$1 Million Dollar Bounty successfully hack the SAFE Network.
Using “Prize Indemnity Insurance” - an insurance policy sold by specialized underwriters (think golf hole-in-one contests which are insured all the time) - pay a small percentage of the $million but it buys some great publicity.
The very challenge of testing the network for the underwriters would garner tremendous publicity.
Bring in KPMG or Deloitte for audit and invite Brian Krebs (Security expert @briankrebs) to be the grandmaster.
Yeah I think we need to set some ground rules that social engineering need not apply. If you get someone’s pass by breaking their social web of trust that doesn’t prove the technical system is broken. It just proves you’re a good con artist.
Also what about types. Look, in a hole-in-one tourney, there’s one hole and one ball, and if the ball goes into the hole, there’s a winner. What constitues a hack of the SAFE Network? For that matter, what constitues a “hack” of the internet? Sorry, that’s too broad a topic to go risking money on.
Also consider - brute-forcing passwords, physical hardware manipulation, and like @Blindsite2k said, social engineering.
[quote=“smacz, post:5, topic:5755”]
What constitues a hack of the SAFE Network? For that matter, what constitues a “hack” of the internet?
[/quote] Excellent questions. Now we need some excellent answers.
1.) >70% accuracy traffic correlation attack.
2.) Data flow tagging attacks.
3.) Vault data decryption.
4.) Forced quorum group attendance.
5.) Sybil attacks.
6.) User to SAFE man in the middle attacks.
7.) Data origin attacks.
8.) Low resourced data flooding attacks.
9.) Forcing disproportionate safecoin allocation from SAFEnet.
10.) Safecoin double spending.
11.) Denial of service attacks.
I guess it should go without saying that attacks on the test-net are somewhat welcomed for the sake of initial hardening but will not qualify for the bounty until safecoin is out of test mode.
@dirvine Are there any attacks you would personally like to see attempted when the network is stable?
I would say that there are three kinds of potential hacks to be worried about.
First would be piercing the veil, that is identifying an IP address or data map, in some way compromising the layers of anonymity.
Second, corruption/deletion, that is compromising all copies of at least a single chunk of a file thereby effectively preventing recovery of that file.
Third, a true hack, where the data map is compromised or by some other means a hacker is able to retrieve a whole decrypted file.
IMO, the million dollars should only be paid out for a true hack. There should be lesser prizes for the other two.
Great thoughts @kirkion ! Thats why KPMG or Deloitte and Krebs are involved … to set / enforce the rules of play. Any way you cut it HUGE publicity for SAFE -
Whether $1M or another means to generate interest, it’s a great idea to announce a challenge on hacking the design integrity of SAFE. There could be a public checklist like @Tonda started where hack participants can attempt an attack, document what was done and the result of the attack then get rewarded for it in some way – i.e. not necessarily with only maidsafe/safecoin. Keeping these hard core tests transparent would garner a lot more public confidence that SAFE is in fact safe and delivers.
I still think these deserve their own bounty and defense strategy. They may not be hacking the actual network but they can be effective none the less. Wetware hacking (social engineering) needs to be defended against just as much as software hacking.
Well for starters if you want to claim the bounty you have to show your work. The attack has to be recorded and/or reproducible so that it can be performed by others, or yourself under controlled conditions. After all the goal here is to find and patch the hole. So if you show your work and you’re see to be using social engineering or malware then you end up being disqualified. If it’s a sybil attack or something it qualifies. Of course it’s entirely possible someone will invent an entirely new attack never seen before but that has yet to be seen. So basically just like anything else here you need proof of work or resource in order to get paid.
Another idea is to pose a friendly challenge to colleges and universities to see if their math and computer science grad (and undergrad) students can hack/attack the SAFE network. An event such as a weekend or week long attack-a-thon might entice some friendly competition as well. That way rules such as use of social engineering can be defined and agreed to formally. Getting academic participation would be huge for quick SAFE adoption and for getting some public press.
The network most certainly must know your IP in order to know where to respond to with the GET fulfillments! However, there is only one point on the network that knows your IP address - the relay node.
So there’s the point of attack - the relay node. That is the part of the network that does have the information. As for the rest of the network, you are correct. It doesn’t have that information. But one element in it does.
So intercepting and decrypting the IP address, convincing the node to transfer the IP (spoofing/replay?), node impersonation, intentionally becoming the node, or otherwise obtaining the IP address of a target client would all be symptoms of a broken implementation of the network resulting in a “hack”. Again, this would be a broken implementation and hopefully a fix would be able to be engineered.
As far as verification goes, the attacker might have to go after a dummy client with a known IP address, and presenting documentation describing the hack and the result.
I am a strong proponent of utilizing a target/Capture The Flag mechanism for a pen test such as this.
EDIT: To further lengthen this already long post, @polpolrene wrote up a quick explanation of this: