Add the missing v-TP3-COM-IMAGE tag, and say why it is placed differently

SUMMARY.md's milestone table named v-TP3-COM-IMAGE but the tag did not exist:
13 rows, 12 tags. Found by diffing the rows against `git tag -l` rather than by
reading the table, which is the only way a claim like that shows up.

It is placed on 7dada57 -- the commit that added Linker.mod, ComTest.mod and
run_com_tests.sh -- not on a commit whose tree contains this table's account of
the milestone, because that documentation was not written until 82c4146, by
which time .COM images had already been executed and the row said "done,
executed". v-TP3-BP-CONTRACT went the other way and was re-pointed for exactly
this reason; the two placements disagree, so the tag message says so rather than
pretending the convention is uniform.

The gap is in the documentation's history, not the code's, and it cannot be
repaired now: the commit that would justify a differently-placed tag does not
exist yet.

SUMMARY.md

@@ -25,6 +25,15 @@ manual, not guessed.
 | Runtime entries under qemu, and the register contract stated | `v-TP3-BP-CONTRACT` | done, **36/36 entry checks; `wrchar`/`wrbool` no longer destroy BP** |
 | `CmdRun` (the `R` key), in-process 8086 interpreter | — | **not started** |
 
+Every row that names a tag has one, and every tag points at a commit on
+`master`; verified by diffing the rows against `git tag -l`, which is how
+`v-TP3-COM-IMAGE` was found to be missing. One tag does **not** follow the
+convention used for the rest: it names the commit that did the work
+(`7dada57`, where `Linker.mod`, `ComTest.mod` and `run_com_tests.sh` appear)
+rather than a commit whose tree contains this table's account of it, because
+the documentation for that milestone was not written until three commits later
+and by then it already said "executed". The reasoning is in the tag message.
+
 ## Build
 
 ```sh