https://github.com/mpuig/claude-code-py-template
Specialized Claude Code agent prompts designed to systematically build proof of concept applications using Python
https://github.com/mpuig/claude-code-py-template
claude-code generative-ai
Last synced: 4 months ago
JSON representation
Specialized Claude Code agent prompts designed to systematically build proof of concept applications using Python
- Host: GitHub
- URL: https://github.com/mpuig/claude-code-py-template
- Owner: mpuig
- License: mit
- Created: 2025-07-06T16:02:05.000Z (8 months ago)
- Default Branch: main
- Last Pushed: 2025-07-06T16:10:51.000Z (8 months ago)
- Last Synced: 2025-09-17T02:55:04.464Z (5 months ago)
- Topics: claude-code, generative-ai
- Homepage: https://medium.com/@mpuig/from-chaos-to-code-how-i-built-a-systematic-approach-to-python-proof-of-concepts-with-claude-code-e55c8d5c1a0f
- Size: 39.1 KB
- Stars: 1
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Claude Code Commands for Python PoC's
This repository contains specialized Claude Code agent prompts designed to systematically build proof of concept applications using Python. These agents work together to analyze requirements, design architecture, and implement complete working systems.
## Overview
These agent commands provide a structured approach to software development, taking you from initial requirements to a working proof of concept through a series of specialized agents:
1. **Problem Analysis** - Deep requirements understanding and problem decomposition
2. **Architecture Design** - Comprehensive system design and technology selection
3. **Task Breakdown** - Converting architecture into actionable implementation tasks
4. **Detailed Planning** - Step-by-step implementation guides
5. **Implementation** - Systematic code development and testing
## Usage
To use these agents, start with your requirements and trigger the software architect agent:
```
I want you to implement a basic clone of called "". It allows .
In terms of tech stack use python on the server, and for the frontend i want you to use react.
For design use tailwind but make sure to make nice components too so it's not a mess.
```
Then trigger the main orchestrator:
```
/software_architect_agent
```
## Agent Commands
### Core Agents
- **`software_architect_agent`** - Main orchestrator that coordinates the entire development process
- **`problem_analysis_agent`** - Analyzes requirements and creates comprehensive problem statements
- **`architecture_design_agent`** - Designs system architecture and selects technologies
- **`task_breakdown_agent`** - Breaks down architecture into implementation phases
- **`detailed_planning_agent`** - Creates detailed step-by-step implementation plans
- **`implementation_agent`** - Executes implementation following the plans
- **`documentation_agent`** - Creates comprehensive project documentation
### Supporting Agents
- **`programming_lead_agent`** - Research lead for high-level research strategy and planning
- **`research_subagent`** - Specialized research tasks and information gathering
- **`citations_agent`** - Adds proper citations to research reports
## Python-Focused Features
These agents are specifically configured for Python development:
- **Frameworks**: FastAPI, Django, Flask
- **Testing**: pytest, unittest with comprehensive test strategies
- **Code Quality**: black (formatting), ruff (linting), mypy (type checking)
- **Package Management**: uv (preferred), poetry, conda
- **Project Structure**: Standard Python project layout (src/, tests/, docs/)
- **Database**: SQLAlchemy, Django ORM, migrations
- **Deployment**: Docker, cloud services, uv virtual environments
## Example Workflow
1. **Start** with your project requirements
2. **Trigger** `/software_architect_agent` to begin the process
3. **Problem Analysis** - Agent analyzes and documents requirements
4. **Architecture Design** - Agent creates comprehensive system design
5. **Task Breakdown** - Agent creates implementation phases
6. **Detailed Planning** - Agent creates step-by-step guides
7. **Implementation** - Agents systematically build the system
8. **Testing & Integration** - Continuous testing throughout development
9. **Documentation** - Agent creates comprehensive project documentation
## Key Principles
- **Backend-First**: Database models → API endpoints → Frontend
- **Incremental Development**: Build and test components progressively
- **Python Best Practices**: PEP 8, type hints, proper project structure
- **Proof of Concept Focus**: Prioritize core functionality over edge cases
- **Systematic Approach**: Never skip phases or implement out of order
## Credits
Original agent prompts were inspired by and adapted from [vibe-minisentry](https://github.com/mitsuhiko/vibe-minisentry) by Armin Ronacher. As noted in the [project issues](https://github.com/mitsuhiko/vibe-minisentry/issues/1), these agent prompts were originally generated by Claude Code itself using Opus, taking research prompts as input and creating specialized prompts for expert software architects and engineers.
This Python-focused version builds upon that foundation with Python-specific tooling, conventions, and best practices.
## Requirements
- Claude Code CLI
- Python 3.9+
- uv (recommended for fast package management)
- Basic understanding of Python development
## Getting Started
1. Place these command files in your `.claude/commands/` directory
2. Start Claude Code in your project directory
3. Provide your project requirements
4. Run `/software_architect_agent` to begin the systematic development process
The agents will guide you through the entire development lifecycle, from requirements analysis to working proof of concept.