InfoResearchPreprintLLM-specific
Hardening JavaScript Programs against LLM-Based Vulnerability Detection
- Published
- Record updated
Summary
AntiAgent is a defense that transforms JavaScript programs before release to impede LLM-based vulnerability detection agents. It evolves a corpus of programs with known vulnerabilities, keeps versions where the agent misses or misreports the flaw, and extracts those transformations into reusable rules. In evaluation on real-world JavaScript applications against four detection agents, including Claude Code and Codex, it reduced detected vulnerabilities by 52.0-60.7%, while transformed programs passed all project tests.
Mitigation
AntiAgent: transform the program before release using transformation rules extracted from an evolved corpus, with the task split between an LLM and the rules to preserve program semantics.
Related items
- InfoCan AI really help spot Black Friday deals? We put it to the testSame vendor · BBC Technology
- InfoIn a year of AI solving maths problems, Anthropic may have just solved the biggestSame vendor · BBC Technology
- InfoAnthropic shares fall after blockbuster opening as wider tech sell-off continuesSame vendor · BBC Technology
- InfoAnthropic mathematician claims to have solved prestigious Millennium Prize problem with ClaudeSame vendor · The Verge (AI)
- InfoQuoting Levent AlpögeSame vendor · Simon Willison's Weblog