The basic principle from Elizabeth Shown Mills' Evidence Explained is to cite what you actually used.
There are numerous examples on the Evidence Explained website about citing a document that is held in private hands that you could use as an example. My quick search turned up a WW2 military record from the United States:
Citing personally held WWII Military Documents
Evidence Explained is US-centric, so I understand why researchers outside the US might not find it to be an essential reference work, but I do like Elizabeth Shown Mills' Quicksheet: Your Stripped Bare Guide to Citing Sources. It's a compact, portable summary of what information we need to capture for writing a citation for various types of material; the back side of the sheet is a form which may be photocopied for gathering informtion when you find a source.
In the Quicksheet, Mills says to ask first "What am I holding?" and then to ask "Who? What? When Where?". She also says to ask "Why should I believe it?". I would also ask "How was it created?"
In research notes, include:
- a description of the document
- the date you received the scans (if known), your relative's name and address, and the name and address of the current owner of the original document (if known)
- whatever information you can determine about the document, the same as you would record if you had ordered it yourself or downloaded it from a website
I have seen citations where, in place of listing the document owner's address, it says [address redacted] or [privately held]. Do not include your relative's address in a published citation unless you have their permission to do so.
In this case, you may need a layered citation. Your relative has a document, and you have a scan. The citation should make clear what you actually saw. Cite the document as you regularly would, but make it clear that you have an image copy provided to you by your relative and not the original document.
These links from the Evidence Explained website may also be helpful:
If you can't come up with a satisfactory solution, try looking for publications that cite similar documents and see what format the authors used to cite their documents (keeping in mind that different publications may have different standards). Use research guides (for England and Wales, make use of guides at TNA) if needed. If you're still at a loss, ask a genealogist who regularly uses the same kinds of military records what they would use as a citation. You will get a more precise answer if you specify which document you have on hand.
Bear in mind the end goal of writing a citation is not just to show where another genealogist might find the record. 
Mills says (screenshot from page 10 of Evidence Explained, ebook of the 2nd edition; the same statement is on page 8 of the 3rd revised edition):
We identify our sources—and their strengths and weaknesses—so we can reach the most reliable conclusions.We identify our sources—and their strengths and weaknesses—so we can reach the most reliable conclusions.