eBPF CVE Analysis
A reproducible workflow for finding and classifying eBPF vulnerabilities across the CVE ecosystem.
The problem
eBPF is a powerful Linux kernel capability, but finding the vulnerabilities that genuinely concern it can be difficult in a broad, inconsistently labelled CVE corpus.
Contribution
I built a Python pipeline to identify and classify eBPF-related vulnerabilities across more than 253,000 CVE records. The work combines structured filtering with hands-on proof-of-concept analysis, turning a large public dataset into a tractable security-research workflow.
Why it matters
The project makes the relationship between vulnerability records and a specific kernel attack surface easier to inspect, reproduce, and extend.