diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml
index e2802994..e64f0f5d 100644
--- a/.github/workflows/release.yml
+++ b/.github/workflows/release.yml
@@ -23,36 +23,58 @@ jobs:
       contents: write
 
     steps:
-      # Checkout the repository at the branch that triggered the workflow.
-      # Python Semantic Release will automatically convert shallow clones to
-      # full clones if needed. We reset to workflow sha to prevent accidentally
-      # releasing un-evaluated changes if branch was updated during workflow.
-      - name: Checkout repository
-        uses: actions/checkout@34e114876b0b11c390a56381ad16ebd13914f8d5 # v4.3.1
+      # Note: We checkout the repository at the branch that triggered the workflow.
+      # Python Semantic Release will automatically convert shallow clones to full clones
+      # if needed to ensure proper history evaluation. However, we forcefully reset the
+      # branch to the workflow sha because it is possible that the branch was updated
+      # while the workflow was running, which prevents accidentally releasing un-evaluated
+      # changes.
+      - name: Setup | Checkout Repository on Release Branch
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           ref: ${{ github.ref_name }}
 
-      - name: Force branch to workflow sha
-        run: git reset --hard ${{ github.sha }}
+      - name: Setup | Force release branch to be at workflow sha
+        run: |
+          git reset --hard ${{ github.sha }}
 
-      - name: Run Semantic Release
+      - name: Action | Semantic Version Release
         id: release
-        uses: python-semantic-release/python-semantic-release@8f4d8c4e3e4fcd2b8df16444b148aa49920cc4d3 # v10.6.1
+        uses: python-semantic-release/python-semantic-release@9a026e9303981c866c3425723009becb2437c757 # v10.6.2
         with:
           github_token: ${{ secrets.GITHUB_TOKEN }}
           git_committer_name: "github-actions"
           git_committer_email: "actions@users.noreply.github.com"
           changelog: "false"
+          # Commit, tag, push and build, but don't create the GitHub release.
+          # We create it ourselves in the next step so that the distributions
+          # are attached before the release is published. See that step for why.
+          vcs_release: "false"
 
-      - name: Upload to GitHub Release Assets
-        uses: python-semantic-release/publish-action@98e6c872368aff4290ef8444db095e5bcd3fba6c # v10.6.1
+      # This repo has immutable releases enabled, which freezes a release's
+      # assets the moment it is published, so assets cannot be attached
+      # afterwards. `gh release create` handles this by creating the release as
+      # a draft, uploading the assets, and only then publishing it:
+      # https://docs.github.com/en/code-security/supply-chain-security/understanding-your-software-supply-chain/immutable-releases
+      - name: Publish | Create GitHub Release with Assets
         if: steps.release.outputs.released == 'true'
-        with:
-          github_token: ${{ secrets.GITHUB_TOKEN }}
-          tag: ${{ steps.release.outputs.tag }}
+        env:
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          # Reuse the release notes python-semantic-release generated for us.
+          RELEASE_NOTES: ${{ steps.release.outputs.release_notes }}
+          TAG: ${{ steps.release.outputs.tag }}
+        run: |
+          # Output the release notes to a file
+          printf '%s' "$RELEASE_NOTES" > "$RUNNER_TEMP/release_notes.md"
+          # Create a draft release, upload the assets, then publish it.
+          gh release create "$TAG" \
+            --verify-tag \
+            --title "$TAG" \
+            --notes-file "$RUNNER_TEMP/release_notes.md" \
+            dist/*
 
-      - name: Upload distribution artifacts
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v7.0.1
+      - name: Upload | Distribution Artifacts
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
         if: steps.release.outputs.released == 'true'
         with:
           name: distribution-artifacts
@@ -64,19 +86,25 @@ jobs:
       version: ${{ steps.release.outputs.version }}
 
   publish_to_pypi:
+    # 1. Separate out the publish step from the github release step to run each step at
+    #    the least amount of token privilege.
+    # 2. Also, publishing can fail, and its better to have a separate job if you need to
+    #    retry, and it won't require reversing the release.
     runs-on: ubuntu-latest
     needs: release
     if: github.ref_name == 'master' && needs.release.outputs.released == 'true'
 
     permissions:
+      contents: read
       id-token: write
 
     steps:
-      - name: Download build artifacts
-        uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v8.0.1
+      - name: Setup | Download Build Artifacts
+        uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1
         with:
           name: distribution-artifacts
           path: dist
 
+      # Publishes to PyPI via OIDC trusted publishing (no API token required).
       - name: Publish to PyPI
-        uses: pypa/gh-action-pypi-publish@ba38be9e461d3875417946c167d0b5f3d385a247 # v1.12.x
+        uses: pypa/gh-action-pypi-publish@dc37677b2e1c63e2034f94d8a5b11f265b73ba33 # v1.14.2