Turns out that CopyQ is a little flakey with the WebAPI copy command I'd opted for. Often it's fine, sometimes not, just a bit here and there and I lacked the or energy to diagnose the nuances, and developed a a general sense of frustration with it alas.
So I looked for a simpler, more reliable and perhaps native solution to inspecting the clipboard contents closely, including it's MIME part breakdown.
As I'm on Linux I found the lowest level of reliable inspection was
xclip easily installed on Debian/Ubuntu derived systems with
sudo apt install xclip.
Subsequently I could see the contents in parts as follows:
xclip -selection clipboard -o -t TARGETS
After a copy operation, for example, I see:
$ xclip -selection clipboard -o -t TARGETS TIMESTAMP TARGETS SAVE_TARGETS MULTIPLE STRING UTF8_STRING TEXT text/plain text/html
And I can see the parts with:
xclip -selection clipboard -o -t text/plain
xclip -selection clipboard -o -t text/html
which works a dream. This can be piped through grep to find things, but the HTML is not formatted.
So I installed HTML tidy.
xclip -selection clipboard -o -t text/html | tidy -qi --wrap 0
Produces a nicely formatted view of the HTML MIME part.
And to get the size of the HTML on the clipboard:
xclip -selection clipboard -o -t text/html | wc -c | numfmt --to=iec
all of which proved very useful inspection and diagnostics when experimenting with what to copy and how.