Reverse-Engineering Exploits from Patches
This is interesting research: given a security patch, can you automatically reverse-engineer the security vulnerability that is being patched and create exploit code to exploit it?
Turns out you can.
What does this mean?
Attackers can simply wait for a patch to be released, use these techniques, and with reasonable chance, produce a working exploit within seconds. Coupled with a worm, all vulnerable hosts could be compromised before most are even aware a patch is available, let alone download it. Thus, Microsoft should redesign Windows Update. We propose solutions which prevent several possible schemes, some of which could be done with existing technology.
Full paper here.