Honeypot-Driven Decoy Block Vulnerability Obfuscation: A Structured Framework for Software Protection Against Symbolic Execution Attacks
Zhizhuang Zhou · 2025
Symbolic execution, a powerful program analysis technique, poses a significant threat to software security by efficiently exploring program paths and exposing vulnerabilities. To address this challenge, we propose a novel Honeypot-Driven Decoy Block Vulnerability Obfuscation (HD-DBVO) framework, which integrates opaque predicates and decoy vulnerabilities to mislead symbolic execution tools. Our approach embeds strategically designed decoy vulnerabilities behind opaque predicates resistant to symbolic execution, thereby diverting analysis tools toward false paths while ensuring the integrity of the actual program logic. Experimental evaluation reveals that our technique increases symbolic execution time by 4.58 times, expands the number of explored paths by 32 times, and achieves an 83.7% false positive rate for decoy vulnerabilities, with a negligible runtime overhead of 6.3% and a 21.9% increase in code size. These results demonstrate that HD-DBVO effectively mitigates the impact of symbolic execution attacks, offering a promising avenue for enhancing software security.