Starter templates for Cognis Digital projects. Copy a template, rename it, and ship.
Each template is self-contained, opinionated, and production-ready out of the box: modern tooling, sensible defaults, no boilerplate to delete.
| Template | Path | What you get |
|---|---|---|
| Python CLI tool | python-cli/ |
A pyproject.toml-based CLI with argparse subcommands, packaged entry point, and pytest. |
| MCP server (Python) | mcp-server-python/ |
A Model Context Protocol server using the official mcp SDK, exposing tools over stdio. |
| Dockerfile | docker/Dockerfile |
A multi-stage, non-root, slim Python image template with a healthcheck. |
| CI workflow | .github/workflows/ci.yml |
GitHub Actions: lint (ruff), type-check (mypy), test (pytest) across a matrix. |
| Dev container | .devcontainer/ |
VS Code / Codespaces devcontainer with Python, uv, and pre-wired extensions. |
| Project README | templates/README.template.md |
A fill-in-the-blanks README for a new project. |
| Issue / PR templates | .github/ |
Bug report, feature request, config, and a PR checklist. |
- Pick the template directory you want.
- Copy it into your new repo (e.g.
cp -r python-cli/ ~/projects/my-tool). - Search-and-replace the placeholder names:
cognis_tool/cognis-tool-> your package / project nameCognis Digital-> kept as-is for first-party repos
- Read the template's own
README.mdfor next steps.
- Python 3.11+ as the floor.
pyproject.tomlis the single source of project config (nosetup.py, nosetup.cfg).- uv is the preferred installer/runner, with plain
pipalways working as a fallback. - ruff for both linting and formatting.
- mypy for type checking, strict where practical.
- pytest for tests.
- Containers run as a non-root user and are multi-stage to keep images small.
flowchart LR
U[You / CI / Agent] --> R[templates]
R --> O[Outputs & artifacts]
R --> M[MCP / JSON]
M --> AI[AI agents]
R --> S[Cognis Neural Suite]
Explore the suite → 🗂️ all tools · ⭐ awesome-cognis · 🔗 cognis-sources
MIT. See LICENSE.