Automating Malware Reverse Engineering with Local LLMs, PyGhidra and Neo4j Graphs

A researcher has developed an automated pipeline that uses local large language models to analyze decompiled malware code extracted via PyGhidra. The system loads functions, call graphs, strings and imports into a Neo4j graph database to preserve context across hundreds of functions. Each function is sent to a local Qwen3 model running in LM Studio together with its neighboring call-graph context, producing structured JSON output on purpose, IOCs, tags and evasion techniques. Aggregated capabilities and behavioral patterns such as file encryption and C2 communication are then derived through graph queries. Testing on a WannaCry sample from MalwareBazaar processed 195 functions in 126 minutes and correctly identified File Encryption, C2 Communication and Anti-Analysis behaviors with 100 percent confidence. The approach keeps all sensitive indicators inside a local environment and avoids context overflow and censorship issues common with cloud-based models.

HabrJuly 27, 2026Translated from Russian

Read at source

A researcher has built an automated system that combines PyGhidra, Neo4j and a local large language model to speed up the initial analysis of malicious binaries. The project addresses the main drawbacks of sending decompiled code to cloud LLMs: leakage of indicators of compromise, possible censorship of sensitive modules, and rapid context overflow when a sample contains 100–200 functions.

Architecture and data flow

The pipeline begins with ghidra_extractor.py, which uses PyGhidra to disassemble and decompile a binary without launching the Ghidra GUI. For every function the extractor records its address, name, size, cyclomatic complexity, decompiled C pseudocode (truncated at 6000 characters with a 30-second timeout), call-graph neighbors, string references and imported API calls. Metadata such as PE format, architecture and all standard file hashes are also stored. All extracted entities are loaded into Neo4j as distinct node types: Sample, Function, String, Import, Analysis, Capability and Behavior.

Next, worker.py and mcp.py iterate over functions and send each one to a locally hosted Qwen3 model running inside LM Studio. The prompt includes both the decompiled code and the function’s position in the call graph, allowing the model to understand caller-callee relationships. The model returns structured JSON describing the function’s purpose, extracted IOCs, tags, network indicators, command-line usage and anti-analysis techniques.

aggregator.py then aggregates tags across all functions into Capability nodes that carry a strength weight reflecting how often each capability appears. builder.py combines these capabilities with the call graph to detect higher-level behavioral patterns such as file encryption combined with network activity. Finally, generator.py produces a human-readable report by querying the enriched graph.

Graph data model

Storing functions in a flat table makes it difficult to answer questions such as whether five anti-debug functions call one another. In the Neo4j model a Function node is connected to other Function nodes via CALLS edges, to String nodes via REFERENCES edges, and to Import nodes via USES edges. Capability and Behavior nodes sit above the function level and are linked through weighted relationships, enabling concise Cypher queries instead of recursive joins.

Validation on WannaCry

The pipeline was tested on a WannaCry sample obtained from MalwareBazaar. Processing 195 functions took 126 minutes on a machine with 60 GB RAM and an Nvidia 5070 Ti. The generated report identified three behavioral patterns with 100 percent confidence: File Encryption (supported by 40 functions), C2 Communication and Anti-Analysis. A publicly available YARA rule confirmed the anti-debug findings discovered by the system.

The complete source code is available in the repository Dvoranchik/MLMalwareAnalyzer. Future work will focus on refining call-graph traversal and prompt engineering to surface additional indicators.

WannaCry Ghidra Neo4j Qwen3 PyGhidra LM Studio

ShareX Facebook LinkedIn Reddit

Meridian digestThe week's most consequential security stories, distilled into one briefing.See the digest

More in Malware & Botnets

Malware & Botnets

Yoba Campaign Uses Signed ChatGPT and Claude Binaries for DLL Sideloading and XMRig Deployment

Researchers at Check Point Research have uncovered the Yoba malware campaign that distributes fake desktop clients for ChatGPT and Claude through GitHub repositories. The attackers bundle legitimate signed executables from OpenAI and Anthropic with a malicious DLL that gets loaded via classic sideloading. Once executed, the loader establishes persistence using the Windows Task Scheduler and retrieves additional payloads through Telegram, DNS over HTTPS, and the Catbox file hosting service. The campaign ultimately deploys the XMRig cryptominer, which pauses mining activity when user interaction is detected to avoid detection. No compromise of OpenAI or Anthropic infrastructure was found; the attackers simply repurpose the companies' signed binaries and reputation. The operation primarily targets home users and small-to-medium businesses with Windows systems.

AntiMalwareOct 28, 2026Translated from Russian

Malware & Botnets

uCreator Spyware Infiltrates Modified Telegram, WeChat and GCash Builds on Android

Security researcher Ovi has published a detailed analysis of uCreator, also known as YingGai, a commercial Android spyware platform designed for surveillance and remote app control. The malware intercepts messages, verification codes and account data directly inside legitimate applications without requiring root privileges. Infection vectors include tampered Telegram, WeChat and GCash APKs distributed through unofficial sources, as well as the Onion Box container that launches the target app inside a controlled environment. Once active, uCreator hooks into the host process, reads Telegram’s local message database through client interfaces, captures two-factor authentication passwords and hides its own session from the user. The operation is supported by cloud-based command panels and exposed GitLab repositories that revealed server-side code and development history. Compatibility problems have been noted on Android 17, and the total number of victims remains unknown.

AntiMalwareOct 26, 2026Translated from Russian

Malware & Botnets

Researchers Deploy Cowrie Honeypot Disguised as NVIDIA H100 Server to Study SSH Botnet Tactics

A cybersecurity team set up a realistic Cowrie honeypot on a fake server equipped with four NVIDIA H100 GPUs, 128 AMD EPYC cores, and terabytes of storage to observe what SSH brute-force bots actually do after gaining access. The trap used a custom StickyAuth module that allowed entry only after several failed attempts and locked in the successful credential pair per IP address. Over the first 12.5 hours the system recorded 1,846 connections from 126 unique IPs, 175 successful logins, and 1,524 commands executed, with the majority of activity traced to a single botnet linked to the Outlaw group. The attackers deployed SSH keys signed mdrfckr, changed the root password, killed competing processes, installed cron jobs, and downloaded a mining archive called dota3 containing an IRC backdoor and its own SSH scanner. The researchers also identified three realism issues in the base Cowrie setup—limited grep flags, outdated filesystem paths in Debian 12, and inconsistent hostname reporting—and patched them to keep sophisticated bots engaged longer. An auxiliary LLM service later summarized each session in Telegram for easier analysis.

HabrOct 25, 2026Translated from Russian