MOSS Bypasser
Demonstrates how easily MOSS plagiarism detection can be bypassed with simple code transformations
Note: Mossad is another tool by UMass that demonstrates how to break MOSS detection, highlighting the need for more robust plagiarism detection systems.
Understanding MOSS Vulnerabilities
🎯 What is MOSS?
MOSS (Measure of Software Similarity) is a plagiarism detection system developed by Stanford University in 1994. It uses text pattern matching to detect code similarity in programming assignments.
🔓 Known Limitations
MOSS can be bypassed using simple transformations: variable renaming, loop restructuring, comment injection, and whitespace changes. These modifications break pattern matching while preserving functionality.
🛠️ Other Tools
Mossad, developed by researchers at UMass, is another tool that demonstrates how MOSS can be defeated. These tools highlight the need for more advanced detection systems that understand code semantics, not just syntax.
Why Modern Detection Matters
- ✓ Text-based matching only - doesn't understand code logic
- ✓ Cannot detect AI-rewritten code (ChatGPT, Copilot)
- ✓ Vulnerable to simple obfuscation techniques
- ✓ Modern systems use semantic analysis to detect similarity even after transformation