Post

@arnaudmontagard

More from @dcuthbert

Daniel Cuthbert

@dcuthbert

Jun 21

I'm going on a web app security rant, so bear with me.

23 years ago OWASP was formed and it tried to help the web application space and those building apps to do so in a secure way. Session management was one of them.

If you had a token, in a header/cookie, make it secure

We've evolved this over the decades, we collectively got better with understanding the nuances and complexities of this identity being thrown around and the consequences of not doing it right.

We have standards, STANDARDS man, github.com/OWASP/ASVS/blo…

and yet, even with all this VC cash thrown at people wearing Patagonia vests and wearing apple watches, I still see most SaaS solutions doing stuff we did when Rain Forest Puppy was hurting SQL

Read 10 tweets

Daniel Cuthbert

@dcuthbert

May 2

Patch ALL teh things we constantly tell CISOs and CIOs.
Thing is, let's be honest with each other right? we can't and this graph is telling.

Patching is a pain, we get it and we do need to revolutionise the approach. Two years ago, @LargeCardinal wrote a phenomenal paper

where, in essence, the idea was to prioritize patches by expressing the connectivity of various vulnerabilities on a network with a QUBO and then solving this with quantum annealing.

Now working with Mark often has me saying 'dafuq you saying bruv?'arxiv.org/pdf/2211.13740

but once he's put down his markers and explained it to me like the child I am, it made sense.

A QUBO problem involves finding values for binary variables (i.e., variables that can only be 0 or 1).

Read 10 tweets

Daniel Cuthbert

@dcuthbert

Apr 11

Strap in, we's going on a ride, a static analysis ride. I recently came across this paper, which looked at a wide variety of SAST tools against a number of Java apps.

Java being the choice of enterprise, and often not the best Java approaches out there, so it's a good choice

First up, what did they use and what did they benchmark it against?

They looked at free tools, tools that specifically supported Java and most importantly, are being actively maintained.

The target was the @owasp project, a good choice imho. They also looked for Java apps with bugs with disclosed CVEs which was around 680 programs.

This gives them a fair amount of apps to look at, which is goodowasp.org/www-project-be…

Read 21 tweets

Daniel Cuthbert

@dcuthbert

Oct 11, 2023

Bugs happen but it's rare you see a bug that grabs you so hard and makes you nod like a little dog..

CVE-2023-44487 did that for me

good god what a bug and here's why

First up is understanding the key differences between HTTP 1.1 and 2, especially how requests work

HTTP 1.1 is a text-based protocol that uses a single connection for each request/response pair. Every time you request the / from , it will be a diff request NSA.gov

for each element of that page (CSS, images etc)

HTTP 2 is a binary protocol that utilises multiplexing, which allows multiple requests and responses to be sent simultaneously over a single connection

much better. more efficient. many wows

Read 16 tweets

Daniel Cuthbert

@dcuthbert

Jun 9, 2023

@Apple

An interesting new feature found in @Apple’s latest privacy and security report is that of Link Tracking Protection and I’ve not stopped thinking about this

First up it’s pretty cool. My views on the pervasive nature of the tracking industry are not something I’ve hidden away: it’s an ugly industry with no real oversight, so any efforts to put a finger in their eye is one to applaud

The approach by Apple is interesting

First up is the deeper inspection (I’m assuming client only) that intercepts any url and does a regex on it to strip out utm and other crap added to the url

If it works like that, I’m impressed. However, how much stuff will it break in the process? I guess time will tell

Read 6 tweets

Daniel Cuthbert

@dcuthbert

Jun 9, 2023

Here’s the thing right: if you are building any application/binary or indeed something that takes input and uses that to form the basis of further functions/actions, you kinda need to think about robustness.

Imagine a HTTP POST request to /remote/portal/bookmarks

What is needed is Content-Length, which indicates the size of the corresponding body. This is how the web works, so to send and indeed accept a zero byte body is odd and you’d check for that right?

Bueller? Right??

@watchtowrcyber

Well it seems not and there’s a brilliant write up of why this was a problem that caused a segfault in a SSL VPN appliance by Aliz Hammond over at @watchtowrcyber

labs.watchtowr.com/fortinet-and-t…

Read 9 tweets

Share this page!

Enter URL or ID to Unroll

Daniel Cuthbert

Try unrolling a thread yourself!

More from @dcuthbert

Daniel Cuthbert

Daniel Cuthbert

Daniel Cuthbert

Daniel Cuthbert

Daniel Cuthbert

Daniel Cuthbert

Did Thread Reader help you today?

Don't want to be a Premium member but still want to support us?

Send Email!