Contracts DO NOT Equal Security – New York and Secure Code

by kriggins on January 16, 2009

in General, programming

Hi folks.  Yesterday, I included this story in my Bits post. It is about new procurement language that says software vendors must "certify" that their software does not have any of the Top 25 Errors released by SANS/CWE early this week.

I have read several blog posts on the topic since and today the topic came up on The Security Catalyst Forums. (You should check those out it if you haven't already. Great conversations and community.)

One of the questions posed was this; does this approach seem like something that should be encouraged?

Below is the response I posted.

Two main things pop out at me with this type of thing.

The first is this phrase "must certify that they have rid their code of the Top 25 Errors." What about the next 25 or the next one? I read a blog post over the last couple days that talked about this very well. Blocking where I saw it. If I find it I will update the thread. The essential bit was that "certifying" that you have addressed the top 25 errors doesn't mean your software is secure. That "26th" error  can be a show stopper too. Say it with me, compliant does not equal secure. Before people yell at me, I am not implying that we shouldn't address the errors listed in the top 25. (side note: Kees and some others have been pointing out that the 25 may not really be 25)

My second concern is this, sayin' it doesn't make it so. Creating contract language like this can lead an organization to a false sense of security. I can see where orgs might go the route of "the contract says the software is secure so we don't need to test it or perform a risk assessment." Again, that 26th error can hurt a whole lot.

Just my 2 cents worth. It's super cold in Iowa, so flame away Smiley

Like it says above, these are my thoughts. What are yours?


{ 3 comments… read them below or add one }

Dave Hull January 16, 2009 at 1:56 pm

Three points.

1) If this is NY’s only means of achieving security in the applications they buy or build, then they are screwed.

2) This is a Top 25 list from the CWE project. There are more than 700 items in the CWE catalog.

3) But, consider that the two of the entries on the list are responsible for more than 1.5 million web site compromises and this seems like a good start that should have a positive impact on web app security over the next few years… until number 26 becomes the new number one.


cji January 16, 2009 at 2:13 pm

The procurement language gets a lot more involved than just “have none of the top 25” – it mandates secure coding training, etc. But still a valid point – #26 could be a show stopper too.

P.S. – He was probably thinking of the Matasano blog post:


kriggins January 16, 2009 at 2:16 pm

@Dave, agree that addressing these errors is a good thing.

@cji Yup, that’s the one. Thanks!



Leave a Comment

Previous post:

Next post: