7.5 KiB
name, description, activityWorkflowFile
| name | description | activityWorkflowFile |
|---|---|---|
| step-04-document-findings | Create a structured architecture document that answers the original questions, includes diagrams, and provides actionable recommendations | ../workflow-analysis.md |
Step 4: Document Findings
STEP GOAL:
Create a structured architecture document that answers the original questions, includes diagrams, and provides actionable recommendations.
MANDATORY EXECUTION RULES (READ FIRST):
Universal Rules:
- 🛑 NEVER generate content without user input
- 📖 CRITICAL: Read the complete step file before taking any action
- 🔄 CRITICAL: When loading next step with 'C', ensure entire file is read
- 📋 YOU ARE A FACILITATOR, not a content generator
- ✅ YOU MUST ALWAYS SPEAK OUTPUT in your Agent communication style with the config
{communication_language}
Role Reinforcement:
- ✅ You are an Implementation Partner guiding structured development activities
- ✅ If you already have been given a name, communication_style and persona, continue to use those while playing this new role
- ✅ We engage in collaborative dialogue, not command-response
- ✅ You bring software development methodology expertise, user brings domain knowledge and codebase familiarity
- ✅ Maintain clear and structured tone throughout
Step-Specific Rules:
- 🎯 Focus only on creating the final architecture document with diagrams and recommendations
- 🚫 FORBIDDEN to redo analysis — use findings from Steps 2 and 3
- 💬 Approach: Collaboratively structure and write the document with user, ensuring original questions are answered
- 📋 Include at least one Mermaid diagram and prioritized recommendations
EXECUTION PROTOCOLS:
- 🎯 Produce a complete architecture document answering the original questions
- 💾 Save the document to the project's output location
- 📖 Reference all findings from Steps 1-3
- 🚫 Do not restart analysis — synthesize existing findings
CONTEXT BOUNDARIES:
- Available context: Question and scope from Step 1; scan from Step 2; architecture map from Step 3
- Focus: Document synthesis, diagrams, risk assessment, recommendations
- Limits: No new analysis — document what was found
- Dependencies: Steps 1, 2, and 3 must be complete
Sequence of Instructions (Do not deviate, skip, or optimize)
1. Create the Architecture Document
Create a document in the output folder that answers the questions defined in Step 01. Use this structure:
# Architecture Analysis: [Project Name]
**Date:** [Date]
**Scope:** [What was analyzed]
**Questions:** [Original questions from Step 01]
## Summary
[2-3 sentence overview of key findings]
## Tech Stack
[Language, framework, database, infrastructure — from Step 02]
## Architecture Overview
[High-level description of how the system is structured]
### Architecture Diagram
[Mermaid diagram — see below]
## Component Map
[Module inventory with responsibilities — from Step 03]
## Data Flow
[End-to-end flow traces — from Step 03]
## Dependencies
[Key dependency relationships, high fan-in/fan-out modules]
## Patterns and Conventions
[Architectural and code patterns observed]
## Risks and Tech Debt
[Issues found during analysis]
## Recommendations
[Actionable next steps]
2. Include Diagrams
Use Mermaid syntax for diagrams that render in Markdown viewers.
Architecture diagram example:
graph TD
Client[Browser] --> API[API Server]
API --> Auth[Auth Service]
API --> Users[User Service]
API --> Payments[Payment Service]
Users --> DB[(PostgreSQL)]
Payments --> Stripe[Stripe API]
Auth --> Redis[(Redis)]
Dependency diagram example:
graph LR
A[Module A] --> C[Core]
B[Module B] --> C
B --> D[Module D]
D --> C
Data flow example:
sequenceDiagram
User->>Frontend: Click "Submit"
Frontend->>API: POST /orders
API->>Validation: Validate input
API->>OrderService: Create order
OrderService->>Database: INSERT order
OrderService->>PaymentService: Charge card
PaymentService->>Stripe: Create charge
Stripe-->>PaymentService: Charge confirmed
PaymentService-->>API: Payment success
API-->>Frontend: 201 Created
Frontend-->>User: Show confirmation
3. Document Risks and Tech Debt
For each risk or debt item found during analysis:
| Risk | Severity | Location | Impact |
|---|---|---|---|
No input validation on /api/admin/ routes |
High | src/routes/admin.ts |
Security vulnerability |
| Circular dependency between User and Order modules | Medium | src/services/ |
Fragile, hard to test |
| No error handling in payment flow | High | src/payments/stripe.ts |
Silent failures |
| Outdated dependencies (2 major versions behind) | Medium | package.json |
Security + compatibility |
4. Write Recommendations
Each recommendation should be:
- Specific — Name the module, file, or pattern
- Actionable — Describe what to do, not just what is wrong
- Prioritized — Order by impact and effort
Example:
## Recommendations
1. **Add input validation to admin routes** (High priority, low effort)
- Add Zod schemas to all `/api/admin/` endpoints
- Estimated: 2-3 hours
2. **Break circular dependency between User and Order** (Medium priority, medium effort)
- Extract shared types to a common module
- Use dependency injection for cross-service calls
- Estimated: 4-6 hours
3. **Add error handling to payment flow** (High priority, medium effort)
- Wrap Stripe calls in try-catch with structured error responses
- Add retry logic for transient failures
- Estimated: 3-4 hours
5. Save Output
Save the document to the project's output location. If no output location is defined, save to:
docs/architecture/in the analyzed project, or- The agent experiences folder for session insights
6. Verify Checklist
- Document answers the original questions from Step 01
- Summary is clear and concise (2-3 sentences)
- At least one Mermaid diagram included
- Risks and tech debt listed with severity
- Recommendations are specific, actionable, and prioritized
- Document saved to output folder
7. Present MENU OPTIONS
Display: "Select an Option: [M] Return to Activity Menu"
Menu Handling Logic:
- IF M: Update design log, then load, read entire file, then execute {activityWorkflowFile}
- IF Any other comments or queries: help user respond then [Redisplay Menu Options]
EXECUTION RULES:
- ALWAYS halt and wait for user input after presenting menu
- ONLY proceed when user selects 'M'
- User can chat or ask questions - always respond and then redisplay menu options
CRITICAL STEP COMPLETION NOTE
ONLY WHEN the architecture document is complete and saved will you then load and read fully {activityWorkflowFile} to execute.
🚨 SYSTEM SUCCESS/FAILURE METRICS
✅ SUCCESS:
- Document answers the original questions from Step 01
- Summary is clear and concise (2-3 sentences)
- At least one Mermaid diagram included
- Risks and tech debt listed with severity
- Recommendations are specific, actionable, and prioritized
- Document saved to output folder
❌ SYSTEM FAILURE:
- Document does not answer original questions
- No diagrams included
- Recommendations are vague or not prioritized
- Document not saved
Master Rule: Skipping steps, optimizing sequences, or not following exact instructions is FORBIDDEN and constitutes SYSTEM FAILURE.