TriVul: Improving Precision of Static Vulnerability Detection through Multi-Agent Reasoning
Published in SAS'26: 33rd Static Analysis Symposium, 2026
TriVul is a multi-agent assisted static analysis framework for Java vulnerability detection that integrates heterogeneous static analyzers with stage-specific LLM reasoning to reduce false positives while preserving recall.
Recommended citation: X. Zhao and Joanna C. S. Santos. (2026). "TriVul: Improving Precision of Static Vulnerability Detection through Multi-Agent Reasoning." Accepted to SAS'26: 33rd Static Analysis Symposium.
Download Paper