Malware Source Code and Decompiled C Pseudocode from Vxunderground

A curated dataset of malware source codes and C-like pseudocode obtained through automated decompilation using Ghidra. The dataset is intended for malware analysis, program analysis research, and machine learning-based detection. It was created by author 'qwewerewq' and last updated on 2026-06-14.

Use Cases

• Train malware detection models based on decompiled source code.
• Conduct program analysis research based on C-like pseudocode.
• Perform reverse engineering experiments using decompiled malware artifacts.
• Benchmark LLM-based code analysis tools on malware samples.

Strengths

• The dataset is curated, suggesting some level of quality control.
• Data is processed using Ghidra, a reverse engineering framework developed by the NSA.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

Based on data from Vxunderground.

Collection Method

Automated decompilation using Ghidra.

Freshness

Last updated 2026-06-14 19:18:08; freshness should be verified.