KeyHog scans source trees, git history, Docker images, S3 buckets, and web assets for leaked credentials. It compiles 891 embedded detectors into a single Hyperscan NFA database, decodes nested encodings before matching, scores findings with ML confidence + Bayesian Beta(α,β) calibration, and routes scans to the fastest hardware backend available: