Making Gemini CLI extensions easier to use

Stop debugging ambiguous startup failures caused by missing API keys, hidden environment variables, or a crashing MCP server. AI tools are most effective when they’re tailored to your specific environment and data but manual configuration is often brittle and confusing. This is why we are launching extension settings. Gemini CLI extensions can now define settings that the user will be prompted to provide upon installation, ensuring extensions have exactly what they need to function from the moment you install them.

Sorry, your browser doesn’t support playback for this video

A first-class experience for extension configuration

Extension settings provide a structured approach that ensures your experience is painless and straightforward. Extension authors can now define exactly what information their tool needs to function; whether it’s an API key, a base URL, or a project identifier. No more digging through README’s or trying to decode error messages.

Extension settings provide the following benefits:

Automated setup: Users are automatically prompted to provide settings during the installation of an extension.
Integrated security: Sensitive settings like API keys are automatically stored in the system keychain rather than plain text files.
Command-line management: The new gemini extensions config command provides a centralized way to view and update settings.
Scoped overrides: Support for both global user and project-specific workspace scopes ensures the right configuration is used in the right context.

Practical example: The AlloyDB extension

The AlloyDB extension is a perfect example of why structured settings matter. To connect with your database and interact with your data, the extension requires several pieces of information: a project ID, region, cluster ID, instance ID, and database.

In the past, you had to manually export these as environment variables. Now, the extension defines them in its manifest. When you install it, Gemini CLI guides you through the setup:

? Project ID: Your GCP project ID
? Region: us-central1
? Cluster ID: my-cluster
? Instance ID: my-primary-instance


✔ Project ID
ID of the Google Cloud project: my-google-cloud-project
✔ Location
Region of the AlloyDB cluster: us-central1
✔ Cluster ID
ID of the AlloyDB cluster: my-cluster
✔ Instance ID
ID of the AlloyDB instance: my-instance
✔ Database
Name of the database: postgres
✔ User
(Optional) Username of the database user (Default: IAM user): postgres
✔ Password
(Optional) Password of the database user (Default: IAM user): *****
✔ IP Type
(Optional) Type of the IP address: PUBLIC, PRIVATE, or PSC (Default: Public):
Extension "alloydb" installed successfully and enabled.

Shell

These values are saved securely and provided to the extension’s MCP server automatically, so you can start querying your data immediately. This setup provides a reliable and secure method for database connection. It ensures you have constant visibility into which database is active and prevents the agent from altering your configuration.

Check out these updates for all Data Cloud extensions: bigquery-data-analytics, cloud-sql-mysql, cloud-sql-postgresql, cloud-sql-sqlserver, firestore-native, looker, spanner.

How to define settings as an author

To add configuration to your extension, include a settings array in your gemini-extension.json file. Each setting specifies a user-friendly name, a description, and the environment variable it should be mapped to.

{
  "name": "my-cool-extension",
  "version": "1.0.0",
  "settings": [
    {
      "name": "API Key",
      "description": "Your secret API key for the service.",
      "envVar": "MY_API_KEY",
      "sensitive": true
    },
    {
      "name": "Endpoint URL",
      "description": "The base URL for the API service.",
      "envVar": "API_ENDPOINT"
    }
  ]
}

JSON

When a user installs this extension, Gemini CLI guides them through the setup process. If a setting is marked as sensitive, the input is obfuscated during entry and stored securely in the system keychain.

Managing settings with the config command

The gemini extensions config command is your primary tool for managing extension configuration after installation. Use it to update settings interactively or target specific values directly.

Update all settings for an extension

To walk through every setting defined by an extension, run:

gemini extensions config <extension-name>

Shell

Update a specific setting

If you know exactly which value needs to change, specify the environment variable name:

gemini extensions config <extension-name> <env-var-name>

Shell

Use workspace overrides

For settings that vary between projects, use the –scope workspace flag to save the value in your current directory’s configuration rather than globally:

gemini extensions config alloydb ALLOYDB_POSTGRES_CLUSTER --scope workspace

Shell

Debugging the configuration

When troubleshooting issues with extensions, it is often helpful to verify that all configuration settings are correct. A simple typo in a project ID or a missing API key can prevent an extension from functioning properly.

You can inspect the currently installed extensions and their active settings using the extensions list command. This functionality is available via the command line and as a built-in command. Both will output a summary of your installed extensions, their version status, and their current configuration.

Command Line

gemini extensions list

# ✓ alloydb (0.1.10)
#  Source: https://github.com/gemini-cli-extensions/alloydb (Type: github-release)
#  Release tag: 0.1.10
# ...
#  Settings:
#   Project ID: my-gcp-project
#   Location: us-central1
#   Cluster ID: my-alloy-db
#   Instance ID: my-primary
#   Database: postgres
#   User: postgres
#   Password: *****
#   IP Type: public

Shell

/extensions list

# Installed extensions:
#   alloydb (v0.1.10) - active (up to date)
#     settings:
#     - Project ID: my-gcp-project
#     - Location: us-central1
#     - Cluster ID: my-alloy-db
#     - Instance ID: my-primary
#     - Database: postgres
#     - User: postgres
#     - Password: ****
#     - IP Type: public

Shell

Best practices for extension settings

Provide clear descriptions: Help users understand exactly where to find the values they need (e.g., “Find this in your GCP Console under IAM”).
Use sensitive for secrets: Always set “sensitive”: true for API keys, passwords, or any data that should not be stored in plain text.
Keep names concise: Use short, user-friendly names that fit well within terminal prompts.
Leverage workspace scope: Use workspace-scoped settings for project- specific IDs to prevent global configuration pollution.

Get started today

Extension settings and the gemini extensions config command are available as of v0.28.0+. Update to the latest version by running:

npm install -g @google/gemini-cli@latest

Shell

Check out our updated documentation to learn more:

Try out extension settings today when building Gemini CLI extensions and if you run into any issues let us know on our GitHub repository or on socials!

You can also follow Gemini CLI on X to stay up to date with the latest news and announcements.

Source_link