Exploring LLMs + symbolic execution with Z3 for reverse engineering ideas

mgostIH · x · 2026-11-23

The author wants to apply the approach to symbolic execution for reverse engineering, potentially expanding to Z3: the more exploitable structure a problem has, the more low-hanging unexplored ideas LLMs might find with just some tokens. He also rediscovered that problems like parity learning are relevant here — though the effort only started today and lacks concrete workflow details yet.

Original post →

More from coding & agent

coding & agent channel →