The code was published on Wednesday by a Vietnamese researcher Nguyen Jang, according to a report in The Record, a website owned by the CIA-backed threat intelligence firm Recorded Future.
But hours later, GitHub took down the code. iTWire has asked the company why it did this.
British security researcher Marcus Hutchins, the first to make mention of the PoC, said in a tweet that it did not work in its original state, but needed some tweaks.
"This is huge, removing a security researchers code from GitHub against their own product and which has already been patched. This is not good."
Hutchins disagreed, saying: "'Has already been patched'. Dude, there's more than 50,000 unpatched Exchange servers out there. Releasing a full ready-to-go RCE chain is not security research, it's recklessness and stupid."
Another researcher, posting under the handle SBS Diva, backed Hutchins' stance, saying: "'Already been patched.' On behalf of the SMBs that may not have been patched yet, can we acknowledge that not everyone is in the same boat?"
But Kennedy disagreed. "That's not the debate here, it wasn't a working exploit it was a PoC," he said.
"Would I have released one yet? No... I would definitely give it a substantially more period of time. Microsoft doesn't get to dictate time windows for their own product, they delayed a patch for over two months."
Well-known Google researcher Tavis Ormandy was also against the removal of the PoC, pointing out how it could benefit other researchers. "It's unfortunate that there's no way to share research and tools with professionals without also sharing them with attackers, but many people (like me) believe the benefits outweigh the risks," he said.
"For example, @maddiestone does variant and root cause analysis on vulnerabilities. Without easy access to research her job is harder, so there's a net benefit to making sure she has access to exploits."
His reference was to Maddie Stone, a fellow researcher at Google's Project Zero. She chimed in: "Yep it's a larger level of investment to patch diff compiled binaries to determine what you *think* the vulnerability is and *then* be able to evaluate if the patch is correct, perform variant analysis, brainstorm and test detections, vs having the details and a POC.
"For attackers, this is a great return on investment. For defenders who have dozens if not more vulnerabilities on their to-do list, it's often not the best use of time. But when a PoC and technical details are shared, we can skip the step of trying to figure out the vuln."
Stone added: "Also from my time on Android's Malware team, when you have a giant queue of things that need to be analysed and signatured, I could deploy new signatures days or weeks faster when a POC has been provided vs trying to figure out what it may look like myself."
There have been cases in the past when exploits for Microsoft software has been posted on GitHub, with PoC code for a wormable flaw known as SMBGhost released in June last year.
But that flaw did not see half the level of exploitation as the Exchange Server flaws.