Inspiration## 🏎️ Inspiration: Formula 1 for Software Teams
In Formula 1, the Halo system protects drivers from fatal crashes. In software development, sprints "crash" all the time—due to hidden dependencies, scope creep, and unplanned "drag."
We realized that modern Jira workflows are reactive. Managers only see a delay after it happens. We wanted to build a predictive safety system—a "Safety Car" for your sprint that flags risks before they become blockers.
Halo Guardian was born from this question: What if your Jira ticket could tell you it was going to fail?
🛡️ What it does
Halo Guardian is an embedded AI Risk Engine that lives directly inside your Jira tickets. It acts as an always-on compliance officer and risk analyst.
- Real-time Drag Detection: It analyzes ticket comments and status changes to calculate a "Risk Score."
- Predictive Intervention: Instead of just tracking tasks, Halo proactively warns the user: "⚠️ High drag detected. This ticket conflicts with the Backend Migration."
- One-Click Fixes: It doesn't just complain; it offers solutions. Users can "Deploy Fixes" (like auto-extending timelines or flagging QA leads) directly from the chat interface.
⚙️ How we built it
We built Halo Guardian using the Atlassian Forge platform to ensure seamless security and integration within the Jira ecosystem.
- Frontend: We used the Forge UI Kit (React-based) to build a native-feeling "Chat Interface" that sits inside the Issue Panel.
- Backend: The logic is powered by Node.js functions running on Atlassian's FaaS (Function-as-a-Service) infrastructure.
- Risk Logic: We implemented a custom weighting algorithm that scans keywords (e.g., "blocked," "waiting," "fail") to trigger dynamic UI states (Red for Danger, Green for Resolved).
📉 The Math Behind the Risk
To make our risk detection scientific, we modeled "Sprint Drag" using the following logic:
$$RiskScore = \frac{\sum (DependencyWeight \times Complexity)}{HistoricalVelocity}$$
When the $$RiskScore$$ exceeds a threshold of 0.8, Halo Guardian automatically deploys the "Yellow Flag" warning to the team.
🧠 Challenges we ran into
The biggest challenge was the "Rovo" Agent integration. We initially aimed to use the beta Rovo agent, but we realized that for a "Safety System," we needed a UI that was always visible and not hidden behind a sidebar.
We pivoted to building a Custom Issue Panel. This required complex state management in React to simulate a "thinking" AI that provides instant feedback without reloading the page.
Built With
- atlassian-forge
- javascript
- jira
- node.js
- react
Log in or sign up for Devpost to join the conversation.