start-it

# Start It When the user says **"start it"** (or **"start"** with a Jira URL): ## 1. Get the Jira ticket - Accept either a **Jira ticket ID** (e.g. `CDF-1234`) or a **browse URL** (e.g. `https://cognitedata.atlassian.net/browse/CDF-1234`). - From a URL, use the issue key after `/browse/`. ## 2. Fetch task details - Use the **Atlassian MCP** server. In Cursor the server id is typically **`user-atlassian`** (not `atlassian`). Read the MCP tool schema if invocation fails. - **Flow:** 1. Call **`getAccessibleAtlassianResources`** on `user-atlassian` and pick the site’s **`id`** as `cloudId` (e.g. Cognite: `cognitedata.atlassian.net`). 2. Call **`getJiraIssue`** with `cloudId` and `issueIdOrKey` set to the ticket key (e.g. `CDF-1234`). - Use the response for summary, description, status, parent epic, components, subtasks, and comments. - Present a brief summary so the user can confirm it is the right ticket. ## 3. Create a working branch - Run `git fetch origin main && git switch main && git pull origin main` to start from the latest main. - Create and switch to a new branch: `git switch -c <branch-name>`. - Derive the branch name from the ticket: lowercase ticket ID + short slug from the summary. - Example: `CDF-1234` with summary "Add data product validation" → `cdf-1234/add-data-product-validation`. - Confirm the branch name with the user before creating it. ## 4. Check if the ticket involves a new resource type - If the Jira ticket is about adding or supporting a **new CDF resource type**, ask the user to provide either: - A **service contract YAML** file, or - A link to the **API documentation page** - Do not proceed to implementation until this information is available, as it is needed to correctly model the resource. ## 5. Plan and start working - Based on the Jira ticket details, create a **todo list** breaking the task into concrete implementation steps. - Explore the codebase to understand relevant files and patterns before making changes. - Begin working on the first todo item immediately.