Behavioral LOLBins detection e... Note

Behavioral LOLBins detection engine with sequential correlation, PE masquerading checks, and Sysmon integration

The author developed a behavioral detection engine specifically targeting Living-off-the-Land (LOLBin) binaries on Windows and is seeking community feedback. Traditional detection methods struggle with LOLBins as they are legitimate, signed executables, often renamed or used with obfuscated command lines. This tool addresses these challenges by collecting process telemetry, ideally from Sysmon, and applying single-event behavioral rules for common LOLBins like certutil and powershell. It also performs sequential multi-stage correlation to identify broader attack patterns, such as a download followed by execution. The engine includes features for detecting binary masquerading, PPID spoofing anomalies, and command-line obfuscation heuristics. To facilitate analysis, it suppresses false positives, de-duplicates events, and reconstructs timelines with process lineage. A local web dashboard aids triage, and automated Markdown incident reports can be sent to various platforms. The modular architecture allows rules to be written in YAML or Python, and a built-in threat emulation harness tests 11 scenarios. Current limitations include polling latency in psutil mode, user-mode-only operation, and a need for more real-world tuning of suppression rules. The author is particularly interested in feedback on rule coverage, false positive rates, missing sequential chains, and the practical usefulness of the dashboard and timeline reconstruction.