🏎️ AI Race Strategist — Atlassian Codegeist 2025

"What if your DevOps infrastructure had the reflexes of an F1 Pit Crew?"

AI Race Strategist is an autonomous DevOps agent for Formula 1 teams, built on Atlassian Forge. It acts as a "Ghost Monitor" inside Jira, bridging the gap between physical hardware telemetry (IoT sensors) and software operations (Jira/Bitbucket/Confluence).

---

🔗 Quick Links

📦 Installation

Get started by installing the app directly from Atlassian:

• 🚀 Install on Atlassian

• 🎬 Demo Video

• 📄 Devpost Submission

---

🧪 Judge's Testing Guide (The "Happy Path")

Follow these steps to experience the full autonomous workflow:

Phase 1: The "Crash Test" (Simulation)

1. Open any Jira Ticket (e.g., KAN-1 or create a new one).
2. Open the "🏁 Pit Wall (real)" Issue Panel.
3. Navigate to the "Sim & Telemetry" tab.
4. Click the "START CRASH TEST" button.
   • Observation: Watch the 3D Digital Twin (Cube) turn RED and shake.
   • Observation: Watch the Graph spike to a critical anomaly.
5. Check the Jira Comments:
   • You will see an automated "🚨 FAILURE DETECTED" comment posted by the system backend.

Phase 2: The Rovo Agent (AI Analysis)

You can test the Agent in two ways:

Option A: The Rovo Simulator (Inside the App)

• Go to the "Rovo Agent (Beta)" tab in the app.
• Type: "Analyze KAN-1" (or your current ticket key).
• Result: The simulated agent will perform a Root Cause Analysis and provide a Bitbucket Fix Link.

Option B: The Native Rovo Agent (If enabled on your instance)

• Open the Rovo sidebar in Jira.
• Select the "Race Strategist" Agent.
• Prompt: "Report Fleet Status" or "Analyze KAN-1".

Phase 3: The Autonomous Fix

1. In the Rovo Chat (or Simulator), click the Bitbucket Link provided.
2. It will take you to a real Pull Request created by the AI (e.g., fix/compliance-KAN-1), showing the code change from 5.0mm to 8.0mm.
3. Check Confluence (if linked) for the auto-generated "Post-Mortem Report."

---

📂 Codebase Navigation

Here is a map of the file structure to help you review the logic:

1. The Backend (src/)

• src/index.js: (CORE LOGIC) This single file contains the entire backend intelligence.
  • Lines 10-50: Bridge Resolvers (Handling the "Crash Test" trigger from frontend).
  • Lines 100-180: prescribeSolutionHandler (The AI Logic: Physics Calc + Compliance Check + Bitbucket PR generation).
  • Lines 250+: Rovo Agent Resolvers (rovoGetStatus, rovoGetIncident).

2. The Frontend (static/)

• static/pit-wall/: The main Jira Issue Panel.
  • script.js: Contains the Three.js logic for the 3D Cube and the TensorFlow.js inference engine.
  • index.html: The UI layout.

3. Configuration

• manifest.yml: The critical configuration file.
  • Defines the rovo:agent ("Race Strategist") and its System Prompt.
  • Defines the action modules (get-live-status) that connect the AI to the backend.
  • Defines the custom OAuth2 Provider (bitbucket-oauth-v2) used for the autonomous PR integration.

---

⚙️ Technology Stack

• Platform: Atlassian Forge (Node.js Runtime)
• AI & Agents: Atlassian Rovo, TensorFlow.js (Local Inference)
• Visualization: Three.js (WebGL Digital Twin), Chart.js
• Integrations: Jira Software, Bitbucket Cloud API, Confluence Cloud API

---

⚠️ Important Notes for Reviewers

1. Bitbucket OAuth2 Integration

The app uses a custom OAuth2 provider to perform actions on Bitbucket (creating branches/PRs). If you fork this repo, you must update the clientId in manifest.yml with your own Bitbucket Consumer credentials.

2. Rovo Simulator vs. Native Agent

Due to access limitations on my development instance during the hackathon, I built a "Rovo Simulator" tab within the app to visually demonstrate the agent's logic for the video. However, the Native Rovo Agent is fully configured in manifest.yml and is functional if installed on a Rovo-enabled instance.

---

Built with ❤️ for the Atlassian Codegeist 2025 Hackathon.