You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Static, dynamic, and sandbox analysis of three real-world malware samples to identify malicious behaviour, extract indicators of compromise (IOCs), and document adversary techniques. Findings were compiled into a formal analysis report.
Tools Used
Tool
Purpose
Ghidra
Static analysis & reverse engineering
Process Monitor
Runtime process & file system monitoring
Process Hacker
Live process inspection & memory analysis
Regshot
Registry & filesystem snapshot comparison
Wireshark
Network traffic capture & analysis
Methodology
Sample 1 — Static Analysis
Reverse engineered a Windows PE sample using Ghidra to
identify malicious functionality without executing the binary. Traced
decompiled function calls across execution flow, uncovering behaviours
including remote registry manipulation, hidden file/process operations,
credential and keystroke capture routines, screen capture capability, and a
range of anti-analysis and defence evasion techniques (VM/sandbox detection,
delay-loaded imports, UI/taskbar hiding, COM object abuse). IOCs were
extracted directly from API calls, string artifacts, and registry paths
referenced in the decompiled code, and organised into categorised tables
(network activity, system modifications, registry, credential harvesting,
keylogging, screen capture, defence evasion, and fake UI initialization).
Indicators of Compromise (IoCs)
Static malware analysis identified the following categories of Indicators of Compromise (IoCs), each backed by specific API calls, registry paths, or string artifacts observed during binary analysis.
Note: Static analysis has inherent limitations — behaviours gated behind runtime-resolved values (e.g. pointer/flag fields filled in only at execution) may be hidden from static review, which can introduce false positives or false negatives.
Executed a confirmed SnakeKeylogger sample (verified
via MalwareBazaar) in an isolated environment to observe live behaviour. Used Process Hacker, ProcMon,
Regshot, and Wireshark to monitor process activity, file system and registry
changes, and network traffic in real time. Documented process hollowing into
RegSvcs.exe, ZoneMap registry manipulation, credential harvesting from
Chrome and Edge, and C2 communication disguised as fake Windows Update
traffic, including geolocation of the C2 infrastructure.
Malware Behaviour
Malware Behaviour
Evidence
Masquerades as a PDF document
Executable uses the filename Sales Order Sheet.pdf.exe to trick users into believing it is a legitimate PDF.
Installs persistence
Creates and modifies numerous registry keys and values (199 keys added, 386 values added, 115 values modified) to maintain execution after reboot.
Steals user credentials
Identified as a credential stealer (SnakeKeylogger) targeting stored usernames and passwords.
Logs keystrokes
Dynamic analysis classifies the sample as a keylogger capable of recording keyboard input.
Performs anti-analysis techniques
Uses execution delays (Sleep) and other evasion techniques to hinder sandbox analysis.
Modifies the Windows Registry
A total of 934 registry changes were observed during execution, indicating extensive system modification.
Injects or interacts with system processes
Behaviour indicates interaction with Windows processes to execute malicious functionality while reducing detection.
Attempts lateral movement
Sandbox behaviour indicates the malware attempts to spread or interact with other systems on the network.
Maintains long-running execution
Remained active for over 11 minutes during analysis, suggesting continuous background activity.
The third sample was submitted to Cuckoo Sandbox for automated dynamic analysis. Identified as a 32-bit Windows GUI executable, the sample returned a maximum threat score (10/10) and could not be hashed via ssdeep — an early indicator of packing or obfuscation, later confirmed by YARA rule matches for UPX and Armadillo packing. Behavioural signatures and antivirus consensus (50/57 on VirusTotal) classified the sample as a banking trojan/spyware consistent with the Zusy and BlackMoon families, exhibiting keylogging, screen capture, registry modification, private profile data access, file operations, and hooking capabilities.
Further analysis revealed the malware allocates memory with full read/write/execute permissions (PAGE_EXECUTE_READWRITE), consistent with Armadillo unpacking and executing its real payload directly in memory at runtime to evade static and signature-based detection. Metadata embedded in the binary suggests a Simplified Chinese-speaking author, though this indicator was treated with appropriate scepticism given the ease of forging or copying version resource metadata. Despite confirmed HTTP communication capability, no network activity was observed during sandbox detonation — assessed as a likely anti-analysis technique (sandbox detection or delayed execution) rather than an absence of C2 functionality.
About
Static and Dynamic Malware Analysis of Real-world Samples, along with Online Sandbox Analysis