We have analyzed the three proof-of-concept documents recently posted on the mailing list, and they don't appear exploitable to us. Specifically: - powerpnt.exe: NULL pointer dereference, unusable. Are we missing anything? - memory corruption: we apparently control the address, but not the contents written to it. Doesn't appear usable - mso.dll: this didn't work for us. We see an integer overflow exception being raised with RaiseException, followed by a C++ exception, but no apparent side effects. PowerPoint complains, but closes cleanly and works fine thereafter Is any of the three supposed to go beyond a local DoS? anyone else has analyzed them?
