Defusing Logic Bombs in Symbolic Execution with LLM-Generated Ghost Code

This collection of PL/systems papers (Mar 2026) includes several items with notable security implications. The standout is Gordian: a hybrid symbolic-execution framework that uses LLM-generated "ghost code" to make SMT solvers handle solver-hostile fragments, improving coverage (52–84% vs. traditional SE; 86–419% vs. prior LLM approaches) while drastically reducing LLM token use. Security impact: substantially stronger symbolic execution improves automated vulnerability discovery and exploit development, and LLM-assisted code synthesis could both aid analysis and introduce risks if ghost code/