the job asks for its runner the way the fleet answers

runs-on was a bare scalar. Jobs written that way queue and are never
picked up — link and quest have waited since 10 August, painter since the
5th, none of them with a single success — while repos writing the same
label as a sequence drain within seconds of a push on the same fleet.

Same label, same runners, different shape. It fails silently: nothing errors,
the job simply never starts, so the lane looks configured and does nothing.

.hanzo/workflows/ci.yml

@@ -31,7 +31,7 @@ on:
 
 jobs:
   go:
-    runs-on: hanzo-build-linux-amd64
+    runs-on: [hanzo-build-linux-amd64]
     timeout-minutes: 30
     steps:
       - uses: actions/checkout@v4
@@ -57,7 +57,7 @@ jobs:
       - run: go test ./...
 
   python:
-    runs-on: hanzo-build-linux-amd64
+    runs-on: [hanzo-build-linux-amd64]
     timeout-minutes: 30
     steps:
       - uses: actions/checkout@v4

.hanzo/workflows/deploy.yml

@@ -41,7 +41,7 @@ jobs:
     # The forge's git-runner advertises this label (GIT_RUNNER_LABELS in
     # universe infra/k8s/git-runner/statefulset.yaml). A label nothing
     # advertises does not fail — it queues to the timeout, which is silence.
-    runs-on: hanzo-build-linux-amd64
+    runs-on: [hanzo-build-linux-amd64]
     timeout-minutes: 45
     steps:
       - uses: actions/checkout@v4