diff --git a/clusters/atlas/flux-system/applications/hermes/kustomization.yaml b/clusters/atlas/flux-system/applications/hermes/kustomization.yaml
index f29ee6c0a..57fed46e4 100644
--- a/clusters/atlas/flux-system/applications/hermes/kustomization.yaml
+++ b/clusters/atlas/flux-system/applications/hermes/kustomization.yaml
@@ -26,6 +26,10 @@ spec:
       kind: Deployment
       name: hermes-switchyard
       namespace: hermes
+    - apiVersion: apps/v1
+      kind: Deployment
+      name: hermes-agent
+      namespace: hermes
     - apiVersion: apps/v1
       kind: Deployment
       name: hermes
diff --git a/services/hermes/agent-deployment.yaml b/services/hermes/agent-deployment.yaml
index 5a6b167cd..b1215abc3 100644
--- a/services/hermes/agent-deployment.yaml
+++ b/services/hermes/agent-deployment.yaml
@@ -25,7 +25,7 @@ spec:
         ai.bstein.dev/execution: Hermes Kanban with durable direct Codex and Claude Code CLI workers
         ai.bstein.dev/model-policy: Jetson-assisted AUTO routing, low through xhigh, cross-provider fallback
         ai.bstein.dev/placement: rpi5 preferred; Jetson deferred until state storage is available
-        ai.bstein.dev/config-rev: "20260813-kanban-degraded-startup"
+        ai.bstein.dev/config-rev: "20260813-kanban-corruption-isolation"
         vault.hashicorp.com/agent-inject: "true"
         vault.hashicorp.com/role: hermes-agent
         vault.hashicorp.com/agent-inject-secret-anthropic-token: kv/data/atlas/hermes/agent-tokens
diff --git a/services/hermes/chat-router.yaml b/services/hermes/chat-router.yaml
index b3e6c9fd2..0b244a9c8 100644
--- a/services/hermes/chat-router.yaml
+++ b/services/hermes/chat-router.yaml
@@ -20,7 +20,7 @@ spec:
         app: hermes-chat-router
       annotations:
         ai.bstein.dev/role: privacy-preserving-chat-tenant-router
-        ai.bstein.dev/config-rev: "20260813-telegram-operator-v2"
+        ai.bstein.dev/config-rev: "20260813-telegram-activated-v1"
         vault.hashicorp.com/agent-inject: "true"
         vault.hashicorp.com/agent-pre-populate-only: "true"
         vault.hashicorp.com/agent-init-first: "true"
diff --git a/services/hermes/scripts/cli_lane_runner.py b/services/hermes/scripts/cli_lane_runner.py
index 13a98b666..5faa23510 100644
--- a/services/hermes/scripts/cli_lane_runner.py
+++ b/services/hermes/scripts/cli_lane_runner.py
@@ -37,6 +37,7 @@ DEFAULT_CLAIM_TTL = 7 * 24 * 60 * 60
 DEFAULT_MAX_RUNTIME = 12 * 60 * 60
 HEARTBEAT_SECONDS = 20
 WORKTREE_LOCK = threading.Lock()
+BOARD_CORRUPTION_ERRORS: dict[str, str] = {}
 CAPACITY_PATTERN = re.compile(
     r"(?:rate.?limit|capacity|overload|usage.?limit|quota|credit|exhaust|429|529|authentication|oauth|token.*expired)",
     re.I,
@@ -741,6 +742,24 @@ def _external(task: Any) -> bool:
     return str(_task_value(task, "assignee", "") or "").startswith(EXTERNAL_PREFIX)
 
 
+def _connect_healthy_board(kanban_db: Any, board: str) -> Any | None:
+    """Open one board without letting localized corruption stop other lanes."""
+    try:
+        conn = kanban_db.connect(board=board)
+    except getattr(kanban_db, "KanbanDbCorruptError", ()) as error:
+        detail = str(error)
+        if BOARD_CORRUPTION_ERRORS.get(board) != detail:
+            print(
+                f"quarantining corrupt Kanban board {board!r}: {detail}",
+                file=sys.stderr,
+                flush=True,
+            )
+            BOARD_CORRUPTION_ERRORS[board] = detail
+        return None
+    BOARD_CORRUPTION_ERRORS.pop(board, None)
+    return conn
+
+
 def recover_orphans() -> None:
     """Return external running tasks to ready after a runner/pod restart."""
     from hermes_cli import kanban_db
@@ -750,7 +769,9 @@ def recover_orphans() -> None:
         if not board:
             continue
         with kanban_db.scoped_current_board(board):
-            conn = kanban_db.connect(board=board)
+            conn = _connect_healthy_board(kanban_db, board)
+            if conn is None:
+                continue
             try:
                 for task in kanban_db.list_tasks(conn):
                     if _external(task) and str(_task_value(task, "status", "")) == "running":
@@ -775,7 +796,9 @@ def claim_ready(active: set[tuple[str, str]], limit: int) -> list[tuple[str, str
         if not board:
             continue
         with kanban_db.scoped_current_board(board):
-            conn = kanban_db.connect(board=board)
+            conn = _connect_healthy_board(kanban_db, board)
+            if conn is None:
+                continue
             try:
                 kanban_db.recompute_ready(conn)
                 for task in kanban_db.list_tasks(conn):
diff --git a/testing/tests/test_hermes_cli_lanes.py b/testing/tests/test_hermes_cli_lanes.py
index 02fe17d78..08c5742b5 100644
--- a/testing/tests/test_hermes_cli_lanes.py
+++ b/testing/tests/test_hermes_cli_lanes.py
@@ -276,6 +276,40 @@ def test_unassigned_ready_task_is_persistently_routed_to_auto_lane(monkeypatch):
     assert assigned == [("t_auto", "cli-auto")]
 
 
+def test_corrupt_board_is_quarantined_without_stopping_healthy_lanes(monkeypatch, capsys):
+    class CorruptBoardError(Exception):
+        pass
+
+    task = SimpleNamespace(id="t_healthy", assignee="cli-auto", status="ready")
+
+    class Connection:
+        def close(self):
+            return None
+
+    def connect(*, board):
+        if board == "cassandra":
+            raise CorruptBoardError("integrity_check failed")
+        return Connection()
+
+    fake_db = SimpleNamespace(
+        KanbanDbCorruptError=CorruptBoardError,
+        list_boards=lambda include_archived=False: [
+            {"slug": "cassandra"},
+            {"slug": "healthy"},
+        ],
+        scoped_current_board=lambda _board: nullcontext(),
+        connect=connect,
+        recompute_ready=lambda _conn: None,
+        list_tasks=lambda _conn: [task],
+        claim_task=lambda _conn, _task_id, **_kwargs: task,
+    )
+    monkeypatch.setitem(sys.modules, "hermes_cli", SimpleNamespace(kanban_db=fake_db))
+    lanes.BOARD_CORRUPTION_ERRORS.clear()
+
+    assert lanes.claim_ready(set(), 1) == [("healthy", "t_healthy")]
+    assert "quarantining corrupt Kanban board 'cassandra'" in capsys.readouterr().err
+
+
 @pytest.mark.parametrize(
     ("result", "expected_action"),
     [