Research Blog

Weekly insights from our cybersecurity research journey

LLM-Powered Fuzzing: Our 23-Strategy Arsenal

A practical breakdown of 10 discovery and 13 patching strategies, orchestrated across multiple LLMs with static and dynamic analysis.

🔎 10 Discovery 🩹 13 Patching 🎭 Multi-Model 📊 Validation Gates

Read Full Post →

September 22, 2025 Live Demo

Watch Claude Code autonomously find a libpng buffer overflow, build a PoC exploit, and ship a clean patch.

Published

September 15, 2025 Research Tools

Our evaluation platform that ranks state-of-the-art models on vulnerability detection and patching.

Published

September 8, 2025 Competition

Highlights from our finalist performance: 28 vulns found, 14 patches, 6 zero-days.

Published

October 6, 2025 Architecture

How we coordinate four services across ~100 VMs to run thousands of concurrent jobs and 100K+ LLM requests.

Coming Next Week