diff --git a/.github/workflows/lifecycle_drift.yml b/.github/workflows/lifecycle_drift.yml
index 2760928d..207c647e 100644
--- a/.github/workflows/lifecycle_drift.yml
+++ b/.github/workflows/lifecycle_drift.yml
@@ -4,7 +4,15 @@ name: Lifecycle Drift
 # complete/YYYY/MM invariant and the generated complete/index.md. Fails on:
 #   - lifecycle check drift (a slug in both active.md and complete.md; a
 #     complete/ record with no complete.md entry; a file in two states);
-#   - a stale complete/index.md (regenerate with `lifecycle.py index --apply`).
+#   - on pull requests, a stale complete/index.md (regenerate with
+#     `lifecycle.py index --apply`).
+#
+# On pushes to main a stale index is SELF-HEALED instead (issue #116): the run
+# regenerates it and pushes a bot commit, and fails only if the repair itself
+# fails. Mind pushes land directly on main from many concurrent agent
+# sessions, so ad-hoc mutations of complete/ (e.g. a manual `git mv`) will
+# keep happening — an alarm-only check emailed the human once per push until
+# the next `record --apply` happened to regenerate the index.
 
 on:
   push:
@@ -24,8 +32,9 @@ on:
       - "scripts/lifecycle.py"
   workflow_dispatch:
 
+# contents: write is needed by the self-heal push on main; PR runs never push.
 permissions:
-  contents: read
+  contents: write
 
 jobs:
   drift:
@@ -34,5 +43,43 @@ jobs:
       - uses: actions/checkout@v4
       - name: lifecycle check
         run: python3 scripts/lifecycle.py check
-      - name: index freshness
-        run: python3 scripts/lifecycle.py index --check
+      - name: index freshness (self-healing on push to main)
+        run: |
+          if python3 scripts/lifecycle.py index --check; then
+            exit 0
+          fi
+          if [ "${GITHUB_EVENT_NAME}" != "push" ]; then
+            echo "::error::complete/index.md is stale — run 'python3 scripts/lifecycle.py index --apply' on this branch and commit the result"
+            exit 1
+          fi
+          echo "complete/index.md is stale on main — self-healing"
+          git config user.name "github-actions[bot]"
+          git config user.email "41898282+github-actions[bot]@users.noreply.github.com"
+          # Each attempt rebuilds on the current tip of main, so a concurrent
+          # push (the usual cause of a rejected push) never needs a rebase or
+          # conflict resolution — just regenerate and try again. The heal push
+          # uses the default GITHUB_TOKEN, which does not trigger workflow
+          # runs, so it cannot loop; convergence is verified before pushing
+          # regardless.
+          for attempt in 1 2 3; do
+            git fetch origin main
+            git reset --hard FETCH_HEAD
+            if python3 scripts/lifecycle.py index --check; then
+              echo "tip of main is already fresh (healed by a concurrent push)"
+              exit 0
+            fi
+            python3 scripts/lifecycle.py index --apply
+            if ! python3 scripts/lifecycle.py index --check; then
+              echo "::error::'index --apply' did not converge — lifecycle.py bug, repair by hand"
+              exit 1
+            fi
+            git add complete/index.md
+            git commit -m "lifecycle: self-heal stale complete/index.md"
+            if git push origin HEAD:main; then
+              echo "healed on attempt ${attempt}"
+              exit 0
+            fi
+            echo "push rejected (attempt ${attempt}) — retrying on the new tip of main"
+          done
+          echo "::error::could not push the healed index after 3 attempts"
+          exit 1