Work Jira data tickets end to end with Claude Code
Watch two Claude Code sessions pick up Jira data analysis tickets, query Snowflake, transition the tickets and commit the work.
Who it's for: Data professionals who work a queue of Jira data analysis tickets.
The problem: A stakeholder submits a ticket, then a data professional reviews it, does the work and gets the output reviewed. That middle section is where the time goes.
Atlassian connects to Claude Desktop with the built-in connector, and Claude Code gets the Atlassian MCP and the Atlassian CLI. Two sessions then work two identical tickets in parallel, one with each tool set, and store output by ticket under each user's folder.
Takeaways:
- The CLI uses fewer tokens; the MCP brings more context into Claude.
- Organize work by username and ticket number so a whole team can share one repository.
- Quality check results: the sessions counted columns differently until one included the information schema.