The built in Jira connector in Microsoft 365 Copilot is read only. You can search and view issues but you cannot create, update, or take action on them. One of our customers raised this as a limitation, so we built a solution.

This is a custom Copilot Cowork plugin that connects to Jira Cloud with full CRUD operations. Users can search issues, create new ones, update fields, transition through workflow states, add comments, log time, and generate sprint reports. All from natural language prompts inside Copilot Cowork.

The plugin includes three skills. The first is issue management, which covers searching, creating, updating, transitioning, and commenting on issues. The second is sprint reporting, which generates progress summaries, team workload breakdowns, and remaining work analysis. The third is time tracking, which lets users log hours and check time spent on any issue.

What makes this interesting is what happens when you combine the Jira plugin with Cowork's built in skills. You can ask Cowork how many tasks someone has in a project and then automatically send them an email with action items. You can upload a project specification document and ask Cowork to create all the Jira issues based on that spec. You can generate a visual dashboard with pie charts and bar charts showing status breakdown, priority distribution, and workload by assignee.

The plugin uses the open source mcp atlassian MCP server to bridge Cowork and Jira, and the Agent Skills open standard to define the skill instructions. This means the same skill files also work in Claude Code, GitHub Copilot, Gemini CLI, Cursor, JetBrains Junie, and 30+ other AI tools.

David Lorenzo and I recorded a full walkthrough video covering the architecture, the code, and a live demo showing every capability in action.

The entire plugin is open source and available on GitHub: https://github.com/rafsan-huseynov/copilot-cowork-jira-plugin