Finding Bugs Using LLMs
Materialize has successfully utilized LLM-based coding agents to identify bugs in code and pull requests since February 2026. The system employs a variety of strategies, including reviewing pull requests and commits, to enhance bug detection. The approach aims to minimize false positives and prioritize significant bugs while leveraging advanced models like Anthropic's Opus 4.7.
- ▪Materialize began using LLM-based coding agents for bug detection in February 2026.
- ▪The system reviews pull requests, commits, and production source code files to find bugs.
- ▪Bugs are categorized by severity, with a focus on high and medium severity issues.
Opening excerpt (first ~120 words) tap to expand
Blog Finding Bugs using LLMs Finding Bugs using LLMs May 22, 2026 Dennis Felsing Software Engineer in Test Table of Contents SessionsPromptTools & SkillsModelsStaying HonestConclusion Table of Contents At Materialize we’ve had success in finding bugs in existing code and open pull requests using LLM-based coding agents since February 2026, coinciding with the release of Anthropic’s Opus 4.6 (now mostly running on 4.7). In this post we’ll look into some of the considerations that went into the system we are currently using as well as lessons learned.SessionsWe have a basic shell script that determines the next unit to operate on and feeds it to claude .
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at Materialize.
Discussion
0 commentsMore from Materialize
