fix(install): restore DGX Station GB300 express setup

docs/inference/set-up-vllm.mdx

@@ -92,40 +92,42 @@ The first run can take 10 to 30 minutes.
 Later runs reuse the cached image and model weights.
 The managed container uses Docker's `--restart unless-stopped` policy, so Docker restarts it unless an operator explicitly stops it.
 
-### Check Storage Before Downloads
+### Check Storage Before the Image Pull
 
-For an uncached image, NemoClaw checks Docker image storage before the watched image pull starts.
-It checks the Hugging Face model cache before that pull and checks it again afterward before `hf download` starts.
-The image estimate includes compressed content, unpacked layers, and pull-staging headroom.
-The model estimate includes the model's full published size plus temporary-write headroom because a local cache cannot prove which files a mutable remote revision will reuse.
+For an uncached pinned image, NemoClaw checks Docker image storage before the watched image pull starts.
+The estimate reserves three times the image's compressed registry size plus `3 GiB` for compressed content, unpacked layers, and pull staging.
 
 For Docker's classic image store, NemoClaw checks the absolute `DockerRootDir` reported by `docker info`.
 For Docker's containerd image store, it checks the top-level `root` from `/etc/containerd/config.toml`, or `/var/lib/containerd` when that configuration file is absent, and also checks `DockerRootDir` for pull staging.
 The check uses the location with the least available capacity.
-The model-cache check uses `~/.cache/huggingface`, or its nearest existing parent filesystem when the cache directory does not exist yet.
 
 NemoClaw recognizes an image cache hit only when Docker contains the exact platform digest.
-A cache hit skips the cold-pull capacity check, but the explicit watched `docker pull` still runs against that immutable digest.
-Before trusting the host model-cache capacity, NemoClaw also verifies that the selected Docker daemon maps bind mounts to the local host; a remote or unverifiable Docker host makes the model-cache check inconclusive.
-An explicit Unix socket is trusted only when it is Docker's default `/var/run/docker.sock`; any other socket can forward to a daemon with a different filesystem and is treated as unverifiable.
-Known container markers are rejected immediately.
-Before trusting client-side image-store capacity, NemoClaw requires the selected default Docker socket peer to have a visible PID and share the client's mount namespace; a nested PID namespace or unverifiable peer fails closed before the image pull.
-Before trusting model-cache capacity, NemoClaw also asks the pinned cached vLLM image to read a random mode-`0600` sentinel through a bounded, read-only Docker bind mount.
-The pull is idempotent for the cached digest and cannot select different layers from a moved release tag.
-The model-download helper and the long-lived vLLM container use `--pull=never`, so an implicit pull cannot bypass the storage gate.
-The second model-cache check detects capacity consumed by the image pull before the model download command can start.
-
-When capacity is insufficient or NemoClaw cannot verify it, the warning identifies the image or model, available capacity, approximate requirement, checked path and source, remediation, and diagnostic commands.
+A cache hit skips the capacity check, but the explicit watched `docker pull` still runs against that immutable digest.
+The model-download helper and the long-lived vLLM container use `--pull=never`, so an implicit pull cannot select another image.
+
+NemoClaw treats `/var/run/docker.sock` and `/run/docker.sock` as the local default Docker socket paths.
+It cannot inspect daemon storage through a remote endpoint, a named non-default Docker context, or another Unix socket.
+When both `DOCKER_CONTEXT` and `DOCKER_HOST` are set, NemoClaw follows Docker CLI precedence and uses `DOCKER_CONTEXT`.
+When the check is inconclusive, NemoClaw prints the reason, approximate image requirement, and diagnostic commands, then continues with the pull without claiming that capacity is sufficient.
+
+When verified capacity is insufficient, the warning identifies the image, available capacity, approximate requirement, checked path and source, remediation, and diagnostic commands.
 Interactive setup continues only when you enter `y` or `yes` explicitly.
-Pressing **Enter**, entering `n` or `no`, or entering any other response stops before the guarded pull or model download starts.
+Pressing **Enter**, entering `n` or `no`, or entering any other response stops before the guarded image pull starts.
 
 <Warning>
-`--yes` and `NEMOCLAW_YES=1` do not accept managed-vLLM storage warnings.
+`--yes` and `NEMOCLAW_YES=1` do not accept a verified insufficient-storage warning.
 Non-interactive setup stops unless you set `NEMOCLAW_IGNORE_VLLM_DISK_SPACE=1` exactly.
-This override accepts both insufficient-capacity and inconclusive checks for Docker image storage and the model cache.
-Use it only after you independently verify that both storage locations can complete the downloads.
+The override applies only when NemoClaw verifies that Docker image storage is insufficient.
+Use it only after you independently verify that Docker storage can complete the image pull.
+Inconclusive checks continue automatically and do not require the override.
 </Warning>
 
+<Note>
+This preflight checks only the pinned vLLM container image.
+It does not estimate or gate the Hugging Face model-weight download into `~/.cache/huggingface`.
+Before the first managed-vLLM run, verify that the filesystem backing your model cache has enough free space for the selected model.
+</Note>
+
 If image pull output stops making progress, a watchdog stops the stalled pull without imposing a fixed wall-clock limit on slow but active downloads.
 If vLLM does not become ready, NemoClaw prints a short tail of the container logs before exiting.
 

docs/reference/commands.mdx

@@ -3250,7 +3250,7 @@ Set them before running `$$nemoclaw onboard`.
 | `NEMOCLAW_INSTALL_TAG` | release tag | For internal installer commands: the release tag to install. Defaults to the admin-promoted `lkg` tag when unset. Overridden by the `--install-tag` flag. |
 | `NEMOCLAW_VLLM_MODEL` | registry slug or Hugging Face model id | Selects the model the managed-vLLM install path serves. Recognised slugs: `qwen3.6-27b`, `qwen3.6-35b-a3b-nvfp4`, `nemotron-3-nano-4b`, `deepseek-v4-flash`, `deepseek-r1-distill-70b`. Unset uses the per-platform profile default. Gated models (e.g. `deepseek-r1-distill-70b`) require `HF_TOKEN` or `HUGGING_FACE_HUB_TOKEN`. |
 | `NEMOCLAW_VLLM_EXTRA_ARGS_JSON` | JSON array of non-blank strings | Appends advanced operator-owned tokens to the managed `vllm serve` command after NemoClaw's registry defaults. Example: `["--max-num-seqs","2"]`. Malformed JSON, non-string tokens, or blank tokens fail before Docker work starts. |
-| `NEMOCLAW_IGNORE_VLLM_DISK_SPACE` | exactly `1` | Allows managed vLLM to continue when Docker image-storage or model-cache capacity is insufficient or cannot be verified; other values do not bypass the checks. |
+| `NEMOCLAW_IGNORE_VLLM_DISK_SPACE` | exactly `1` | Allows managed vLLM to continue when verified Docker image-storage capacity is insufficient; other values do not bypass the check. Inconclusive image-storage checks report diagnostics and continue automatically. The check does not cover Hugging Face model-cache capacity. |
 <AgentOnly variant="openclaw">
 | `NEMOCLAW_MINIMAL_BOOTSTRAP` | `1` to enable | Skips default OpenClaw workspace-template seeding for new pristine workspaces. Existing files are not deleted; refer to [Understand Runtime Changes](../manage-sandboxes/configure-sandboxes/understand-runtime-changes). |
 </AgentOnly>

scripts/install.sh

@@ -2655,7 +2655,7 @@ detect_express_platform() {
   fi
   case "$model" in
     *DGX*Spark*) printf "DGX Spark" ;;
-    *DGX*Station*) printf "DGX Station" ;;
+    *DGX*Station* | *Station*GB300*) printf "DGX Station" ;;
     *) ;;
   esac
 }

src/lib/inference/vllm-models.test.ts

@@ -16,25 +16,6 @@ import {
 } from "./vllm-models";
 
 describe("vllm model registry", () => {
-  it("records a finite positive Hugging Face download size for every model", () => {
-    for (const model of VLLM_MODELS) {
-      expect(Number.isFinite(model.downloadSizeBytes)).toBe(true);
-      expect(model.downloadSizeBytes).toBeGreaterThan(0);
-    }
-  });
-
-  it("pins the official Hugging Face repository totals", () => {
-    expect(
-      Object.fromEntries(VLLM_MODELS.map((model) => [model.envValue, model.downloadSizeBytes])),
-    ).toEqual({
-      "qwen3.6-27b": 30_900_000_000,
-      "deepseek-r1-distill-70b": 141_000_000_000,
-      "nemotron-3-nano-4b": 5_280_000_000,
-      "deepseek-v4-flash": 160_000_000_000,
-      "qwen3.6-35b-a3b-nvfp4": 23_500_000_000,
-    });
-  });
-
   it("returns null when NEMOCLAW_VLLM_MODEL is unset so the caller can fall back to the profile default", () => {
     expect(selectVllmModelFromEnv({} as NodeJS.ProcessEnv)).toBeNull();
   });

src/lib/inference/vllm-models.ts

@@ -35,8 +35,6 @@ export interface VllmModelDef {
   label: string;
   /** Stable identifier accepted via `NEMOCLAW_VLLM_MODEL`. */
   envValue: string;
-  /** Approximate full Hugging Face repository download size in bytes. */
-  downloadSizeBytes: number;
   /** `--max-model-len` flag value. */
   maxModelLen: number;
   /** Model-specific flags appended after the shared serving flags. */
@@ -65,7 +63,6 @@ export const VLLM_MODELS: readonly VllmModelDef[] = [
     id: "Qwen/Qwen3.6-27B-FP8",
     label: "Qwen3.6 27B FP8",
     envValue: "qwen3.6-27b",
-    downloadSizeBytes: 30_900_000_000,
     maxModelLen: 262144,
     modelArgs: [
       "--gpu-memory-utilization",
@@ -88,7 +85,6 @@ export const VLLM_MODELS: readonly VllmModelDef[] = [
     id: "deepseek-ai/DeepSeek-R1-Distill-Llama-70B",
     label: "DeepSeek-R1 Distill Llama 70B",
     envValue: "deepseek-r1-distill-70b",
-    downloadSizeBytes: 141_000_000_000,
     maxModelLen: 32768,
     modelArgs: [
       "--gpu-memory-utilization",
@@ -108,7 +104,6 @@ export const VLLM_MODELS: readonly VllmModelDef[] = [
     id: "nvidia/NVIDIA-Nemotron-3-Nano-4B-FP8",
     label: "NVIDIA Nemotron-3 Nano 4B FP8",
     envValue: "nemotron-3-nano-4b",
-    downloadSizeBytes: 5_280_000_000,
     // Matches the model card's `max_position_embeddings` and the vLLM
     // example NVIDIA publishes for this checkpoint. The previous value
     // (262000) was an undocumented round-down with no headroom rationale.
@@ -138,7 +133,6 @@ export const VLLM_MODELS: readonly VllmModelDef[] = [
     id: "deepseek-ai/DeepSeek-V4-Flash",
     label: "DeepSeek V4 Flash",
     envValue: "deepseek-v4-flash",
-    downloadSizeBytes: 160_000_000_000,
     maxModelLen: 1048576,
     modelArgs: [
       "--kv-cache-dtype",
@@ -179,7 +173,6 @@ export const VLLM_MODELS: readonly VllmModelDef[] = [
     id: "nvidia/Qwen3.6-35B-A3B-NVFP4",
     label: "Qwen3.6 35B-A3B NVFP4",
     envValue: "qwen3.6-35b-a3b-nvfp4",
-    downloadSizeBytes: 23_500_000_000,
     maxModelLen: 262144,
     // Additive flags on top of the shared serving defaults. The shared flags
     // already cover --tensor-parallel-size/--pipeline-parallel-size/