Nobody really knows where worm authors go shopping for exploits to develop, but its widely assumed that they are greatly assisted by exploit code released by legitimate researchers.
Go look at most vulnerability reports, and youll see references to where exploit code may be obtained. Why would a "legitimate" researcher do such a thing?
If you think about how youd want to run security management at a large organization with some time and budget behind it, its not hard to see how exploit code is valuable. We—meaning vendors and researchers and media—tell you to test patches before installing them on the production network.
Of course, a lot of you would get laughed at by your bosses if you proposed creating a test network, but assume you were testing a patch before deploying it. You would want to test your software with the patch installed to make sure it still worked correctly, but youd also want to test the patch to make sure it worked correctly.
For this, you need exploit code. You could also use it to test workarounds in case you dont want to deploy the patch immediately. Finally, you cant just trust your perimeter defenses, so you need exploit code to run penetration tests, also known as "pentests."
Of course, if youre a malicious "script kiddie" looking to impress your fellow vandals, canned exploit code makes your "job" a lot easier. This was the conclusion of Johnny Cyberpunk, a researcher at The Hackers Choice. Mr. Cyberpunk recently announced that he will "personally not publish any further exploits to the public."
The basic story for Mr. C was that the legitimate uses "didnt work." He doesnt explain how they failed, although he does say many users didnt know how to do the customization necessary to make exploit code work correctly. And he saw too much risk of it being used for untoward purposes.