Posted in Threat Intelligence

GitHub - jp101216/LiveResponse-NG: Live-response triage agent for Windows and Linux DFIR: ~120 ATT&CK-mapped…

jp101216 — Security Engineer | Offensive Security Research & Custom C2 Tooling

LiveResponse-NG v0.16.0 is live: Multi-Report Case Workspaces, WFP EDR-Silencing Detection, ClickFix Analytics, and Offline Cloud/LLM Abuse Triage

The latest releases of LiveResponse-NG (lrng) push the tool beyond single-host snapshots into multi-report case correlation, interactive evidence graphing, and modern initial-access and defense-evasion detections—all while keeping zero new Go dependencies and preserving report schema 1.2.

What’s New in v0.16.0

• Multi-Report Case Workspaces & Persistence Diffing (-case-report, -compare -same-host): Group multiple reports into an offline case workspace with source-file digests, preserved host/run boundaries, and cross-report finding search. Use -compare -same-host for operator-attested, chronological persistence comparisons across services, scheduled tasks, autoruns, and drivers without false assertions on one-sided or partial observations.

• Windows Filtering Platform (wfp) Enumeration: Read-only native enumeration of enabled app-specific BLOCK filters, provider/condition metadata, and findings for filters targeting security agents (EDR silencing).

• ClickFix Lure Analytics & runhistory Correlation: Account-scoped Run-dialog command correlation that reads loaded user hives and distinguishes registry key timestamps from actual execution times.

• DLL Sideloading & IDE Tunnel Detection: Identifies signed-host/adjacent-DLL sideloading candidates with bounded signature checks (honoring -no-sign) and suspicious-child correlation gated on exact process identity. Also detects abused IDE tunnels using verified process ancestry and bounded original-filename inspection for renamed Windows binaries.

• Offline Cloud & AI/LLM Abuse Heuristics (-cloud-audit, -llm-burst, -agent-trace): Evaluates AWS, GCP, and Azure audit metadata (or normalized JSONL) completely offline for model-resource abuse and credential/permission changes without collecting prompts, responses, or raw secret values. Pairs with v0.15.0’s -agent-trace for offline OTLP/JSON AI agent telemetry ingestion and span-lineage tracking.

• Interactive Evidence Graph & Adaptive Triage (-graph, -adaptive): Build a bounded evidence graph with explicit observed vs. inferred relationships and neighbor navigation in the offline HTML viewer, and use -adaptive for a deterministic, budget-capped follow-up pass on executable paths tied to suspect or critical live records.

• Hardened Performance & Testing: Validated against hostile-case Chromium viewer checks, a repeatable 10,000-service comparison benchmark, and standalone Linux live-process test executables that avoid multicall coreutils exits.

Repo & Changelog: https://github.com/jp101216/LiveResponse-NG

GitHub - jp101216/LiveResponse-NG: Live-response triage agent for Windows and Linux DFIR: ~120 ATT&CK-mapped…

3 reactions · 2 comments

Comments

Victor Font: Interesting development. I particularly appreciate the distinction between observed and inferred relationships in the evidence graph, along with source-file integrity checks and offline analysis. From a governance and incident response perspective, preserving evidence provenance and investigative context is essential. One question: does LiveResponse-NG also support documenting analyst decisions, e

jp101216: Victor Font Thanks, Victor! The current release supports documenting analyst decisions through per-finding review statuses, notes, an analyst summary, and an exportable handoff. That handoff includes assessments, source references, available hashes, and collection gaps, while preserving the original evidence unchanged. Formal chain-of-custody tracking and case-disposition approvals still need to b

Read and join the discussion on Cyberr