But, isn’t an OSINT Report designed to take OS info and “connect the dots”, providing ^likely^ conclusions? The conclusion itself and the “how” of connecting those dots would be classified, but NOT the OS info.
Example would be akin to our connecting the dots involved with ADD students, medication provided by a doctor with ties to INTEL staff, MKULtra, an axe to grind, and known cellphone fookaree that might be used to trigger.
Leading some to conclude a likely conclusion that the Parkland HS shooting was designed by our own INTEL and triggered when the situation seemed appropriate.
Now, instead of that conclusion being classified, it is dismissed as “conspiracy theory”.
But, isnt an OSINT Report designed to take OS info and connect the dots, providing ^likely^ conclusions? The conclusion itself and the how of connecting those dots would be classified, but NOT the OS info.
You are talking of two things. The open source material and the 'product' that results from it that satisfies a Request for Information (RFI). One could be classified, and unless the OSINT portions were in unclass paragraphs, you would have to develop the OSINT from a different source before using it in the clear.
Your MKUltra example is different in that a) there is no RFI from any agency, and b) you can only assume the MKUltra capabilities and other links even if you see them in writing since you have no provenance of their origin. The difference between 'assessed capability' proven and therefore must be planned for by Operations and suspected 'unproven.
I apologize, it has been awhile since I did collections vs analysis. LP likely has more recent or deeper experiences.