frank-small: --attention-backend flag (env var removed in this build)

The VLLM_ATTENTION_BACKEND env var does not exist in this vLLM dev
build (warns 'Unknown vLLM environment variable' and is ignored), so
the pin landed nowhere and Gemma4's hook force-selected TRITON_ATTN
again. Pass --attention-backend TURBOQUANT as a CLI flag instead.

tests/frank_small_unit.bats

@@ -113,10 +113,12 @@ QUADLET_LARGE="$REPO_ROOT/dankclaw/frank-large/quadlet/frank-large.container"
     # (head_dim 256) and full-attention (head_dim 512) layers, and fp8 is
     # hardware-dead on sm_86 anyway.
     [ "$kvdtype" = "turboquant_k8v4" ]
-    # The TURBOQUANT backend must be pinned explicitly: Gemma4's config
-    # hook force-selects TRITON_ATTN (mixed 256/512 head dims) whenever no
-    # backend is selected, and TRITON_ATTN rejects the turboquant dtype.
-    grep -q '^VLLM_ATTENTION_BACKEND=TURBOQUANT$' "$ENV_FILE"
+    # The TURBOQUANT backend must be pinned via --attention-backend:
+    # Gemma4's config hook force-selects TRITON_ATTN (mixed 256/512 head
+    # dims) whenever no backend is selected, and TRITON_ATTN rejects the
+    # turboquant dtype. This vLLM build has no attention-backend env var.
+    grep -q '^ATTENTION_BACKEND=TURBOQUANT$' "$ENV_FILE"
+    grep -q -- '--attention-backend "$ATTENTION_BACKEND"' "$SERVE"
     # The 3090 is dedicated to this service, so the engine may take most of
     # it; the KV pool, not the seq ceiling, caps full-context concurrency.
     [ "$util" = "0.95" ]