Showing posts with label litchfield. Show all posts
Showing posts with label litchfield. Show all posts

Friday, September 12, 2008

not sure if this is good or bad...

full disclosure foo... so litchfield is a ninja and all, but i'm torn on this one...

here is a no-auth remote compromise of oracle db's from a few months back...

NGSSoftware Insight Security Research Advisory

Name: PLSQL Injection in Oracle Application Server
Systems Affected: Oracle Application Server 9.0.4.3, 10.1.2.2, 10.1.4.1
Severity: Critical
Vendor URL: http://www.oracle.com/
Author: David Litchfield [ davidl at ngssoftware.com ]
Reported: 9th October 2007
Date of Public Advisory: 15th July 2008
Advisory number: #NISR15072008
CVE: CVE-2008-2589


do you see what i saw?

it was publicly disclosed in july 08, but reported to the vendor in *oct 07*. no-auth remote compromise just hanging for the better part of a year...

i'm sorry, but if it really takes that long to dev a security patch, oracle is doing something really really wrong.

this is one of those times where (imho) dropping 0day to kick vendors in the arse is completely justified. not weaponized or anything, but get that info out there. how many other peeps found that vuln and didn't disclose? no one will ever know...

Thursday, April 24, 2008

lateral sql injection

so litchfield just posted a pdf on what he calls lateral sql injection...

basically, the attack focuses on situations where you can affect a function which doesn't take any parameters. normally you'd assume such functions were immune to attack. but he takes a side-channel approach and alters the output of internal commands called by the function which are used in sql queries.

as he says at the end of the paper, the attack vector probably isn't going to be seen all that often. i'm def not a sql/db expert, but it seems like you'd need a decent amount of knowledge about the underlying code being used in a system to attack it via lateral sql injection... of course, there are probably some really common stored procedures, and perhaps an attacker could make reasonable guesses as to what a developer called in his or her code...

anywho, it's always fun to see people looking at things in new ways...