🏎️ PitCrew AI for Atlassian

Formula 1-inspired Engineering Telemetry for Jira & Bitbucket

PitCrew AI transforms Jira and Bitbucket into an F1-style telemetry system for engineering teams. It analyzes pull requests in real time, detects risks, recommends reviewers, flags issues, and generates weekly race-style reports — all powered by Atlassian Forge, Rovo agents, and a fully custom React dashboard.

Designed for Codegeist 2025: Williams Racing Edition.

---

🚀 What is PitCrew AI?

PitCrew AI acts like a Formula 1 pit crew for your development workflow:

• Measures the “risk level” of each PR using a lightweight scoring model
• Detects critical files, missing tests, and sprint-end danger zones
• Calls Rovo to summarize PRs and highlight what reviewers should check
• Flags high-risk or blocked PRs directly in Jira & Bitbucket
• Generates a weekly “Race Report” that summarizes team performance
• Displays all engineering telemetry inside a Williams-themed dashboard

This provides technical clarity and keeps teams flowing like a perfectly timed pit stop.

---

🧠 Tech Stack

Frontend (Dashboard)

• React
• TypeScript
• Vite
• Tailwind CSS v4 (Williams Racing Theme)
• Atlaskit

Backend / Integrations

• Atlassian Forge (Custom UI)
• Forge Functions, Resolvers, Scheduled Triggers
• Rovo Agents
• Bitbucket & Jira Webhooks

---

🏗️ Repository Structure

pitcrew-ai-atlassian/
  apps/
    forge-app/      # Backend Forge app (webhooks, resolvers, Rovo integration)
    dashboard/      # React dashboard (Williams F1 style, Tailwind 4)
  docs/             # Documentation, diagrams, planning
  package.json      # NPM Workspaces config
  README.md

---

🔧 Setup & Installation

1. Clone the repository

git clone <repo-url>
cd pitcrew-ai-atlassian
npm install

2. Configure the Forge app

cd apps/forge-app
forge login
forge deploy
forge install

3. Run the dashboard (local development)

To iterate on the UI with hot-reload (standalone mode):

npm run dev:dashboard
# Dashboard available at http://localhost:5173

4. Develop with Forge Tunnel (Integration)

To see the app inside Jira with live backend changes:

npm run dev:forge

5. Build & Deploy Production

To build the dashboard and deploy the full Forge app:

npm run build:forge

This command runs the dashboard build (outputting to apps/forge-app/static/dashboard) and then runs forge deploy.

---

🎨 UI Theme: Williams F1 Inspired

• Dark navy background (bg-williamsBlueDark)
• Cyan & white accents
• Gauge-style UI components
• Flag metaphors (green/yellow/red)

---

✅ What is Implemented

PitCrew AI currently delivers a production-grade PR risk analysis system:

• Real-time PR Analysis – Webhooks trigger instant analysis on every PR update (create, update, merge, decline)
• Intelligent Risk Scoring – Sophisticated algorithm (0-100 score) with 3-tier classification (🟢 Green, 🟡 Yellow, 🔴 Red) based on files changed, lines modified, test coverage, reviewers, and timing
• Smart Gating – Skip redundant analyses when commits haven't changed, with intelligent caching (5min TTL)
• Automated Bitbucket Comments – In-place comment updates with risk breakdown, metrics, and actionable factors
• Williams F1 Dashboard – Full React telemetry dashboard (7 views, 26+ components, Tailwind v4) with KPIs, charts, and PR timeline
• Robust Infrastructure – Retry logic with exponential backoff, structured JSON logging, Zod validation, and comprehensive error handling
• Security-First Design – Least-privilege scopes, no PII storage, GDPR-compliant, full threat model documented

---

🚧 What is Planned

Next phase for Codegeist 2025:

• Rovo Agents – AI-powered PR summaries and smart reviewer recommendations
• Weekly Race Report – Automated F1-style team performance reports with podiums, trends, and insights
• Jira Deep Integration – Auto-flag high-risk PRs as Jira issues, custom fields, sprint board metrics
• ML Analytics – Predictive models for review time, bottleneck detection, and quality forecasting

---

📜 License

Apache License 2.0