Microsoft · GitHub · GitHub Blog
Each candidate then goes to a fine-tuned ModernBERT classifier, which reads the value together with its file path
Compiled by KHAO Editorial — aggregated from 1 source. See llms.txt for citation guidance.
★ Tier-1 Source
Repository content, commits, pull requests, and other scanned surfaces.
Key facts
- The team developed a fast classifier that reduced median post-push scan execution time by 86% compared with the previous LLM-based pipeline
- The trained model is exported to ONNX format with 8-bit quantization, enabling fast inference
- As a Partner Applied Science Manager in Microsoft CoreAI, the reporter lead an applied research team that measurably boosts developer productivity with ML and AI
- As a Senior Research Scientist in Microsoft CoreAI, the reporter develop agentic model orchestration and routing techniques for GitHub Copilot
Summary
Compact model vs. remote LLM calls. Secret scanning works best when developers are confident in the findings. To address this challenge, they trained a secret classifier specifically for GitHub’s product needs: distinguishing real credentials from lookalikes while meeting the speed and efficiency requirements of scanning at scale. In their companion post, Secret protection must scale with software, they explained why secret protection must keep pace with AI-accelerated development. Push protection helps prevent secrets from entering a repository by scanning changes before a push is accepted, giving developers an opportunity to address the finding.