From: Eugene Leitl (email@example.com)
Date: Mon Mar 13 2000 - 18:10:44 PST
Joseph S. Barrera III writes:
> While at Microsoft I thought about how easy it would have been to embed code
> in the various Windows variants in places that would not easily be
You really think M$ OS/apps have no remotely exploitable backdoors?
(I.e. via "overseen" buffer overruns; NSA has certainly the expertise
to code extremely devious things requiring very special set of
conditions to triggered and hence virtually impossible to detect even
by close scrutiny), done both by disgruntled employees, and
fed-mandated ones? Have you never wondered about these xDSL modems,
which always check on company's server whether there is a fresher OS
to flash? About cell phones possibly with remote software upgrade
hooks? (if they're not yet out there, they soon will). ISDN phone
systems? Great tools for packet sniffing/portable audio bugs with ~100
feet precision locating capability. These Nextel things Rohit so fond
of have certainly sensitive enough mikes, as we today inadvertely
found out at the company I'm with.
> discovered. Cracking security to get the code into the source tree was, of
> course, the easiest part. I used to rant about the number of machines with
> null Administrator passwords, but no one ever cared. Transitive closure is a
> powerful concept... no less so when applied to security...
One of best way to insert above hidden backdoors in OpenSource stuff
is to become a contributor good enough to sneak a few lines of
crypto-malicious code into an otherwise highly functional module.
Speaking of which, just acting as a source of confusion at
organisation/communication in an OpenSource project could be an
considerable source of FUD. I'm sure the bad guys will try that route,
This archive was generated by hypermail 2b29 : Mon Mar 13 2000 - 18:15:06 PST