AI SDLC - Python template

A Cookiecutter template for Python applications and libraries built with a controlled, approval-gated, AI-assisted Software Development Life Cycle workflow.

Why this template

This template helps teams start Python projects with:

• a reproducible repository layout;
• built-in quality tooling;
• skill-driven SDLC stages;
• human approval gates between workflow stages;
• centrally maintained Agent Skills that can be updated independently of the template.

Requirements

Before generating a project, install:

• Python;
• Cookiecutter;
• Node.js and npm.

Node.js/npm are required because the template installs the AI SDLC Agent Skills during project creation with the skills CLI through npx.

Verify that npx is available:

npx --version

If npx is not available, install a current Node.js release using your operating system package manager or the official Node.js distribution.

Quick start

1) Install Cookiecutter

Choose one option:

uv tool install cookiecutter

pipx install cookiecutter

python -m pip install cookiecutter

2) Verify npx

npx --version

3) Generate a project

cookiecutter https://github.com/ecarrenolozano/ai-assisted-python-template.git

During project generation, the template downloads the AI SDLC skills from:

https://github.com/ecarrenolozano/ai-sdlc-skills

The skills are installed with:

npx skills add ecarrenolozano/ai-sdlc-skills --all

4) Run initial checks

cd <your-project-slug>
uv sync --all-groups
uv run pre-commit install
uv run pytest

Cookiecutter basics for first-time users

Cookiecutter asks for prompt values and generates a new repository from this template.

Run with prompts:

cookiecutter https://github.com/ecarrenolozano/ai-assisted-python-template.git

Reuse previous answers:

cookiecutter https://github.com/ecarrenolozano/ai-assisted-python-template.git --replay

What you get

• src/ package layout with Hatchling;
• uv, pytest, pytest-bdd, Ruff, mypy, coverage, and pre-commit configuration;
• reusable AI SDLC Agent Skills installed during project creation;
• a provider-neutral canonical skills location under .agents/skills/;
• compatibility with supported AI coding agents through the skills CLI;
• staged SDLC documentation under sdlc_docs/;
• workflow traceability and approval-gated transitions;
• documentation placeholders for MkDocs or Zensical;
• issue and change-request templates.

The Agent Skills are maintained separately from this Cookiecutter template in the ai-sdlc-skills repository. This allows existing generated projects to receive newer versions of the workflow without regenerating the project.

Updating Agent Skills

Generated projects can update their installed skills independently of the Cookiecutter template.

From the root of a generated project, run:

npx skills update

You do not need to regenerate the project to receive skill updates.

SDLC navigation in generated projects

Start here after generation:

1. sdlc_docs/trace_workflow.md
2. sdlc_docs/00_inception/sources/
3. WORKFLOW.md

The Agent Skills under .agents/skills/ drive stage transitions from inception through release preparation.

Troubleshooting

• cookiecutter: command not found: reopen your shell, or run with uvx cookiecutter.
• npx: command not found: install Node.js/npm and verify with npx --version.
• Skill installation fails during project generation: verify internet access and run npx skills add ecarrenolozano/ai-sdlc-skills --all from the generated project directory.
• Destination folder already exists: choose another output folder, or remove the old generated repository.
• Wrong prompt values: re-run with --replay or pass explicit key-value arguments.

Template scope

This template provides reusable process and infrastructure. Project-specific requirements, architecture, code, datasets, and exercises are added after generation.

---

Authors

Edwin Carreño	Scientific Software Center	Heidelberg, Germany
Maxim Scheremetjew	Max Plank Institute of Molecular Cell Biology and Genetics	Dresden, Germany