Our team recently prototyped a project we’ve been calling the Open Source Swagger Feedback Triage Assistant. It was born from a simple reality many open-source maintainers face: GitHub issues never stop coming in, and keeping up with them can quickly turn into a full-time job.
The Problem
Swagger’s open-source repositories receive a steady flow of community feedback. New issues, comments, feature ideas, and bug reports. While that engagement is great, manually triaging all of it isn’t. Developers have to constantly check GitHub, read through threads, and decide whether each post is a bug, a feature request, or just a question.
On top of that, someone has to determine if the issue should become a Jira ticket, be handled immediately, or simply acknowledged. This back-and-forth eats up valuable engineering time, slows responses to the community, and risks losing great ideas or important bug reports in the noise.
The Idea
To solve this, we built a Slack bot prototype that connects GitHub, Slack, and SmartBear tools through MCP to automate the triage process, or at least make it a whole lot easier.
Whenever a new issue or comment appears on Swagger’s open-source repositories, the bot posts it into a dedicated Slack channel. From there, AI steps in: it reads the text, classifies the feedback (Bug, Feature Request, or Question), and suggests a possible priority or next action.
The team can then confirm or edit the classification directly in Slack. Once confirmed, the bot can automatically create a Jira ticket with all the relevant details or comment back on GitHub to acknowledge that the issue is being looked into.
It’s simple, transparent, and keeps the entire triage process where teams already collaborate in Slack.
The Prototype
Since this was a hackathon project, our focus was on building a working MVP that proved the concept. The prototype included:
- GitHub → Slack notifications for new issues and comments
- AI classification directly in Slack, suggesting the issue type
- Quick actions to confirm, edit, or ask for more details
- Optional follow-up comments back to GitHub once confirmed
As a stretch goal, we experimented with automatically creating Jira tickets from Slack, pre-filling metadata like issue type, priority, and description.
While still early, the prototype successfully demonstrated that AI can handle a large portion of repetitive triage work — freeing maintainers to focus on solving problems instead of sorting them.
Why It Matters
This kind of automation has a huge potential impact. It saves developer time, reduces delays in responding to contributors, and ensures valuable community insights aren’t missed. It also helps maintainers keep triage consistent and fair. The AI doesn’t forget, overlook, or get tired of reading long threads.
Ultimately, it supports the open-source ethos: faster feedback, clearer communication, and better collaboration between the community and the maintainers who make Swagger thrive.
What’s Next
For now, Open Source Swagger Feedback Triage Assistant remains a functional prototype, but the potential is exciting. The next steps could include deeper integration with Jira workflows, fine-tuning AI accuracy based on real feedback, and expanding to other open-source projects within SmartBear’s ecosystem.
We’d love to hear your thoughts. How do you handle open-source issue triage today? What would make this kind of assistant genuinely useful in your workflow?