Application Security This Week for July 29

Venmo, a social payment system, defaults to public disclosure of payments made on the system.


Scott Simmons has some terriffic advice about using Same-Origin policy as a control for CSRF.


Open redirect flaw in Electron exploites in the new Google Hangouts Chat application.


F5 has released their annual Application Protection report.  Worth a read.


DOMpurify, a common control for DOM based XSS, has a vulnerability - update if you are using it (you probably are).

Application Security This Week for July 22

It has come to my attention that one of Paul Asadoorian's Security Weekly broadcasts is titled Application Security Weekly! I had no idea. It's good too, you should listen.  I caught up with the last few weeks when I drove over to Indianapolis to chat with the Indy Software Artisans meetup.  Anyway, I am changing the title of this recurring series of posts to Application Security This Week because of the mixup.


Interesting discussion over at El Reg about the weakest link in software security.


Oracle addressed 334 security vulnerabilities in its latest patch series.


Shape Security did the math, and 9 out of 10 login attempts on the web are bypass attempts.

Raspberry Pi project: network speed tester

For my home internet, I have one choice - Spectrum, née Time Warner Cable.  I don't completely live in the boonies but it is far enough away from everything that I can't get anything but cable, and Time Warner is the only cable company I got.  Honestly it hasn't been bad - I have been a customer since they were Roadrunner, and have had nearly zero problems.  Had a mixup when we moved our phone service to them, which led to us dropping phone altogether and just using mobile (and the alarm system for emergencies). Had a couple useless repair folks.  Over a 20 year history with them, I'm pretty pleased.

However, we have started having brownouts.  We'll be tooting along, then the feed will go to a dribble. I pay for 50 Mbps down, and usually get 70, but then it will be 500 Kbps for hours.  Then it will be fine.  Is it Spectrum?  My modem?  My router?  Something else on my network? I wasn't sure, so I would lug my laptop back there, plug the ethernet from the cable model directly in, and run Speedtest. Started to get some data.  Then, the laptop that had an ethernet port went titsup.  Now what?

Enter the Speedtest Pi.  I scrapped a screen from an old POS appliance that I tested and was allowed to keep, an old mini keyboard, and one of the dozen Pis that I have laying around from various programs I have taught, and built a semi-permanent speed test appliance that I can go and use anytime I want.

I did need to buy an AV shield at Microcenter so I didn't have to soldier the heck out of the Pi, but that's OK. This was a quick and dirty job, and the shield was only $15.

Now - best way to do the speedtest?  Well, did you know there is a Speedtest API? I DID NOT.  To make it even better, there is a speedtest command line interface app in the install paths for Raspbian.  So I simply:

sudo apt-get install speedtest-cli

and that's all there was to it.

Next next step is to put a fast switch in, so that I can leave it on all the time, and then have it run every hour or so, and show a graph on the screen.  Should be a fun Python project.

Application Security Weekly for July 15

npm is a dumpster fire.  Yet another malicious package discovered that it automagically brought into many projects thanks to dependencies.  In other news, I learned about snyk, which is a pretty cool tool.


In dev news, the #1 development GUI of all time is being updated.  Notepad!


Apple wrote some code to appease the Chinese government and it was kind of a mess.


Vuln-lab found a neat XSS vulnerability on an AT&T site's profile feature.


Remember when I said "Spectre is not exploitable"?  Yeah, I was wrong.  Again, and again, and again...


New variation of my favorite Weblogic vuln - CVE-2017-10271.

I wrote the tests for this vulnerability for Nikto.


And that's the news.

Application Security Weekly for July 8

LTE has a bug.  Who knew? One more strike for IoT devices, methinks.


Cool XXE Vulnerability in WeChat Pay SDK.


UK's National Health Service had a breack due to a currently unspecified coding flaw, keep an eye on the story for more info.

Application Security Weekly for July 1

It's the "Bill accidentally skipped a week" edition.  I didn't even DO anything last Sunday, I just forgot!


The IETF calls for formal revocation of the TLS 1.0 and 1.1 standards.  This will effectively cut mobile users on Android 4.4 and earlier off the web.  Guess who this hurts: developing countries. And why?  Because it's possible to decrypt a message BEFORE the heat death of the universe.  We have a priority problem.


Rhino Security put together a good article about privilege escalation on Amazon Web Services, and it is juicy.

They have an open source AWS scanning tool too!!


This isn't a security story explicitly, but it is about why security in apps for mobile is so important, and it features Columbus, where I am based.  And it is The Atlantic, one of my favorite papers.


There's a 7-month-unpatched vulnerability in Wordpress that allows for unauthorized access.  Considering what Wordpress has grown into I'm kind of shocked by this.


A breach bigger than Equifax?  SURE WHY NOT.


While I am eating up your Wired soft-paywall allowance, they have another good article on how the Mirai botnet was just some kids trying to cheat at Minecraft.  Great long read.  Don't screw with malware, folks!

By the way, Wired has great reporting and is worth the $10 a year.  You should subscribe.


And that's the news.  Have a great 4th, if you are in the US.  Otherwise, have a great week!

Veracode partnership

I don't do a lot of advertising on this blog, but this is a pretty important part of my "walk the talk" campaign.  I have for years been espousing a four part analysis pattern, including manual dynamic analysis (vulnerability analysis), manual static analysis (code review), automatic dynamic analysis (scanning the app with something like ZAP), and automatic static analysis (code scanning).  Well, I have added this last one, automatic static analysis to the list of products that POINT offers, with a partnership with Veracode.  Veracode offers automatic static binary analysis, and is the best product I've found for web applications and mobile applications.  What's more, I can triage the findings for you before delivery. (I'll of course also give you the original test results).  I spoke on this in my talk from a couple of years ago, Developers: Care and Feeding.

I also spoke at length on the topic on the Brakeing Down Security podcast

So now, I offer this for real.  It's not free, but it's a great addition to a vulnerability analysis, and I'm pleased to be able to add it to the suite of offerings we have here at POINT.

Application Security Weekly for June 17

The Android Debig Bridge (ADB) feature is even less secure than we thought.  Avoid those "recharge stations"


A tale of the disclosure of WebUSB vulns.


In the "let's be clear" department, Microsoft explains what is will and will not fix.


And that's the news.

Application Security Weekly for June 10

Firstly, I have had a MASSIVE chest cold that has kept me down for the count, so I have been reading a lot of news.  Thus, long newsletter.


Microsoft bought Github.  This might seem to not be a security issue, but 'tis.  Why did they buy them? Github doesn't make money.  However: 1) Microsoft wants devs on their platform and 2) they are really into machine learning.  So, let's get all of the devs and all of their code and ... profit?


This is a little older but was new to me - Bruce Schneier writing for Lawfare (recommended reading by the way) about the implications of Efail.


A cartoon intro to DNS over HTTPS.  We need more of these.


Building malicious zip files.  Remember, mess with malware in a virtual machine, and NOT on your company network please.


Didier Stevens is oft referenced in these missives, and he had a really productive May.  I'll just link to his own overview.  Lots of great appsec content.


XSS on ESPN's site.  Stuff is just everywhere:


Oh man, I forgot about this one.  Remote Code Execution on a voice-based AI.  You know, one of those smart speakers?  Incredible stuff.  Now I wanna go test my Echo.


And we'll finish up with a breakdown by El Reg of all of the week's data breaches.


Have a good week, everyone. I'm going back to bed. Oh, and that's the news.

Application Security Weekly for June 3

My good friends at AppSec Consulting tipped me off this this really neat finding .  It's a SAML bypass - they didn't discover it but they have been using it in tests and it works well.


Remember JScript, that attempt by Microsoft to take over ECMAscript?  Yeah, neither does anyone else but it is still in Windows and it has an RCE vulnerability.


Apparently it's the theme today, so I'll point out that an RCE vulnerability was found in the Steam client, and has a good writeup.


In a previous post I mentioned the sheer mass of Redis servers left open on the Internet.  Someone has now written a worm for them, and 75% are infected.


And that's the news.


Bill Sempf

Husband. Father. Pentester. Secure software composer. Brewer. Lockpicker. Ninja. Insurrectionist. Lumberjack. All words that have been used to describe me recently. I help people write more secure software.



profile for Bill Sempf on Stack Exchange, a network of free, community-driven Q&A sites