QotW #22: What are legal/ethical concerns to bear in mind, when hacking websites with open invitations?

2012-04-06 by . 2 comments

Post to Twitter

This weeks question of the week was asked by user Yoav Aner, who wanted to understand the legal and ethical concerns of executing an attack on a web site which carried a notice inviting attacks. Yoav specifically wanted to know what, if any contractual implications there were and if it were not specified, how far would be too far. This spun off into two further questions - What security measures to have before openly allowing security researchers to hack your site and What security concerns should one bear in mind when hacking open-invitation websites? so this post will look at all three.

Before we start, I must re-iterate: we are security professionals here, not lawyers, so if in doubt, consult a lawyer.

Legal and ethical concerns:

The answer Yoav accepted was provided by Rory McCune, who raised the point that ultimately, a no holds barred approach is extremely unlikely to be acceptable in any circumstance. Rory highlighted the importance of ensuring the page in question was written by the administrators of the site, as opposed to being supplied through user content. Clearly, unless it is clear the page was written by someone with the authority to make that kind of invitation, attacking it would definitely be hostile.

Another excellent point raised in this answer was that some companies actively invite finding bugs in their web sites and products, provided you follow a number of guidelines. More on this can be found in the answer itself.

Finally, Rory touched on what many people may forget – although the website may invite attempted break-ins, it may actually be illegal where you are even to make the attempt.

Our next answer was provided by Security moderator, user and blogger Rory Alsop. According to Rory, one problem when dealing with this kind of issue is that no test cases have yet passed through the courts – so until they do, it’s unlikely any precedent has been established for dealing with these kinds of issues. Rory also raised the criminal activity point again. Always understand that the law of your country/jurisdiction still applies.

Next up, Rory explained that during a penetration test a contract established for the work may include rules about what should happen, how far a test may go, who should be notified if a vulnerability is found etc. Even with this safeguard, there is still the potential for legal action should something break. Rory advised logging absolutely everything that was going on so as to have proof of actions.

When applied to the website scenario, Rory pointed out that in this case there is no signed contract establishing this understanding, just an implication of one which neither party is legally bound to.

That is all for answers on this question. I tend to miss questions on ethics on the main site, so writing them up is actually quite interesting. As such, I am going to summarise the key points below:

  1. The rules of engagement are not well established. Assuming a “feel free to hack this message” we have no idea to what extent that is actually what they mean. By contrast, penetration testing is usually better scoped.
  2. The author of the page might not have the authority to make such an invitation. As I was reading this, I did wonder – if this is a shared host, the administrator of the site is a different person from the company who owns and maintains the box. So even though the site author can put up this message, they’re not actually entitled to make that call (and it probably violates the ToS on their hosting package).
  3. It may be illegal to engage in the act of attempting, whether or not the site in question has given you permission.
  4. Some sites actively encourage hunting for bugs.

Security concerns when hacking open-invitation websites:

Iszi raised the following worries on his question

  • The site could be a honeypot, run by government or other entities looking to gather information about active (or would-be) hackers.
  • The site could be set up by a black-hat as a honeypot to gather a list of interesting, hackable amateurs to target.
  • A third-party black-hat could potentially access the site’s logs and farm them for data about interesting, hackable amateurs to target.

Lucas Kauffman confirmed that he had a school project where he faked an open sendmail relay and

just piped all the incoming emails to a python script that got all the destinations out, generating my own spamlist. I think in the end after about 3 weeks I had close to 300.000 different email addresses.

Rory Alsop focused on the reputational and professional risks, as the host of the site will be able to see everything you did in their logs…do you ever mistype commands, use dir instead of ls, accidentally stray outside the scope of the test? This will be recorded and could negatively impact you.

Think about what you are divulging when hacking a website….

  • Your methodology
  • Your tools
  • Your mistakes?
  • etc

Finally – Yoav also asked a question focusing on the other side,

What security measures should I have in place before inviting people to hack my website?

Ttfd’s answer went into some considerable detail on the practical logistics – how you think about the problem is probably as important as actually implementing security in this situation:

  1. Do you have the money to do that?
  2. Do you have the resources? (servers, security teams and etc)
  3. How far can you limit the damages a hacker can make to your system? I.E. If a hacker hacks into your server what access will he have ? Will he be able to connect to your database and retrieve/store/update data? Is your data encrypted ? Will he be able to decrypt it? (and so on)
  4. Can your security team find how a hacker exploited your system?
  5. Does your security team have the skills to fix problems that may occur ?
  6. Probably many more questions that you need to ask and answer before you decide.

M15K gave a summarised answer

I can’t imagine very many positive scenarios in declaring open season on you’re front door will result in something useful. But let’s say you do, and you do get some positive feedback. Are you and your security team in a position to remediate those vulnerabilities?

Interesting stuff! All in all, this looks like a relatively risky business on both sides, so core to the decision must be a full understanding of the risks, and how these match to your risk appetite. If you are hosting such a site, you may get some valuable information into attack techniques, but you need to protect yourself from an escalation from the attack environment to your own systems. If you are testing the site, think about the risks you may be facing, and plan accordingly.

Liked this question of the week? Interested in reading it or adding an answer? See the question in full. Have questions of a security nature of your own? Security expert and want to help others? Come and join us at security.stackexchange.com

2 Comments

Subscribe to comments with RSS.

  • Yoav Aner says:

    Proud to having made it to the ‘question of the week’! Thanks for a great summary. It’s nice to see how questions spin others (mine was by itself triggered by another great question, so perhaps this one deserves the question of the week…

    Personally however, I was a little discouraged from the “bottom line” message of the answers. I believe there is benefit in having an open approach, encouraging security-research and experimentation, and of course full disclosure of vulnerabilities. Somehow if I’m reading through all those risks people mentioned, the underlying message is that both researchers and website should perhaps avoid using such an open collaborative approach, and I think it’s a real shame if this will always be the case.

    • roryalsop says:

      I think you are right Yoav – they do encourage research, but I think the message was one of assessing the risks and planning accordingly. In many cases the risks will be minimal or non-existent but it pays to carry out sanity checks prior to this sort of thing.

  • Leave a comment

    Log in
    with Stack Exchange
    or