Skip to content

DEVELOPER.md

Latest commit

 

History

History
104 lines (74 loc) · 4.08 KB

DEVELOPER.md

File metadata and controls

104 lines (74 loc) · 4.08 KB

DEVELOPER.md

This document provides instructions for setting up your development environment and contributing to the Data Agent Kit Starter Pack project.

Prerequisites

Before you begin, ensure you have the following:

  1. Gemini CLI: Install the Gemini CLI version v0.6.0 or above. Installation instructions can be found on the official Gemini CLI documentation. You can verify your version by running gemini --version.
  2. Cloud SQL for PostgreSQL Instance: For testing data plane tools, you will need access to an active Cloud SQL for PostgreSQL instance.

Developing the Extension

Running from Local Source

The core logic for this extension is handled by a pre-built toolbox binary. The development process involves installing the extension locally into the Gemini CLI to test changes.

  1. Clone the Repository:

    git clone https://github.com/gemini-cli-extensions/data-agent-kit-starter-pack.git
cd data-agent-kit-starter-pack

  2. Download the Toolbox Binary: The required version of the toolbox binary is specified in toolbox_version.txt. Download it for your platform.

    # Read the required version
VERSION=$(cat toolbox_version.txt)

# Example for macOS/amd64
curl -L -o toolbox https://storage.googleapis.com/genai-toolbox/geminicli/v$VERSION/darwin/amd64/toolbox
chmod +x toolbox

    Adjust the URL for your operating system (linux/amd64, darwin/arm64, windows/amd64).

  3. Install the Extension Locally: Use the Gemini CLI to install the extension from your local directory.

    gemini extensions install .

    The CLI will prompt you to confirm the installation. Accept it to proceed.

  4. Testing Changes: After installation, start the Gemini CLI (gemini). You can now interact with the data-agent-kit-starter-pack tools to manually test your changes against your connected database.

Adding or Updating Skills

To add a new skill or update an existing one:

  1. Navigate to the skills/ directory.
  2. Create or update the skill folder (must be kebab-case).
  3. Ensure there is a SKILL.md with the required frontmatter (name and description).
  4. Validate your changes locally by running gemini extensions validate ..
  5. Create a pull request following the Conventional Commits format (e.g., feat: add new bigquery optimization skill).

Testing

Automated Presubmit Checks

A GitHub Actions workflow (.github/workflows/presubmit-tests.yml) is triggered for every pull request. This workflow primarily verifies that the extension can be successfully installed by the Gemini CLI.

Currently, there are no automated unit or integration test suites within this repository. All functional testing must be performed manually. All tools are currently tested in the MCP Toolbox GitHub.

Other GitHub Checks

  • License Header Check: A workflow ensures all necessary files contain the proper license header.
  • Conventional Commits: This repository uses Release Please to manage releases. Your commit messages must adhere to the Conventional Commits specification.
  • Dependency Updates: Renovate is configured to automatically create pull requests for dependency updates.

Maintainer Information

Team

The primary maintainers for this repository are defined in the .github/CODEOWNERS file:

  • @gemini-cli-extensions/data-cloud-admin

Releasing

The release process is automated using release-please.

  1. Release PR: When commits with conventional commit headers (e.g., feat:, fix:) are merged into the main branch, release-please will automatically create or update a "Release PR" containing the updated changelog.
  2. Merge Release PR: A maintainer approves and merges the Release PR. This action triggers release-please to create a new GitHub tag and a corresponding GitHub Release.