Appearance
Custom Workflows
Responsible Vibe supports custom workflow definitions that you can create, install, and share. The system uses a directory-based approach with domain filtering for better organization.
How Custom Workflows Work
The .vibe/workflows/ Directory
Custom workflows are stored in your project's .vibe/workflows/ directory:
.vibe/
├── workflows/
│ ├── my-custom-workflow.yaml
│ ├── team-review-process.yaml
│ └── client-specific-flow.yaml
└── development-plan-main.mdWorkflow Domains
Workflows are organized by domains to keep things manageable:
code: Software development workflows (default)architecture: System design and architecture workflowsoffice: Business process and documentation workflows
Control which domains are loaded:
bash
export VIBE_WORKFLOW_DOMAINS="code,architecture"
# Only loads workflows from code and architecture domainsCreating Custom Workflows
1. Basic Workflow Structure
Create a YAML file in .vibe/workflows/:
yaml
name: 'my-custom-workflow'
description: 'Custom workflow for my specific needs'
initial_state: 'start'
metadata:
domain: 'code'
complexity: 'medium'
bestFor: ['Custom processes', 'Team workflows']
useCases: ['Specific project needs']
examples: ['Custom review process']
states:
start:
description: 'Initial phase'
instructions: |
Start your custom process here.
Define what the AI should focus on in this phase.
transitions:
- trigger: 'ready_for_next'
to: 'next_phase'
transition_reason: 'Ready to move forward'
next_phase:
description: 'Next phase'
instructions: |
Continue with the next step of your process.
transitions:
- trigger: 'workflow_complete'
to: 'start'
transition_reason: 'Workflow complete, ready for new task'2. Advanced Features
Phase-specific instructions:
yaml
states:
design:
instructions: |
You are in the design phase. Focus on:
- System architecture decisions
- Component interfaces
- Data flow design
Reference existing architecture: $ARCHITECTURE_DOC
Document your design in: $DESIGN_DOCConditional transitions:
yaml
transitions:
- trigger: 'design_approved'
to: 'implementation'
additional_instructions: 'Call setup_project_docs to create implementation templates before starting'
- trigger: 'need_more_design'
to: 'requirements'
additional_instructions: 'Review and update requirements.md based on design feedback'Installing Workflows
Using CLI
The CLI provides convenient commands to list and copy workflows:
bash
# List all available workflows
npx responsible-vibe-mcp workflow list
# Copy a built-in workflow to customize it
npx responsible-vibe-mcp workflow copy waterfall my-custom-waterfall
# Copy any available workflow
npx responsible-vibe-mcp workflow copy epcc my-team-processThis copies the workflow to .vibe/workflows/my-custom-waterfall.yaml where you can customize it for your needs.
The copied workflow will have:
- Updated
namefield to match your custom name - All original states and transitions
- Metadata you can modify (description, complexity, etc.)
From URLs or Files
bash
# Install from URL
"Install workflow from https://example.com/my-workflow.yaml"
# Install with custom name
"Install the waterfall workflow as 'detailed-waterfall'"Workflow Discovery
List Available Workflows
bash
# See all workflows available to your project
"List available workflows"
# See all workflows regardless of domain filtering
"List all workflows including unloaded ones"Domain Filtering in Action
bash
# Default: only 'code' domain workflows
VIBE_WORKFLOW_DOMAINS="code"
# Multiple domains
VIBE_WORKFLOW_DOMAINS="code,architecture,office"
# All domains
VIBE_WORKFLOW_DOMAINS="code,architecture,office"Project-Specific Configuration
.vibe/config.yaml
Control which workflows are available for your project:
yaml
enabled_workflows:
- 'waterfall'
- 'my-custom-workflow'
- 'team-review-process'This filters the available workflows to only those specified, regardless of domain settings.
Workflow Metadata
Enhanced Discoverability
yaml
metadata:
domain: 'code' # Which domain this belongs to
complexity: 'high' # low, medium, high
bestFor: # What this workflow is good for
- 'Large features'
- 'Design-heavy projects'
useCases: # Specific use cases
- 'Building new systems'
- 'Complex integrations'
examples: # Example scenarios
- 'Create authentication system'
- 'Build reporting dashboard'This metadata helps your AI automatically select the right workflow for different scenarios.
Real-World Example
yaml
name: 'api-development'
description: 'Workflow for developing REST APIs with proper testing'
initial_state: 'api_design'
metadata:
domain: 'code'
complexity: 'medium'
bestFor: ['API development', 'Backend services']
useCases: ['REST API creation', 'Microservice development']
states:
api_design:
description: 'Design API endpoints and contracts'
instructions: |
Design your API:
- Define endpoints and HTTP methods
- Specify request/response schemas
- Document authentication requirements
- Plan error handling approach
Document in $DESIGN_DOC
transitions:
- trigger: 'api_design_complete'
to: 'implementation'
additional_instructions: 'Create API implementation templates and set up testing framework'
implementation:
description: 'Implement API endpoints'
instructions: |
Implement the API following your design:
- Create route handlers
- Implement business logic
- Add input validation
- Include proper error handling
transitions:
- trigger: 'implementation_complete'
to: 'testing'
additional_instructions: 'Set up test environment and create test data fixtures'
testing:
description: 'Test API endpoints'
instructions: |
Test your API thoroughly:
- Unit tests for business logic
- Integration tests for endpoints
- Test error scenarios
- Validate against API design
transitions:
- trigger: 'testing_complete'
to: 'api_design'
additional_instructions: 'Document API completion and prepare for next API development cycle'
transition_reason: 'API complete, ready for next API'Why This System Works
Directory-based: Easy to see and manage all your custom workflows
Domain filtering: Only load workflows relevant to your work
Project-specific: Each project can have its own custom workflows
Shareable: Workflows can be installed from URLs or shared between projects
Discoverable: Rich metadata helps AI select appropriate workflows
Your custom workflows integrate seamlessly with the built-in ones, giving you complete control over your development process.
Next: Git Commits – Configure automatic commit behavior