Skip to main content

8. Example Workflows

Learn how to apply Athanor's features in real-world scenarios with these step-by-step workflow examples, from simple feature additions to complex system integrations.

Simple Feature: Dark Mode Toggle​

Let's implement a dark mode toggle for a web application.

Step 1: Define the Task​

Task Description:

Implement a dark mode toggle switch in the settings panel that saves the user's preference to localStorage and applies the theme immediately.

Step 2: Smart File Selection​

Instead of manually selecting files, let's use Athanor's intelligence:

  1. Don't select any files initially
  2. Click "Autoselect" prompt template
  3. Copy the generated prompt to your AI assistant
  4. Paste response back into Athanor
  5. Click "Apply AI Output" to auto-select relevant files

The AI might suggest files like:

  • components/SettingsPanel.tsx
  • hooks/useTheme.ts
  • styles/globals.css
  • contexts/ThemeContext.tsx

Step 3: Implement the Feature​

  1. Review the auto-selected files
  2. Click "Coder" prompt template
  3. Copy prompt to AI assistant
  4. Receive implementation with proper React hooks, CSS variables, and localStorage persistence

Step 4: Apply Changes​

  1. Copy AI response
  2. Click "Apply AI Output"
  3. Review each proposed change in the diff viewer
  4. Accept changes that look correct
  5. Reject any unwanted modifications

Expected Outcome​

  • New theme toggle component
  • Hook for theme management
  • CSS variables for dark/light modes
  • Persistent user preferences

Complex Feature: Stripe Payment Integration​

Let's integrate Stripe for handling monthly user subscriptions with a multi-commit approach.

Phase 1: Architecture Planning​

Task Description:

Integrate Stripe for handling monthly user subscriptions. This should include:
- Creating subscription plans
- Handling webhooks for payment success/failure
- Updating user subscription status
- Managing billing portal access
- Error handling and user feedback

Step 1: Get the Architecture Plan​

  1. Click "Autoselect" to identify relevant files
  2. Apply AI Output to select suggested files
  3. Click "Architect" prompt template
  4. Copy to AI assistant for strategic planning

The AI will provide a commit-by-commit breakdown:

Commit 1: Setup Stripe SDK and API keys Commit 2: Implement subscription plan management Commit 3: Create checkout session handling Commit 4: Implement webhook endpoint Commit 5: Build user billing portal Commit 6: Add error handling and UI feedback

Step 2: Copy Architecture to Context​

  1. Copy the AI's architecture response
  2. Paste the commit plan into the Context field
  3. This will be included in subsequent prompts

Phase 2: Implementation (Commit by Commit)​

Commit 1: Setup Stripe SDK​

  1. Select "Commit 1" from Task Context: 
    Implement Commit 1: Setup Stripe SDK and API keys.
    Configure environment variables and initialize Stripe client.
  2. Use existing file selection (you rarely need to rerun Autoselect)
  3. Click "Coder" prompt
  4. Apply changes after review

Commit 2: Subscription Plans​

  1. Either:
    • Continue in your current AI chat by telling the AI "Continue with Commit 2"; or
    • Select "Commit 2" from Task Context, click "Coder prompt" and paste in a new chat
    • This choice depends on the complexity and length of the current chat
  2. Review and apply changes

Continue Pattern for Remaining Commits​

Repeat the process for each commit:

  • Update task context with current commit (or continue in the chat)
  • Use "Coder" prompt for implementation
  • Apply and test changes
  • Move to next commit

Phase 3: Integration Testing​

After completing all commits:

  1. Task Description: 
    Review the complete Stripe integration implementation and provide:
    - Integration testing checklist
    - Potential security considerations
    - Performance optimization opportunities
    - Error scenarios to test
  2. Select all modified files
  3. Use "Query" prompt for analysis
  4. Follow AI recommendations for testing

Understanding Existing Code: Authentication Flow​

Learn how to use Athanor to understand unfamiliar codebases.

Step 1: Broad Overview​

Task Description:

Explain the current user authentication flow, including:
- Token generation and validation
- Session management
- Role-based access control
- Security measures implemented
- Which files are primarily involved?

Process:​

  1. Click "Autoselect" to find auth-related files
  2. Apply AI Output to select suggested files
  3. Click "Query" prompt for analysis-focused output
  4. Review AI's explanation of the authentication system

Step 2: Deep Dive into Specific Components​

Based on the AI's overview, dive deeper on the JWT token validation middleware.

Process:​

  1. Select specific middleware files identified in Step 1
  2. Use the copy action from the File Manager to copy just these files in the existing chat
  3. Ask follow-up questions in the same AI chat

Step 3: Document Your Understanding​

Process:​

  1. Ask the AI for a Summary of your chat
  2. Paste the summary into the Task Description
  3. Autoselect all authentication-related files
  4. Use "Writer" prompt for documentation generation
  5. Apply changes to create/update documentation files

Best Practices for Workflow Success​

Planning Phase​

  1. Start with Autoselect when unsure about file selection
  2. Use Architect for complex features requiring multiple steps
  3. Break large tasks into smaller, manageable commits
  4. Document your plan in the Context field

Implementation Phase​

  1. Review AI suggestions carefully before applying
  2. Test incrementally after each commit
  3. Use version control to track changes
  4. Ask follow-up questions when AI responses are unclear
  5. Stay in the chat if the context is still relevant

Quality Assurance​

  1. Use Query prompts to verify understanding
  2. Generate tests for critical functionality
  3. Review security implications of changes
  4. Document important decisions and changes

Team Collaboration​

  1. Share custom templates that work well for your team
  2. Use consistent commit messages when applying changes
  3. Document workflows that prove effective
  4. Train team members on successful patterns

Troubleshooting Common Workflow Issues​

AI Responses Don't Match Expectations​

  • Refine task descriptions with more specific requirements
  • Add relevant context about coding standards or patterns
  • Include examples of desired output format
  • Break complex requests into smaller parts

Changes Don't Apply Cleanly​

  • Check that files haven't changed since generating the prompt
  • Regenerate prompts with current file state
  • Review merge conflicts carefully
  • Apply changes in smaller batches

Workflow Takes Too Long​

  • Use preset tasks for common operations
  • Create custom templates for repeated workflows
  • Leverage API integration for simple requests
  • Batch similar operations together

Results Lack Project Context​

  • Include PROJECT.md or documentation files in selection
  • Add project-specific requirements to task descriptions
  • Use project-specific custom templates
  • Provide architectural context in prompts

Ready to master Athanor? Practice these workflows with your own projects and experiment with creating custom templates that match your team's specific needs!