If you’re not gluing your vulnerabilities together you are likely missing the bigger picture

If you’re not gluing your vulnerabilities together you are likely missing the bigger picture

?

TL;DR

If like most, you separate your security estate and vulnerabilities into atomic chunks, by project, business unit, and CVE, then you will likely be missing critical attack chains, and that will leave your assets exposed to attack.

Context

If you’ve worked within security for any length of time, you’ll have seen the cute memes that circulate, poking fun that just because you have risk-accepted an issue, or categorised something out-of-scope, it doesn’t mean the hackers are playing by the same rules. And whilst these memes are meant as a bit of light-hearted humour, they are also 100% correct.

In the same way, people (and by extension) the tooling, like to break things up into manageable chunks. Technical vulnerabilities are resolved down into individual CVEs. The budgets for assessments and reporting are broken down by team, business unit, and application, and so the scope will likewise be defined along the same lines.

However, that’s not how the hackers actually work. They have no interest in your arbitrary departmental lines, or industry-standard categorisation scheme. They are just looking for an effective attack they can use.

Example One – Triage by CVE Impact

CVE classification is great for tool efficiency, and clarity of root-cause reporting, but it really doesn’t reflect the reality of how exposed your estate is. For a start, at the CVE level, most organisations are only interested in high impact issues and above. I have lost count of the number of project wash-ups I have sat-in on, where the other vulnerabilities aren’t discussed at all. However, it is often trivial to combine a few lows together into an effective attack, with a high impact.

As a simple example of this, you often see a finding for allowing the email address attached to an account to be changed, without requiring authentication. And this will generally be rated as a low. ?You also often see CORS misconfigurations, which again will be rated as a low. But a path that is affected by both could allow a full account take-over, locking a valid user out of their own account, and leaking all their data. Not a low impact.

And there are thousands of other combinations like this. Ooops.

Example Two – Triage by Organisational Unit

Budgets and reporting are generally organised along departmental, or project boundaries. And as such, so is the security testing and reporting. However, the actual vulnerabilities can happily be combined across projects. More oooops.

As a simple example of this, something like a cookie XSS vector will be reported as just a reflected value, and probably only a low or informational impact, as on its own, it is indeed unexploitable. But to make it useful, all that is required is another low impact, response header injection (CRLF) vulnerability, anywhere in the eTLD. That’s right: anything that shares your TLD: the dev platforms that you don’t care about, because they don’t have live data in them; the marketing blog; the third-party e-commerce platform. And likewise, all of those may have rated the response header injection low, because there is nothing in their scope that it can be combined with.

Improving your Workflow

If your tooling, suppliers and workflow do not currently consider chaining vulnerabilities across your entire estate into meta-attacks, then you are missing important issues that may eventually come back to bite you on the butt.

Introducing some form of cross-platform workflow, where you regularly review all the outstanding vulnerabilities, to see if they can be combined, should help pick up a lot of these kind of issues. But doing this at scale will likely be unmanageable. This is a problem that is crying out for some form of automation!

Nota bene: it is worth noting, that any red team on the payroll should be looking at precisely this kind of thing. If not, then you should be questioning why that is so.

About the author

Martin O’Neal has spent the majority of his life gainfully employed in the shady world of information security. He's seen things you people wouldn't believe. Dandruff in drifts on the shoulders of Ryan. Watched LEDs glitter in the dark near Bishopsgate. All those moments will be lost in time... like tears in rain...


?

?

要查看或添加评论,请登录

Martin O'Neal的更多文章

  • Find a Zero-Day, Win a (Shit) Sandwich

    Find a Zero-Day, Win a (Shit) Sandwich

    TL;DR: the bug bounty model is fundamentally broken. The platforms profit, the programmes save on security costs, and…

  • TL;DR no matter what anyone says, it’s not practical to find all the security bugs in pre-prod

    TL;DR no matter what anyone says, it’s not practical to find all the security bugs in pre-prod

    As you already know, I’m a bit of a security dinosaur, and have been working red and blue team gigs pretty much since I…

    8 条评论
  • Traumadar

    Traumadar

    TL;DR Trauma leaves its mark on those of us who have experienced it, shaping behaviours, emotional patterns, and even…

  • Is your instrumentation a liability?

    Is your instrumentation a liability?

    TL;DR It is now common to use various instrumentation tools to track user engagement, but if you’re not monitoring and…

  • Eating broccoli makes you a vegetarian, right?

    Eating broccoli makes you a vegetarian, right?

    TL;DR If someone has a varied CV, it doesn't mean anything more than they have a varied CV. Blah blah blah I’m out…

  • Physician heal thyself!

    Physician heal thyself!

    It’s funny, but I had a little ironic grin to myself yesterday, about how easy it is to fall into the trap of thinking…

    5 条评论
  • Buzzword bingo

    Buzzword bingo

    It has been really interesting to follow the recent okta hack, along with the fallout that has come from it (where okta…

  • Building bridges, not burning them

    Building bridges, not burning them

    TL;DR As a wandering consultant, I often see security teams that make their own job difficult, by being confrontational…

    1 条评论
  • Eating your own dog food

    Eating your own dog food

    You may not realise it from my youthful looks (I wish), but I’ve been hacking other people’s stuff for over 40 years…

    1 条评论
  • Learning to value your mistakes

    Learning to value your mistakes

    I’m regularly asked what I think are the most useful tools and skills for breaking into apps. And my answer is always…

    1 条评论

社区洞察

其他会员也浏览了