This guide walks you through setting up the AI-Native SDLC environment on your machine. We will set up Spec-Kit for planning and Superpowers for execution.
Before installing the tools, make sure you have the following installed on your system:
uv (Fast Python Package Manager)Spec-Kit is distributed as a Python package. We recommend using uv to manage Python CLI tools because it is significantly faster and more isolated than standard pip.
# macOS and Linux
curl -LsSf https://astral.sh/uv/install.sh | sh
# Windows (PowerShell)
powershell -c "irm https://astral.sh/uv/install.ps1 | iex"
Verify that uv is installed:
uv --version
# Expected: uv x.y.z
[!NOTE] Windows users: After installing
uv, close and reopen your terminal before runninguv --version. To reload the PATH without restarting, run:$env:PATH = [System.Environment]::GetEnvironmentVariable("PATH","Machine") + ";" + [System.Environment]::GetEnvironmentVariable("PATH","User")
Use uv to install the Spec-Kit CLI tool globally. This ensures that the CLI and its dependencies are placed in an isolated, dedicated environment.
uv tool install specify-cli --from git+https://github.com/github/spec-kit.git
Verify that Spec-Kit is installed:
specify --version
# Expected: specify-cli x.y.z
Superpowers is installed directly inside your AI coding agent (e.g. Claude Code, Cursor, Gemini CLI).
graph TD
A[Choose AI Agent] --> B(Claude Code)
A --> C(Cursor)
A --> D(Gemini CLI)
A --> E(Codex CLI)
B --> F["/plugin install superpowers@claude-plugins-official"]
C --> G["/add-plugin superpowers"]
D --> H["gemini extensions install https://github.com/obra/superpowers"]
E --> I["Use /plugins -> Search 'superpowers' -> Click Install"]
[!IMPORTANT] Command accuracy: Plugin install commands change as agents evolve. If the command above fails for your agent, check the official Superpowers repository for the latest install instructions for your specific agent version.
After installing, type the following in your agentβs chat interface to confirm itβs active:
/sp-status
If Superpowers is installed correctly, you will see a list of available skills including using-git-worktrees, test-driven-development, and requesting-code-review.
Once the tools are installed globally, you must initialize Spec-Kit in your project root directory.
If you are starting a completely new project in a blank directory, run:
mkdir my-new-project
cd my-new-project
specify init . --integration claude
[!NOTE] Change
--integration claudeto matches your target agent (e.g.gemini,copilot,cursor).
If you are adding AI-assisted workflows to an existing project, navigate to the project directory and run:
cd my-existing-project
specify init . --force --integration claude
[!IMPORTANT] The
--forceflag is safe to use. It only creates the.specify/directory structure and does not modify or delete any of your existing source files.
After initialization, you should see a new .specify/ folder in your project root containing:
your-project/
βββ .specify/
βββ memory/
β βββ constitution.md # [Editable] Project principles & stack rules
βββ specs/ # Contains individual feature specs
βββ templates/ # Custom formatting templates
βββ scripts/ # Integration bash/PowerShell scripts
If these files are present, you are successfully set up and ready to create your first specification!