Daniel, would you mind taking a stab at increasing the granularity of our metrics? Currently, we're measuring things at a very high level in https://cs.chromium.org/chromium/src/third_party/WebKit/Source/core/dom/ClassicScript.cpp?l=19

It would be helpful to get more detailed in two ways:

1.  HTML bans a few types that we don't currently ban. We should add specific metrics for `text/plain`, `text/xml`, `application/octet-stream`, and `application/xml`. I imagine `text/html` is going to be massive, but we should split that out too.

2.  It would be helpful to split out document contexts from `importScripts()` in Workers. Perhaps we can be stricter with the latter than we can with the former.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/dea1cddff24486b92732bedf6e98f7f1d5a3ed48

commit dea1cddff24486b92732bedf6e98f7f1d5a3ed48
Author: Daniel Vogelheim <vogelheim@chromium.org>
Date: Tue Oct 10 11:08:51 2017

Add UseCounters to measure legacy mime type usage with 'nosniff'.

Bug: 765544
Change-Id: I40feae9fa7e93f87592415db267ac3149f128a60
Reviewed-on: https://chromium-review.googlesource.com/702475
Reviewed-by: Mike West <mkwst@chromium.org>
Commit-Queue: Daniel Vogelheim <vogelheim@chromium.org>
Cr-Commit-Position: refs/heads/master@{#507635}
[modify] https://crrev.com/dea1cddff24486b92732bedf6e98f7f1d5a3ed48/third_party/WebKit/Source/core/BUILD.gn
[modify] https://crrev.com/dea1cddff24486b92732bedf6e98f7f1d5a3ed48/third_party/WebKit/Source/core/dom/ClassicScript.cpp
[add] https://crrev.com/dea1cddff24486b92732bedf6e98f7f1d5a3ed48/third_party/WebKit/Source/core/loader/AllowedByNosniff.cpp
[add] https://crrev.com/dea1cddff24486b92732bedf6e98f7f1d5a3ed48/third_party/WebKit/Source/core/loader/AllowedByNosniff.h
[add] https://crrev.com/dea1cddff24486b92732bedf6e98f7f1d5a3ed48/third_party/WebKit/Source/core/loader/AllowedByNosniffTest.cpp
[modify] https://crrev.com/dea1cddff24486b92732bedf6e98f7f1d5a3ed48/third_party/WebKit/Source/core/loader/BUILD.gn
[modify] https://crrev.com/dea1cddff24486b92732bedf6e98f7f1d5a3ed48/third_party/WebKit/Source/core/loader/resource/ScriptResource.cpp
[modify] https://crrev.com/dea1cddff24486b92732bedf6e98f7f1d5a3ed48/third_party/WebKit/Source/core/loader/resource/ScriptResource.h
[modify] https://crrev.com/dea1cddff24486b92732bedf6e98f7f1d5a3ed48/third_party/WebKit/Source/core/workers/WorkerScriptLoader.cpp
[modify] https://crrev.com/dea1cddff24486b92732bedf6e98f7f1d5a3ed48/third_party/WebKit/public/platform/web_feature.mojom
[modify] https://crrev.com/dea1cddff24486b92732bedf6e98f7f1d5a3ed48/tools/metrics/histograms/enums.xml

The *OriginWorker* numbers look pretty good in favor of blocking, although they may get worse (as workers become more prevalent) unless we start blocking them soon?