth
- Tree-sitter analysis activated
- Knowledge enhancement enabled
- Explicit confidence scoring"

When Prompts Are Too Vague​

**Clarification Request:**
"Please use get_workflow_guidance to:

- Clarify the optimal approach for [GOAL]
- Recommend specific tools and parameters
- Provide step-by-step workflow
- Estimate effort and timeline"

📋 Best Practices & Templates​

Template Library​

1. Comprehensive Project Analysis​

"I need a complete architectural analysis of [PROJECT_TYPE]. Please:

1. Use analyze_project_ecosystem with enhanced mode and knowledge enhancement
2. Focus on [TECHNOLOGY_STACK] with comprehensive depth
3. Include environment analysis and security assessment
4. Provide confidence-scored recommendations
5. Suggest missing ADRs based on findings
6. Generate actionable TODO items

Expected outcome: Complete architectural assessment with implementation roadmap."

2. Security-First Analysis​

"Perform a security-focused analysis of [PROJECT]. Please:

1. Use analyze_content_security with tree-sitter analysis
2. Detect secrets, vulnerabilities, and compliance issues
3. Generate content masking strategies
4. Validate masking effectiveness
5. Create security-focused ADRs
6. Provide deployment security guidance

Expected outcome: Comprehensive security assessment with remediation plan."

3. ADR Generation & Management​

"Help me establish ADR management for [PROJECT]. Please:

1. Discover existing ADRs using discover_existing_adrs
2. Review current ADRs using review_existing_adrs with compliance scoring
3. Suggest missing ADRs using suggest_adrs with architectural focus
4. Generate implementation tasks using generate_adr_todo
5. Create ADR validation rules using generate_rules

Expected outcome: Complete ADR management system with validation."

Performance Optimization Tips​

  1. Always Enable Enhanced Mode

    enhancedMode: true
    knowledgeEnhancement: true

  2. Use Specific Technology Focus

    technologyFocus: ["react", "nodejs", "docker", "kubernetes"]

  3. Request Confidence Scoring

    "Provide confidence scores for all recommendations"

  4. Enable Tree-Sitter Analysis

    enableTreeSitterAnalysis: true

Common Pitfalls to Avoid​

❌ Don't:

  • Use vague prompts without context
  • Skip enhanced mode for complex analysis
  • Ignore confidence scores in responses
  • Mix incompatible tool combinations

✅ Do:

  • Provide clear context and goals
  • Enable advanced AI features
  • Request explicit confidence levels
  • Follow recommended tool workflows

🎯 Prompt Templates by Use Case​

Enterprise Architecture Review​

"As an enterprise architect, I need to assess [SYSTEM] for compliance and modernization. Please use analyze_project_ecosystem with enterprise focus, then suggest_adrs for governance patterns, and generate_rules for compliance validation. Provide confidence-scored recommendations suitable for executive reporting."

DevOps Pipeline Integration​

"I'm integrating ADR analysis into our CI/CD pipeline. Please use deployment_readiness with strict validation, analyze_content_security for automated security checks, and smart_git_push for safe deployments. Focus on automation-friendly outputs and error handling."

Team Collaboration Setup​

"Our team needs collaborative ADR management. Please use interactive_adr_planning for team workflows, generate_adr_todo for task distribution, and tool_chain_orchestrator for automated workflows. Emphasize team coordination and knowledge sharing."

📚 Additional Resources​

💡 Pro Tip: The MCP ADR Analysis Server uses methodological pragmatism - always request confidence scores and explicit acknowledgment of limitations for the most reliable results.

🔄 Keep Updated: This guide evolves with the server. Check back regularly for new prompting patterns and techniques.