Back to Work
Hackathon 3× Awards — SFxHH 2025

ShipSure

AI-powered pull request risk intelligence — analyzes GitHub PRs, auto-generates tests via CodeRabbit, runs them in isolated Daytona sandboxes, and delivers a GPT risk assessment.

ShipSure

The Problem

Development teams merge PRs daily without systematic risk assessment. Manual code review misses systemic issues across Security, Performance, and Reliability dimensions — especially under time pressure.

The SFxHH Hamburg 2025 hackathon, sponsored by Coderabbit, Daytona, and Windsurf, challenged teams to build developer tooling that automates critical workflows.

The Solution

A fully automated agentic pipeline. Provide a repo and API keys: ShipSure fetches open PRs, commissions CodeRabbit to generate unit tests, executes them in clean Daytona sandboxes, and feeds everything to GPT for a structured risk report.

Key features built in 48 hours:

  • Repository & PR Selection — GitHub API integration to fetch and analyze open pull requests
  • Automated Test Generation — CodeRabbit API creates unit tests for the PR changes
  • Isolated Test Execution — Daytona cloud sandboxes run tests in clean environments
  • AI Risk Analysis — GPT evaluates 5 risk categories: Security, Performance, Maintainability, Reliability, Compatibility

Technical Implementation

The backend is built with Python and Flask, handling async job processing with status polling via jobId endpoints. The GitHub client integrates repository fetching and PR analysis.

The frontend features a modern cyber-themed UI with real-time progress tracking — showing each stage from PR analysis to test execution to risk report generation. Built with vanilla JavaScript for minimal footprint.

Deduplication logic prevents creating duplicate CodeRabbit test PRs, and the pipeline polls CodeRabbit every 60 seconds for up to 15 minutes during test generation.

Results & Impact

Awards Won

5

Risk Categories

48hrs

Development Time

Won three awards at SFxHH Hamburg 2025 from sponsors Coderabbit, Daytona, and Windsurf. Built as part of a competitive team — I contributed to the Flask backend architecture and the frontend risk visualization UI.

Project Details

Timeline

48 hours (Hackathon)

Role

Backend & Frontend

Team

Competitive team project

Event

SFxHH Hamburg Dec 2025


Tech Stack

Python Flask GitHub API OpenAI GPT CodeRabbit Daytona