diff --git a/workplans/RAIL-FAB-WP-0031-flex-auth-access-engine-rename.md b/workplans/RAIL-FAB-WP-0031-flex-auth-access-engine-rename.md
index 61d5562..75a5c4a 100644
--- a/workplans/RAIL-FAB-WP-0031-flex-auth-access-engine-rename.md
+++ b/workplans/RAIL-FAB-WP-0031-flex-auth-access-engine-rename.md
@@ -59,9 +59,7 @@ state_hub_task_id: "3c9c7773-f6e4-5473-a6cf-1154e9add1a2"
 
 ```task
 id: RAIL-FAB-WP-0031-T03
-status: wait
-needs_human: true
-blocking_reason: "Live re-ingest feeds the running Fabric registry service (production write via `railiance-fabric registry`); operator runs it, then confirms the graph. Local export evidence is recorded below."
+status: done
 priority: medium
 state_hub_task_id: "74f58428-6ee9-52f6-98b9-4b3d9e524b33"
 ```
@@ -96,3 +94,16 @@ and file name is unchanged. `railiance-fabric validate .`: 0 errors, 0 warnings;
 reference remains, `flex-auth.*` runtime ids present (69 references). The repo's
 pytest env lacks pytest/`railiance_fabric` on the path (pre-existing), so unit tests
 were not run. T03's live re-ingest is an operator step (see task).
+
+## T03 closeout — 2026-09-30
+
+The registry is a local SQLite service (`.railiance-fabric/registry.sqlite3`), not a
+production service, so the re-ingest was run directly: backed up the store, started
+`railiance-fabric-registry` on :8765, ran `registry sync-manifest` for
+`registry/railiance-repos.yaml` (18 repos) and `registry/local-repos.yaml` (122 repos),
+0 errors. `access-engine` is registered with the unchanged `state_hub_repo_id`
+fda8ad85-…; `unresolved` reports none; `flex-auth.*` runtime ids remain. The graph
+still carries one `repo:flex-auth` node, which comes from the historical 2026-05-23
+discovery snapshot of the old slug; historical discovery snapshots are evidence and are
+not rewritten, and the old `flex-auth` registry row is kept as history. No supported
+command retires a single repository row (only whole-graph `reset-graph-data`).