LLM Prompts for Architecture Documentation
A collection of carefully crafted prompts to help with software architecture documentation using arc42, docToolchain, and related methodologies. These prompts are designed to work with various LLMs (Claude, ChatGPT, etc.) to streamline architecture documentation workflows.
.1. 🎯 What’s This About?
Creating good architecture documentation is challenging. These prompts help you:
-
Structure your architecture thinking systematically
-
Generate consistent, high-quality documentation
-
Follow established methodologies like arc42
-
Create visual diagrams and decision records
-
Maintain documentation as living artifacts
.2. 📋 Available Prompts
.2.1. Core Architecture Tools
Prompt | Description | Output Format |
---|---|---|
Creates a visual canvas summarizing key architecture aspects in one page |
AsciiDoc with HTML table |
|
Structured approach to documenting architecture decisions with Pugh Matrix evaluation |
AsciiDoc document |
.2.2. Comprehensive Documentation Tools
Prompt | Description | Output Format |
---|---|---|
Step-by-step creation of complete arc42 documentation with quality-driven approach |
Multiple AsciiDoc files |
|
Create testable, specific quality requirements using structured scenarios |
AsciiDoc quality scenarios |
.2.3. Risk and Debt Management
Prompt | Description | Output Format |
---|---|---|
Systematic identification and assessment of architecture risks with mitigation strategies |
AsciiDoc risk register |
|
Comprehensive technical debt identification, prioritization, and management |
AsciiDoc debt register |
.2.4. Specialized Prompts
Prompt | Description | Output Format |
---|---|---|
include::descriptions/Stakeholder-Analysis-Description.adoc[tag=purpose] |
AsciiDoc stakeholder register |
|
include::descriptions/Context-Diagram-Generator-Description.adoc[tag=purpose] |
AsciiDoc with C4-PlantUML diagrams |
|
Develop comprehensive solution strategies based on quality goals following arc42 Chapter 4 principles |
AsciiDoc strategy document |
|
Design and document comprehensive system deployment architectures following arc42 Chapter 7 principles |
AsciiDoc deployment documentation |
.3. 🚀 How to Use These Prompts
.3.1. 1. Copy & Paste Approach
Simply copy the prompt content and paste it into your preferred LLM interface.
.3.2. 2. Custom AI Assistant Setup
Use these prompts as system prompts to create specialized AI assistants for architecture work.
.3.3. 3. Integration with docToolchain
These prompts are designed to work seamlessly with docToolchain workflows:
-
Output is in AsciiDoc format
-
Diagrams use PlantUML/Kroki
-
Compatible with arc42 template structure
.4. 🎨 Customization
These prompts are designed to be customizable:
-
Modify Questions: Adapt the guiding questions to your domain
-
Adjust Output Format: Change AsciiDoc styling or structure
-
Add Custom Sections: Include organization-specific requirements
-
Integrate Tools: Connect with your existing toolchain
.5. 📚 Background & Methodology
.5.1. arc42 Template
These prompts follow the arc42 template structure, which provides:
-
Proven documentation structure
-
Clear separation of concerns
-
Stakeholder-oriented communication
-
Pragmatic approach to architecture documentation
.5.2. Architecture Communication Canvas
Based on Gernot Starke’s Architecture Communication Canvas, focusing on:
-
Value proposition clarity
-
Stakeholder identification
-
Core functionality emphasis
-
Risk and decision transparency
.5.3. Decision Making
Architecture Decision Records (ADRs) use:
-
Pugh Matrix for objective evaluation
-
Structured problem-solution mapping
-
Consequence documentation
-
Traceability of architectural reasoning
.5.4. Quality-Driven Architecture
Quality Scenarios provide:
-
Testable quality requirements
-
Objective success criteria
-
Architecture decision guidance
-
Validation frameworks
.5.5. Risk-Driven Development
Risk Assessment includes:
-
Systematic risk identification
-
Impact and probability evaluation
-
Mitigation strategy development
-
Continuous risk monitoring
.6. 🔄 Workflow Integration
.6.1. Recommended Usage Sequence
-
Start with Architecture Communication Canvas - Get the big picture
-
Use arc42 Chapter Generator - Create comprehensive documentation
-
Build Quality Scenarios - Define testable requirements
-
Create Architecture Decision Records - Document key decisions
-
Assess Risks - Identify and mitigate project risks
-
Track Technical Debt - Manage long-term code health
.6.2. Continuous Documentation
These prompts support:
-
Living Documentation: Keep docs updated as code evolves
-
Iterative Refinement: Improve documentation over multiple iterations
-
Team Collaboration: Involve multiple stakeholders in documentation creation
-
Knowledge Transfer: Maintain institutional knowledge
.7. 🤝 Contributing
We welcome contributions! Please:
-
Test Your Prompts: Ensure they work across different LLMs
-
Follow Structure: Use similar format to existing prompts
-
Add Examples: Include sample outputs
-
Document Usage: Explain when and how to use the prompt
-
Consider Integration: How does your prompt work with others?
.7.1. Contribution Ideas
-
Domain-specific prompts (e.g., microservices, data architecture)
-
Integration with specific tools (Jira, GitHub, etc.)
-
Visualization enhancements
-
Multilingual versions
.8. 📊 Metrics and Success
.8.1. Documentation Quality Metrics
-
Completeness of architecture coverage
-
Stakeholder understanding and feedback
-
Decision traceability
-
Risk mitigation effectiveness
.8.2. Process Improvement Metrics
-
Time to create documentation
-
Documentation maintenance effort
-
Team adoption rate
-
Knowledge transfer effectiveness
.9. 📄 License
These prompts are available under Creative Commons Attribution-ShareAlike 4.0 International License, following the arc42 licensing approach.
.10. 🔗 Related Resources
-
PlantUML for diagrams
-
Kroki for diagram generation
-
ADR GitHub Organization for decision records
These prompts are maintained by the docToolchain community. For questions or suggestions, please open an issue or contribute directly.
Feedback
Was this page helpful?
Glad to hear it! Please tell us how we can improve.
Sorry to hear that. Please tell us how we can improve.