Code Project: CVE Hunter with Cursor AI

A Practical Side Project for Real-World Cyber Threat Intelligence

In the age of AI-assisted development, I challenged myself to build something that not only showcases technical skills but also solves a real problem I face everyday. That’s how CVE Hunter was born.

As a cybersecurity professional, I frequently look up CVEs (Common Vulnerabilities and Exposures). But here’s the issue: the information I need is fragmented across the internet. One website has CVSS scores but no exploit data, another has exploit references but lacks EPSS probabilities, and CVEDetails—once a go-to source—now hides most useful features behind a paywall. It was frustrating.

So I decided to build a unified CVE intelligence dashboard that aggregates all the relevant details I care about—faster, cleaner, and AI-powered.

Step 1: Setting Up the CVE Data Pipeline (NVD API)

I started with the National Vulnerability Database (NVD), which offers a public API for querying CVEs. To make queries faster and more flexible, I wrote a Python script that:

• Pulls CVE data via the NVD API
• Saves all CVEs in a local database (with full JSON blob for flexibility)

• Auto-syncs every 3 hours to fetch new or modified CVEs (NVD’s limit is 1 request every 2 hours)

Check out NVD API Best Practices for Developers

This created a robust and scalable backend to power the rest of the application.

Step 2: Building the CVE Hunter Interface

With the data foundation in place, I started designing the frontend with a clean, focused UI. The core interface includes:

• CVE Search Bar (by ID or keyword)
• CVE Details (summary, publish date, vendor info)
• Mitigation Tips (if available)
• EPSS Scores (for real-world exploit probability)
• CVSS Scores (base, temporal, and environmental)
• CWE Data (classification of the vulnerability)
• Known Exploits (links to Proof of Concept code, malware usage, etc.)
• Risk Matrix (custom scoring based on impact and likelihood)
• References (blog posts, vendor advisories, security research)
• AI-Powered Analysis (using OpenAI API to summarize the threat, suggest mitigations, and even generate detection ideas)

The result? A single-pane-of-glass for CVE analysis.

Step 3: Design & Extras

I themed the app with a Cyberpunk aesthetic, to keep the vibe fresh and developer-friendly. To make the platform more robust, I added Shodan CVE data, which brings in exploit context—whether the CVE is being actively exploited, associated with ransomware, or exposed in the wild.

What the Site Shows Today

• 🔢 Total CVEs Indexed and Beautiful Search Box Design

• 🔍 Detailed CVE View, CWE Classification, CVSS Scores, EPSS Scores

• ⏳ Beautiful Loading Screen

• 🧮 Custom Risk Matrix (Impact x Likelihood) and References

• ⚠️ Threat Intel and Threat Context

• 🔓 Known Exploits and AI Analysis

• 🌐 Shodan CVE Database and Intel

• 🕵️ Search History & Stylish Footer

Honestly? I’m pretty happy with how it turned out.

Roadmap: What’s Next

This is just the beginning. Here’s what I plan to add next:

• 🗺️ Shodan Search Integration with Interactive Map

  Visualize IPs, locations, and open ports tied to CVEs in real time.

• 🔗 More Threat Intel Sources

  Integrate APIs from vendors like GreyNoise, AlienVault OTX, and Recorded Future for richer context.

• 🔍 Full-text CVE Search & Filters

  Severity, vendor, exploitability score, affected product—you name it.

• 🔐 Red Team/Blue Team Tabs

  Different views for defensive mitigations vs. offensive testing strategies.

---

CVE Hunter started as a side project, but it quickly became something I now use everyday. It not only saves time, but it also improves the depth of my CVE analysis by combining structured data with AI insights.

It’s a reminder that even small personal projects—when built with intention—can have a big impact.