Implementing features with Antigravity CLI and JIRA¶
This guide demonstrates how the Antigravity CLI accelerates feature development across the Software Development Life Cycle (SDLC) by integrating with external third-party task and requirements management systems using the Model Context Protocol (MCP). By pulling user stories from JIRA and technical specifications from Confluence directly into the Antigravity context, developers can understand unfamiliar codebases, plan architectural modifications, generate verified application code, such as adding ratings and validations to a Java microservice, run unit tests, and transition JIRA ticket statuses without leaving the development environment.
Requirements¶
To follow this guide, you need:
- A Google Cloud project with the
Ownerrole. - A JIRA and Confluence project.
- Antigravity CLI: Installed and configured. For installation instructions, visit antigravity.google.
- Java Development Kit (JDK): Version 21 or higher
Atlassian projects setup (JIRA and Confluence)¶
-
For this guide, you need to have JIRA and Confluence projects configured. Sign up.
-
Create a new Confluence page:
-
Create a new JIRA user story and assign it to yourself:
Title: Update Menu service Update Menu service: 1. Add new fields: description and rating to Menu entity. 2. Update other dependencies where Menu entity is used in the code, eg MenuResource. 3. Add unit tests for all methods, including new fields. Link to Confluence page: https://YOUR-ORG.atlassian.net/wiki/spaces/SPACE-KEY/pages/NNNN/Menu-Service+Rating+capabilities
Clone Git repository and navigate to workspace¶
-
Open your local terminal, clone the demo repository, and navigate to the menu-service directory:
git clone --filter=blob:none --no-checkout https://github.com/GoogleCloudPlatform/cloud-solutions cd cloud-solutions git sparse-checkout set --cone projects/build-with-gemini-demo/gemini-powered-development/menu-service git checkout cd projects/build-with-gemini-demo/gemini-powered-development/menu-service
Configure Antigravity MCP server in local environment¶
-
The Antigravity CLI discovers workspace-scoped MCP servers located under the
.agents/directory. Create the.agentsfolder and writemcp_config.jsondefining the Atlassian MCP server:
Authentication with service account API key (headless environments)¶
Follow these steps when configuring the MCP server in a CI/CD pipeline.
-
Create a new service account for your JIRA instance in the Atlassian admin console, under
Directory / Service accounts. -
Create a credential for the service account using the
API tokenauthentication type. -
Enable
Allow API token authenticationunderRovo / Rovo MCP server. -
Configure MCP server:
You will use
Service account API key (Bearer token)to configure the Atlassian MCP server. Additional details.
Prepare the environment for Atlassian integration¶
-
Update JIRA/Confluence instance details in
AGENTS.md:## Atlassian MCP When connected to atlassian-mcp server: - **MUST** use Jira project key = YOURPROJ - **MUST** use Confluence spaceId = "123456" - **MUST** use cloudId = "https://yoursite.atlassian.net" (do NOT call getAccessibleAtlassianResources) - **MUST** use `maxResults: 10` or `limit: 10` for ALL Jira JQL and Confluence CQL search operations.
Check MCP server configuration¶
-
Run Antigravity CLI:
-
List configured MCP servers and tools:
The output confirms that the
atlassianserver is configured and ready. On the first run, you need to authenticate, select your JIRA instance and grant permissions.Press Esc or exit the overlay to return to the prompt panel.
Codebase exploration and onboarding¶
-
Prompt Antigravity to analyze the existing codebase and produce an onboarding guide:
Role & Context: Assume the role of Technical Lead to onboard a developer joining this codebase. Objective: Evaluate the project repository and generate a comprehensive Onboarding Guide containing: - High-Level Architecture: Details on the technical stack and component interactions. - Core Functionality: Overview of the 3 primary features delivered by the system. - Directory Layout: Concise descriptions outlining the intent of major folders. - Request Pipeline: Step-by-step lifecycle of an incoming request from entry point through database persistence and return.
(Optional) If you did not configure JIRA and Confluence projects¶
-
Send a prompt with the task requirements:
/plan Review the code and prepare the implementation plan for the requirements below. I will approve the plan before you can start implementation. Update Menu service: 1. Add new fields: description and rating to Menu entity. 2. Update other dependencies where Menu entity is used in the code, eg MenuResource. 3. Add unit tests for all methods, including new fields. Rating must be an integer value from 1 to 5. Can't be null. Can't be empty. Can't be zero.
Bring requirements into the context of an Antigravity CLI session¶
-
Send a prompt to list assigned JIRA tasks or get the context for a specific issue:
-
Replace the JIRA issue details and send the prompt below:
-
Send a prompt to query the context of the linked Confluence page:
Plan and implement feature changes¶
-
Send a prompt to create an implementation plan:
/plan Review the code and prepare the implementation plan for this user story. I will approve it before you can start implementation.Start implementation by approving the plan using the
/artifactcommand.Review and approve tools and suggested code changes. If the Antigravity CLI runs into issues, for example, test validation, multiple iterations might be required to fix and rerun until the generated code is valid.
-
Send a prompt to update the JIRA user story:
-
Exit the Antigravity CLI before moving to the next section.