diff --git a/.github/workflows/auto-release.yml b/.github/workflows/auto-release.yml
index e0c7fd8..f4ea0a2 100644
--- a/.github/workflows/auto-release.yml
+++ b/.github/workflows/auto-release.yml
@@ -1,15 +1,13 @@
-# Auto-tags and cuts a GitHub Release whenever pyproject.toml's version changes
-# on main and the Python Test workflow has passed for that commit. Publishing
-# a Release (types: [published]) is exactly what python-publish.yml already
-# listens for, so this is the only piece needed to make merging to main (with
-# a version bump) publish to PyPI, with no separate manual release step.
+# Auto-tags, cuts a GitHub Release, and publishes to PyPI whenever
+# pyproject.toml's version changes on main and the Python Test workflow has
+# passed for that commit. No manual "cut a release" step needed.
 #
-# The release-creation step deliberately uses the RELEASE_TOKEN secret (a
-# PAT), not the default GITHUB_TOKEN: GitHub Actions does not fire other
-# workflows' triggers (release: published included) for events performed
-# with the automatic GITHUB_TOKEN, to prevent recursive runs. A release
-# created with a PAT is attributed to a real actor instead, so it triggers
-# python-publish.yml normally. See README "Releasing" for one-time setup.
+# The publish job calls python-publish.yml directly via workflow_call rather
+# than relying on the release: published event to cascade into it: GitHub
+# Actions doesn't fire other workflows' triggers for events performed with
+# the automatic GITHUB_TOKEN (anti-recursion safeguard), so a release created
+# here wouldn't otherwise trigger it. Calling the job directly sidesteps that
+# entirely -- no PAT/secret needed, since it's one run, not two.
 
 name: Auto Release
 
@@ -30,6 +28,8 @@ jobs:
     runs-on: ubuntu-latest
     permissions:
       contents: write
+    outputs:
+      released: ${{ steps.check.outputs.exists == 'false' }}
     steps:
       - uses: actions/checkout@v4
         with:
@@ -55,13 +55,17 @@ jobs:
       - name: Create tag and GitHub release
         if: steps.check.outputs.exists == 'false'
         env:
-          GH_TOKEN: ${{ secrets.RELEASE_TOKEN }}
+          GH_TOKEN: ${{ github.token }}
         run: |
-          if [ -z "$GH_TOKEN" ]; then
-            echo "::error::RELEASE_TOKEN repo secret is not set (needed so this release triggers python-publish.yml -- see README 'Releasing')."
-            exit 1
-          fi
           gh release create "v${{ steps.version.outputs.version }}" \
             --title "v${{ steps.version.outputs.version }}" \
             --target "${{ github.event.workflow_run.head_sha }}" \
             --generate-notes
+
+  publish:
+    needs: release
+    if: needs.release.outputs.released == 'true'
+    permissions:
+      id-token: write
+      contents: read
+    uses: ./.github/workflows/python-publish.yml
diff --git a/.github/workflows/python-publish.yml b/.github/workflows/python-publish.yml
index 00325f5..8c35b0d 100644
--- a/.github/workflows/python-publish.yml
+++ b/.github/workflows/python-publish.yml
@@ -12,6 +12,7 @@ on:
   release:
     types: [published]
   workflow_dispatch:
+  workflow_call:
 
 jobs:
   deploy:
diff --git a/README.md b/README.md
index fe908bf..5ec238c 100644
--- a/README.md
+++ b/README.md
@@ -300,21 +300,17 @@ async with pool.connection() as con:
 Bump `version` in `pyproject.toml` as part of your PR, same as any other
 change. Once that PR merges to `main` and the `Python Test` workflow passes
 for that commit, `.github/workflows/auto-release.yml` automatically tags it
-`vX.Y.Z` and cuts a GitHub Release (skipping if that version was already
-released, e.g. a merge that didn't touch the version) — no manual release
-step needed. Publishing a Release is what `.github/workflows/python-publish.yml`
-already listens for, so the new release then builds and publishes to PyPI
-via trusted (OIDC) publishing automatically.
-
-`workflow_dispatch` on `python-publish.yml` still works as a manual
-fallback if you ever need to re-publish a version without going through a
-fresh tag/release.
-
-**One-time setup:** `auto-release.yml` needs a `RELEASE_TOKEN` repository
-secret — a PAT (classic, `repo` scope, or fine-grained with this repo's
-Contents: Read and write) belonging to someone with write access. This is
-required because GitHub Actions doesn't fire other workflows' triggers
-(`release: published` included) for a release created with the automatic
-`GITHUB_TOKEN`, to prevent recursive runs — a release created with a PAT is
-attributed to a real actor instead, so it triggers `python-publish.yml`
-normally. Add it under Settings → Secrets and variables → Actions.
+`vX.Y.Z`, cuts a GitHub Release (skipping if that version was already
+released, e.g. a merge that didn't touch the version), and publishes to
+PyPI via trusted (OIDC) publishing — no manual release step, and no extra
+secret to configure. It calls `python-publish.yml`'s `deploy` job directly
+(`workflow_call`) rather than relying on the release it just created to
+trigger that workflow on its own — GitHub Actions doesn't fire other
+workflows' triggers for events performed with the automatic `GITHUB_TOKEN`,
+so a manually-created-via-Action release wouldn't otherwise cascade into a
+publish; calling the job directly sidesteps that instead of working around
+it with a PAT.
+
+`workflow_dispatch` (or an actual GitHub UI release) on `python-publish.yml`
+still works as a manual fallback if you ever need to re-publish a version
+without going through `auto-release.yml`.