LATIDIA · Ciberseguridad

SyzHarness: Reproducción de errores de kernel basada en parches con arneses de difuminado sintetizados por LLM

arXiv: 2609.23889v2Announce Type: replace Resumen: La reproducción automatizada de vulnerabilidades del kernel es esencial para el triaje de errores, la validación de parches y las pruebas de regresión, pero aún carece de una solución efectiva y eficiente.

La noticia

arXiv:2609.23889v2 Announce Type: replace Abstract: Automated kernel vulnerability reproduction is essential for bug triage, patch validation, and regression testing, but still lacks an effective and efficient solution. The core challenge is twofold: a reproducer must first recover the trigger scaffold needed to reach the vulnerable state and determine the precise concrete values that actually trigger the bug. Existing directed fuzzing approaches are ineffective at recovering the necessary trigger scaffold, while LLM-only generation is brittle because it struggles with concrete-value discovery and runtime nondeterminism. We design SyzHarness, a framework that combines LLM reasoning with coverage-guided fuzzing for patch-based Linux kernel vulnerability reproduction. Given a patch, SyzHarness uses an LLM agent grounded