Blog
Blog
Notes from LeoTrace on what AI coding agents get wrong about security, and on building LeoPrevent. Benchmark results have their own section.
Coding agents copy your security mistakes
What two months of watching GPT-6 and Claude Opus 5.5 write code on flawed apps taught us, from interview challenges to real CVEs, and fixing a benchmark along the way.
Hello, world
What this blog is for: what coding agents get wrong about security, building LeoPrevent, and how we measure whether it helps.