Prompt file imported from danilonovaisv/SCENES-GENERATOR (
.windsurf/workflows/core/feature/feature-request.md). Copyright stays with the author.
Feature Request Workflow
The Feature Request system provides comprehensive tracking and annotation capabilities for feature requests with granular command aliases and automatic code/shell script annotation.
Usage
Granular Command Aliases
The system supports multiple levels of command granularity:
# Full form
max:feature-request create "Add new feature" "Description here" --priority high
# Short form
max:fr list --status open
# Minimal form
m:fr show FR-20250220-001
# Legacy forms
feature-request update FR-20250220-001 --status in_progress
fr create "Quick feature" "Simple description"
Basic Operations
Create Feature Request
# Basic creation
max:feature-request create "Add dark mode" "Implement dark theme for UI" --priority high
# With tags
max:fr create "API endpoints" "Add REST API for user management" --tags api,backend --priority medium
# From file
m:fr create "Complex feature" --file feature_description.md --priority critical
List and Filter
# List all requests
max:feature-request list
# Filter by status
max:fr list --status open
# Filter by priority
feature-request list --priority high
# Combined filters
m:fr list --status open --priority high
View Details
# Show specific request
max:feature-request show FR-20250220-001
# Short form
m:fr show FR-20250220-001
Update Requests
# Update status
max:fr update FR-20250220-001 --status in_progress
# Update priority
feature-request update FR-20250220-001 --priority critical
# Both status and priority
m:fr update FR-20250220-001 --status completed --priority low
Delete Requests
# Delete a request
max:feature-request delete FR-20250220-001
Annotation Features
Automatic Code Block Annotation
The system automatically detects and annotates code blocks in feature request descriptions:
Python Code Annotation
def process_data(data):
# 📝 Function: process_data - defines reusable logic
for item in data:
# 📝 Loop - iterates over sequence
if item.is_valid():
# 📝 Conditional logic - controls program flow
yield transform(item)
Shell Command Annotation
cd /Users/ethandellaposta/Documents/git/MAx
# 📋 Navigate to directory: /Users/ethandellaposta/Documents/git/MAx
git log --oneline --since="2 weeks ago"
# 📋 Git operation: log
&& git show --name-only HEAD
# 📋 Chain command - execute if previous succeeds
Mixed Content Annotation
The system handles mixed content with multiple annotation types:
Based on my analysis, I can see that the workflow needs updates.
Command cd, git
cd /Users/ethandellaposta/Documents/git/MAx && git log --oneline
# 📋 Navigate to directory: /Users/ethandellaposta/Documents/git/MAx
# 📋 Chain command - execute if previous succeeds
# 📋 Git operation: log
```python
# 📝 Import: specific module import - targeted functionality
from feature_request import FeatureRequestSystem
def main():
# 📝 Function: main - defines reusable logic
system = FeatureRequestSystem()
# 📝 Import: FeatureRequestSystem - brings in external functionality
Command Reference
Actions
| Action | Description | Example |
|---|---|---|
create |
Create new feature request | max:fr create "Title" "Description" |
list |
List feature requests | max:fr list --status open |
show |
Show specific request | max:fr show FR-20250220-001 |
update |
Update request status/priority | max:fr update FR-001 --status in_progress |
delete |
Delete a request | max:fr delete FR-20250220-001 |
Options
| Option | Description | Values |
|---|---|---|
--title |
Request title | String |
--description |
Request description | String |
--priority |
Priority level | low, medium, high, critical |
--status |
Status filter | open, in_progress, completed, rejected |
--tags |
Categorization tags | List of strings |
--file |
Read description from file | File path |
Alias Hierarchy
max:feature-request (Full form - most explicit)
↓
max:fr (Short form - balanced)
↓
m:fr (Minimal form - quickest)
↓
feature-request (Legacy form - compatible)
↓
fr (Shortest form - minimal)
Annotation Patterns
Code Block Detection
The system detects:
- Markdown code blocks with language specification
- Shell command blocks with command prefixes
- Mixed content with multiple annotation types
Annotation Types
| Type | Trigger | Example |
|---|---|---|
| Function definition | def or class |
# 📝 Function: name - defines reusable logic |
| Import statements | import or from |
# 📝 Import: module - brings in external functionality |
| Control structures | if , for , while |
# 📝 Conditional logic - controls program flow |
| Shell commands | cd , git , ls |
# 📋 Navigate to directory: path |
| Command chaining | &&, ` |
Annotation Logic
The annotation system:
- Detects blocks using regex patterns
- Analyzes content for specific patterns
- Adds contextual annotations after relevant lines
- Preserves original code while adding insights
- Handles multiple languages (Python, Bash, JavaScript, etc.)
Integration
With Git Hooks
Add to pre-commit hooks for feature request validation:
# .pre-commit-config.yaml
repos:
- repo: local
hooks:
- id: feature-request-check
name: Feature Request Validation
entry: python3 .max/feature_request.py list
language: system
With IDE
The command aliases work directly in IDE command palette:
max:feature-request- Full feature request operationsmax:fr- Quick feature request actionsm:fr- Rapid feature request management
With CI/CD
Integrate with GitHub Actions:
# .github/workflows/feature-requests.yml
name: Feature Request Check
on: [pull_request]
jobs:
check-requests:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v2
- name: List open feature requests
run: python3 .max/feature_request.py list --status open
Data Storage
Feature requests are stored in .max/feature_requests.json with the following structure:
[
{
"id": "FR-20250220-001",
"title": "Feature title",
"description": "Detailed description with code blocks",
"priority": "high",
"status": "open",
"created_at": "2025-02-20T12:00:00",
"updated_at": "2025-02-20T12:00:00",
"annotations": ["Found code_block at lines 5-10"],
"code_blocks": [
{
"type": "code_block",
"language": "python",
"original": "def example():\n pass",
"annotated": "def example():\n # 📝 Function: example - defines reusable logic\n pass",
"line_start": 5,
"line_end": 6
}
],
"tags": ["enhancement", "ui"]
}
]
Best Practices
1. Use Granular Aliases
- Full form for documentation and scripts:
max:feature-request - Short form for daily use:
max:fr - Minimal form for quick operations:
m:fr
2. Write Annotatable Descriptions
Include code blocks and commands in descriptions:
Need to implement the following:
```python
def new_feature():
# Implementation here
pass
Shell commands for testing:
cd project && python test_new_feature.py
3. Use Tags Effectively
Categorize requests for better organization:
max:fr create "API feature" "Description" --tags api,backend,high-priority
4. Regular Status Updates
Keep requests current:
m:fr update FR-001 --status in_progress
# Work on feature...
m:fr update FR-001 --status completed
Examples
Complete Workflow
# Create feature request with detailed description
max:feature-request create "Add user authentication" """
Implement OAuth2 authentication with the following:
```python
# 📝 Import: specific module import - targeted functionality
from authlib.integrations.flask_client import OAuth
def setup_oauth(app):
# 📝 Function: setup_oauth - defines reusable logic
oauth = OAuth(app)
# Configure OAuth providers...
Shell commands for setup:
# 📋 Navigate to directory: /project
cd /project
# 📋 Git operation: checkout
git checkout -b feature/auth
# 📋 Chain command - execute if previous succeeds
&& pip install authlib
""" --priority high --tags security,authentication
List high priority open requests
max:fr list --status open --priority high
Update status when starting work
m:fr update FR-20250220-001 --status in_progress
Mark as completed when done
feature-request update FR-20250220-001 --status completed
### Quick Operations
```bash
# Quick status check
m:fr list
# Fast feature creation
fr create "Quick fix" "Minor bug fix" --priority low
# Rapid status update
m:fr update FR-001 --status completed
🚀 The Feature Request system provides comprehensive tracking with intelligent annotation and granular command control!
