Skip to main content

Jira Skill and Dataset

The jira-project-tracking skill creates and maintains a realistic Scrum and Kanban dataset while remaining idempotent.

Workflow​

  1. Resolve the Jira cloud ID with getAccessibleAtlassianResources.
  2. Discover current read, create, edit, and transition schemas.
  3. Search for the cowork-rovo-repro label before writes.
  4. Match records by stable seedId and exact summary.
  5. Create parent records before children.
  6. Request confirmation before bulk creation or transitions.
  7. Return created, updated, and unchanged records.

Canonical dataset​

The dataset contains 20 records:

ProjectModelRecordsExpected distribution
SWEScrum103 To Do, 3 In Progress, 4 Done
KANKanban107 To Do, 2 In Progress, 1 Done

SWE-1 is an epic parent for the remaining SWE work. KAN items are independent flow records. Every seeded issue uses cowork-rovo-repro and demo-data labels.

Safety model​

  • Never delete as part of synchronization.
  • Never modify unrelated issues or project settings.
  • Read current transitions before changing status.
  • Prefer unchanged results over duplicate creation.