... but that doesn't mean that people weren't working on options in the background...
You are again entering religious realms with that argumentation ("I have not seen God splitting the Red Sea, but that doesn't mean he can't do that").
How large is the R&D department of SpaceX? Today SpaceX has 3000 employees, if you assume a rather high upper limit, that 40% of these are in R&D (you also need people for building, flying and selling the stuff), that means 1200 engineers today (in the past it had been way less, 2005 SpaceX had only 130 employees after starting 2002 with 30 employees). More realistic for a company like SpaceX would be just 25% R&D.
Now look at the many known projects of SpaceX. You have the Falcon 9 1.1, the Dragon, the Falcon Heavy, the Grasshopper. These all also have ground segments as part of their system. The Merlin engine series is its own project there.
I estimate that SpaceX has about 30 major R&D programs running, and about one dozen R&D departments with around 6-8 teams grouped by subject (Means a team size of around 12-16 engineers at the high limit, which is not that large, a smaller R&D would have less teams, not smaller teams).
Now, I work in R&D, and I know what a team of engineers in that size with optimal tools can handle in a given time. It is a lot depending on the technological risk and how standardized the actual project is (Is it what you do every month, or something new, that you need to figure out), but not that much. Every bigger program at SpaceX will have just 3-4 teams assigned to it at the same time. (At the more realistic scale, just 2-3 teams)
There is not much capacity left for doing anything like you claim for further investigation of recovery options.
It is more likely, that the R&D had from the end of the last year capacity left (I would say, maybe as soon as November) and assigned teams to do a number of feasibility studies in that field. Maybe the Falcon 9 1.1 was getting developed faster than expected. Maybe the Falcon Heavy program was slowed down a bit. The number of R&D tasks is usually highest midway in development, and drops slowly shortly before flight testing begins.
Why exactly the capacity was then there, is something nobody outside SpaceX can really know now. What is known: One team was tasked with investigating the possibilities. Maybe more, but that should not be needed.
At the beginning of the year, they reported to the powers that be, presented their findings. And the powers had been impressed by the recovery test idea and decided to push ahead.
In March, when Musk annouced it, they had finished the early concept work for the modification. They had a test plan, an integration concept and the required software modifications by assistance from the Grasshopper team. They could say that there is a 95% chance that it will be ready for the launch date and only a low risk that it would fail the primary mission. If it would have been there sooner, Musk would have annouced it sooner. Musk is not known for keeping exciting things secret for long, he only has his lips sealed when it comes to failures: There he only talks as much as he really must do.
Today, they are busy integrating the whole extra package into the flight hardware. Guidance simulations are done, maybe they also do further tests for engine restarts in flight. Procedures are developed for the launch director team. All stuff that has to be done. SpaceX is not Armadillo Aerospace, which can afford a rather experimental atmosphere, SpaceX has investors sitting in its back.
It is no "uncontrolled hack" as it maybe suggests. But it is rapid accelerated R&D program. Maybe similar to the change of Apollo 8 from LEO test to lunar orbit, but at half the time.
Now tell me: Should I really prove you, that such a program did not exist and did not follow the realistic history, that I deducted? Or can you maybe accept, that SpaceX has limitations and can't research everything everytime.
And if you accept this, calling for evidence that something major was already in development, when SpaceX R&D department was just half of its current size is simply more logical, than calling for evidence, that such a program was already done at a time when there had been far less people and far more tasks to be done - and thus much less likely to exist at that point.
EDIT: And looking at the office space available at their two major R&D sites (Hawthorne & McGregor), it is doubtful that they would really have more than 750 engineers today. For 1200, they would need to stack them in the offices.