Visual TL;DR
Secret Scanning NoiseDriver
too many alerts that look like secrets but aren't
From the article 3 mentionsGitHub is leveraging Large Language Models (LLMs) to make its secret scanning more trustworthy.
GitHub Secret ScanningContext
existing pipeline combining pattern and AI detection
From the article 3 mentionsThis enhancement builds upon GitHub's existing secret scanning pipeline, which combines pattern-based and AI-based detection.
Microsoft Security & AICore
From the article 4 mentionsTo combat this, GitHub partnered with Microsoft Security & AI to integrate more contextual reasoning into the verification process.
Developer FrictionDriver
wasting time triaging false positive security alerts
From the article 4 mentionsThis friction erodes confidence in automated security systems.
LLM Contextual AnalysisCore
examining how detected values are used within the code
From the articleThe new approach focuses on the verification step, adding LLM-based contextual analysis.
Boosted Alert AccuracyEffect
slashing false positives in secret scanning
Developer EfficiencyOutcome
over 75% improvement in focusing on genuine threats
From the article 4 mentionsThis focused context, rather than more raw code, is crucial for accuracy and efficiency.