I don't think this is the right set of tags to use here. While it might be technically correct (I'm not sure whether that is the case), it is inconsistent with the tagging used on every other road in the terminal (as well as every other port I've looked at). It currently misleads people looking at the rendered, who will likely be led to believe (as I was) that the link from the A20 is closed (perhaps due to some need to manage queues).
I think it would be better to remove the current access tags, and instead add bicycle=no, foot=no and horse=no. (Are any of these explicitly signposted?)

This looks like a mechanical edit (i.e. one systematically with minimal checking of what is actually on the ground). Have you discussed this on the relevant mailing lists, as required by the Automated Edits Code of Conduct (osm.wiki/Automated_Edits_code_of_conduct).
Also, note that page for the entrance key explicitly warns against making automated edits in this scenario. (entrance=*).

Some of the issues that arise can be hard to foresee. For instance, this particularly changeset could potential have broken the University of Cambridge map (map.cam.ac.uk), although I think entrances were already broken by a similar mechanical edit three years ago.

This has previously been commented on by robw on changeset/78190542 and Borbus on changeset/78120031, and I made this change partly on the basis that they were suggesting similar things. I think the remark about "only tag construction sites" is more about not mapping roads as closed if they are going to reopen again soon because it might complicate routing decisions. The issue here is different, because it's comparing two different ways of tagging the road while representing it as closed.

You seem to have accidentally moved a node near Bradford as part of this changeset. Was that intended to be a separate changeset, or was the movement just a mistake?

In the imagery this looks like a vehicle routes within a temporary A14 works compound. Since the works at this location are complete, have you got any evidence that there is still a compound there, and that the vehicle routes match this layout? I would expect that it is already being restored to farmland.

This link probably also no longer exists (see my comment on 78800735). I think this changeset should probably also be reverted (it maps part a temporary construction related to construction that has already finished).

This is almost certainly not the current layout - since the A14 opened two weeks ago, I highly doubt that there is still a link onto the carriageway used by construction traffic (and if it does exist, then it won't for much longer). I think this needs reverting, unless you have more recent evidence (i.e. from the past two weeks) to show it still exists.

You seem to have accidentally mapped a school in the middle of the north sea. Presumably this was supposed to be in/near Abidjan. Can you fix this please? Also, I'd be interested to know how this mistake happened in the first place.

Also, can you write more informative changeset comments please - it makes it much easier for people to work out what changesets are supposed to do without having to look at all the details. For instance, a much better comment for this changeset would have been "Tag users prohibited from new A14 bypass, and tag viaduct as bridge=viaduct". A similar remark applies to many of your other changesets - mentioning that you've changed some tags doesn't really say anything about what the changeset actually represents.

A couple of questions:
Why did you only add these tags to a few parts of the new bypass? Surely they ought to be added to the entire route between New Ellington and Swavesey (including the slip roads).

Why did you add a new node to one of the ways? This looks like a misclick to me.

Did you intend to move a trunk_link node onto an admin boundary? I've deleted the node since it seemed to be redundant in its new location, and I don't think the two ways should share nodes here.

A small tip about fixing errors - check the details of the changeset that introduced those errors. In this case there were a number of mistakes introduced in one changeset (https://overpass-api.de/achavi/?changeset=50973528), and you only fixed some of those mistakes. Checking the history also allows you to give feedback to people who make mistakes (as I have just done for this particular changeset).

I think I've now reverted all the mistakes in the changeset (aside from those that were already fixed).

You seem to have changed a lot of roads around Dover from highway=tertiary to highway=service. I'm pretty sure that most of these were errors, and you only intended to make changes to the port roads. I have therefore reverted this change in changeset/78448114 for all the roads outside the port that hadn't yet been reverted.

While I'm here, I have a more general remark about your changeset comments. These are often not very descriptive - for example I looked at four different changesets with the changeset comment "Road(s) amended" and saw that they were doing very different things. Better changeset comments in those cases would have been something like "Change roads in Dover Port to highway=service", "Add X road to map", "Update speed limit on part of X road" and "Correct name of X road". i

More precise comments (like those I suggested) make it a lot easier to check what changesets are intended to do without looking at them in detail, and would make it clear that (a) editing the other roads outside the port in this changeset was a mistake; and (b) no other changesets at the same time were doing similar things, so there's no reason to check those changesets for the same sort of mistake.

You've reinstated the name, I think, but I'd like to point out that the existence of another road with the same name 2 miles away definitely does not constitute any evidence at all that a way has been misnamed. Also, in this case it is easy to check that the county's highway records have this as Brampton Road (although they probably still reflect the pre-A14C2H layout).
https://www.cambridgeshire.gov.uk/residents/travel-roads-and-parking/roads-and-pathways/highway-records/

This changset has also added some untagged ways on the coast of The Wash - what is going on there?

Why did you add those tags to only two of the ways in the bypass? Why did you add a random node to one of the ways? What is your source for these restrictions (I think I've heard about them several times, but a clear, authoritative source would be nice). And why does your changeset description include the word "minor"?

Anyway, good spot about the access restrictions; we just need to get this mapped consistently on the whole bypass, and ideally with a reference to a good source that allows people to verify the exact nature of the restrictions.

That road is certainly not a dead end - it leads through to Grafham. I haven't looked into it enough to form a strong opinion, but I think tertiary would be more appropriate for the entire length. (Does the tertiary tag generally mean a C-road classification in the UK? I'm not sure.) I don't agree that marking it as unclassified makes the map 'look more appealing'.

Can you please try to use more descriptive changeset comments? Railway fixes could mean that you're fixing broken ways, or changing incorrect tagging. In this case it seems you're adjusting the geometry of the railway. What source are you using as the basis of your adjustments? Also, you've actually made the alignment noticeably worse in parts - at the northern extent of your changeset the spacing between the track is now very inconsistent, with two of the tracks almost crossing over.

Oh, apologies, I just realised that you didn't mark the roundabouts as open - that was someone else a few hours earlier. In any case, the rest of my comments still apply, and I've now reverted the relevant part of your changeset so I can mark the roundabout as partially closed again.

You marked the roundabouts as open again, whereas many comments on Facebook say that they are still closed. Since you didn't mention that change in your comment, and didn't include a source, I've reverted that in 78184278. This is the second time I've reverted that change since Sunday night (and the third time in total). Please check the changeset history around the A14 works before future edits - there are many people contributing, and it's unhelpful if they keep making the same errors. Also, make sure your changeset description describes *all* of the changes, and if you split it into separate changesets if you're doing multiple separate things remotely, then it makes it easier to revert errors.