diff --git a/docs/UNIFIED-TASK-TRACKER.md b/docs/UNIFIED-TASK-TRACKER.md
index 1d974621..e9b02a5b 100644
--- a/docs/UNIFIED-TASK-TRACKER.md
+++ b/docs/UNIFIED-TASK-TRACKER.md
@@ -82,11 +82,46 @@ those are marked ✅ below with the commit that did it, so we stop re-litigating
     still pruned — that ceiling is shared by every non-.228 node), but **fully
     synced** (`ibd:false`, blocks==headers 956,240). Bootstrapped bats 1.11.1 +
     jq 1.7.1 onto it 2026-07-01 and **launched the 5× destructive gate
-    (`ARCHY_ITERATIONS=5 ARCHY_ALLOW_DESTRUCTIVE=1`) — running now**, log at
-    `/tmp/gate.log` on .5, background poller watching for the `RESULTS` banner.
-  - Once .5's gate reports: bring the rest of the fleet to precondition, then the
-    cross-node federation/mesh/transport suites. This is the literal
-    "next exit criterion" called out in `CLAUDE.md`.
+    (`ARCHY_ITERATIONS=5 ARCHY_ALLOW_DESTRUCTIVE=1`).
+  - ✅ **Gate finished 2026-07-01: 5/5 iterations, technically all "FAIL" per
+    run-gate.sh's tally — but only because .5's pruned-bitcoin limitation
+    (expected, permanent, accepted going in) fails one test every single
+    iteration.** By iteration 4-5 that was down to exactly 2 failures per run:
+    the expected pruned-bitcoin one, plus a reproducible `lnd` proxy timeout
+    (`https://host/app/lnd/`, distinct from the DNS bug below — happened
+    consistently on both of the last 2 iterations, worth its own investigation,
+    not yet root-caused). Iterations 1-3 also hit test-suite bugs since fixed
+    live mid-run (see Tier 0/below) and one ~2h hang (also below) — none of
+    those are real product bugs.
+  - 🐛 **Real hang found + fixed live**: `lnd` cached a dead IP for
+    `bitcoin-knots` after an earlier restart gave it a new container IP —
+    every RPC needing chain data blocked forever (client-side `timeout`
+    wrappers don't reliably kill `podman exec`'s in-container process).
+    Blocked iteration 4 for ~2 hours before diagnosed + fixed (`podman
+    restart lnd`, forces fresh DNS resolution). **Product-level gap, not
+    fixed at the code level**: dependent services should reconnect/re-resolve
+    after a backend container is recreated, not cache indefinitely. Logged as
+    a follow-up, not yet implemented.
+  - Next: bring the rest of the fleet to precondition, then the cross-node
+    federation/mesh/transport suites. This is the literal "next exit
+    criterion" called out in `CLAUDE.md`.
+- [x] ~~**Real bug found + fixed 2026-07-01**: boot reconciler left any
+      genuinely-installed-but-fully-absent app stuck forever unless it was one
+      of 8 hardcoded "required baseline" apps~~ — surfaced by indeedhub's
+      backend containers (minio/postgres/relay) never recovering on .116 after
+      going absent. Root cause: `ensure_running_with_mode()`
+      (`prod_orchestrator.rs`) only called `install_fresh()` for
+      `is_required_baseline_app()` apps in the fully-absent case; every other
+      installed app was left as `Left("absent")` with no path back short of an
+      explicit reinstall. Fixed: self-heal now applies to any app that
+      reaches this point (i.e. already confirmed NOT user-stopped / NOT
+      user-uninstalled earlier in the same function — those markers are
+      properly set/cleared on uninstall/reinstall, so this can't resurrect a
+      deliberately-removed app). Deleted the now-dead
+      `is_required_baseline_app()`, updated/renamed the test that had locked
+      in the old behavior. Compiles clean; test suite run in progress.
+      indeedhub itself not yet manually recovered on .116 — the code fix
+      will self-heal it on the next reconcile tick once deployed there.
 - [ ] **Phase-3 Quadlet default-flip** — code is validated + opt-in via
       `ARCHIPELAGO_USE_QUADLET_BACKENDS=true` on .228/.198 already (confirmed live
       2026-07-01). Ready to flip (`config.rs:256` + its test) the moment the .5 gate