Take action in Swarmia from the MCP

We recently introduced actions in Swarmia AI, making it easier to turn insights into concrete next steps. Now, you can take those same actions from the MCP, too.

Swarmia’s MCP server already lets you explore and analyze your engineering data using your preferred AI tools. With this update, you can also make changes in Swarmia directly from those conversations.

Say you’re investigating a spike in cycle time. Instead of just finding out what happened, you can ask your agent to add an annotation to capture the explanation. Or, if you uncover work spread across multiple teams, you can ask it to create an initiative to track it together.

Here are a few things you can try:

• Create an initiative to bring related work together, such as all the cross-functional work happening around a particular topic.
• Set up a working agreement for unlinked PRs or other team practices. Notifications are included in working agreements, so you don’t need to configure them separately.
• Add an annotation to a pull request, issue, or signal to capture important context. AI-written annotations are clearly marked, and you can accept, deny, edit, or replace them.
• Save a key metric from your analysis under your organization-wide goals to keep important measures visible.
• Create an explore view to save a custom report from your analysis, so you can come back to the same data later.

You’re always in control of what changes. Before taking action, the MCP asks for any missing details and shows you a preview of what it’s about to do. Nothing changes until you confirm, and all actions follow your existing Swarmia permissions.

Once the action is complete, you’ll see what changed and which objects were affected. Where possible, you can also inspect the result.

We’re starting with actions that are easy to review and understand. Broad settings changes and actions that rewrite historical data aren’t supported yet, but more actions are coming soon.