You don't just download zero day exploits. They aren't known, they are in people's pockets. As well the exploits on the web usually have the port numbers changed in the shellcode, or something along those lines. Many times you would have to know quite a bit to even use an exploit. Vae's original question stated goals of being a security researcher, and I noted the best researchers would proove the problem, so development effort that requires resources would be awarded. Pretty funny if you ask me, you ask me to proove the password can be cracked, when I meant to say something else, and it wasn't my point, while I was saying why prooving things was good for security research. I"m also fairly sure I can generate a collision with MD5. There are known weaknesses in the hashing function. This makes the output of the function not so unique.
Both a security researcher and penetration tester should know software security. An automated scan isn't going to be able to check that every web application doesn't check its input. Why would you test only for common passwords, and ignore out of date software that could root a box, or a web application that could produce a shell? Take a look at metasploit. It's a common penetration testing tool that throws a bunch of exploits at something. You could probably even get more granular with it using nmap to enumerate services (lots of work was put into exposing what's really listening on a TCP/UDP port).
Here's a list of sites I want you to launch your zero-days at:...tastyspleen.net...Let me know how many work and how it all works out for you when you root them.
metasploit is used a lot yes, but those arent zero day exploits exactly... they are often known exploits that some one failed to patch that make breaking into certain versions of software easy, they were zero day at one point in time, but honestly they are easy to execute with the help of something like metasploit rather than having to code your own malware to exploit it or something to that extentand you dont have to be a pro coder to exploit most of the known exploits around lol
and imo nmap is very easy to use, and you can verify service banners with something like httprint to see what exactly is running on a certain port via foot printing rather than banner grabbing i did this against tastyspleen's port 80 and even had a conversation with quadz on it as his banner gives way more information than needed and then we got into a discussion about a security through obscurity thing and blah blah blah... yeah you can even go one step beyond nmap and use something like hping to do some very cool nitty gritty stuff like audit firewall state tables by making custom packets fly through under different protocols to see what exactly the firewall wants to filter so you can bypass it by masking traffic with something else like fragrouter lol (hping does have fragmentation built in, and so does nmap though)
you might have needed to be an epic coder and know some over the top things to hack systems in years past, but that simply isnt the case anymore... you need moderate coding skills to be really pro, but you can get by no problem with very little coding skills due to the amount of tools that exist to make security experts lives easy and a pain all at once