Unblock CI and address mxinO review on remote-execution.md

Two fixes:

1. Exclude vendored upstream skills from markdownlint.

   `.claude/skills/launching-evals/` and `.claude/skills/accessing-mlflow/`
   are vendored verbatim from NVIDIA-NeMo/Evaluator and re-synced via
   .claude/scripts/sync-upstream-skills.sh. Markdownlint wanted to
   reformat them (trailing blank lines, spacing around fences), but
   fixing would violate the "verbatim" property documented in their
   frontmatter. Add an `ignores:` glob to `.markdownlint-cli2.yaml`.

2. Reframe the checkpoint/storage note on `skills/common/remote-execution.md`.

   Reviewer @mxinO noted (PR #1239) that the previous "compute nodes may
   not share the same filesystem as login nodes" framing is misleading —
   compute nodes on a given cluster do share storage with the login
   node. The real issue is that workstation filesystems aren't mounted
   on the cluster at all. Also drops the dlcluster-specific row, which
   @mxinO flagged as an internal quirk that shouldn't ship publicly.

Signed-off-by: Zhiyu Cheng <zhiyuc@nvidia.com>

Author: Edwardf0t1

.claude/skills/common/remote-execution.md

@@ -28,16 +28,15 @@ clusters:
 default_cluster: my-cluster
 ```
 
-### Checkpoint and storage availability
+### Staging checkpoints from your workstation
 
-Cluster compute nodes may not share the same filesystem as login nodes or other clusters. Before running any workload that references a checkpoint path, verify the path is accessible from compute nodes:
+Workstation filesystems (`/home/scratch.*`, local NFS) are **not** mounted on the cluster. If a checkpoint was produced on your workstation, copy it to the cluster's own storage before submitting any job that references it — NEL and SLURM do NOT sync checkpoints automatically.
 
-| Cluster type | Compute-node storage | NOT accessible from compute nodes |
-|-------------|---------------------|----------------------------------|
-| JET clusters (oci-hsg, cw, oci-nrt) | `/lustre/fsw/...` | Workstation NFS (`/home/scratch.*`), other cluster mounts |
-| dlcluster | `/home/omniml_data_*`, `/home/scratch.*` | `/lustre/` paths |
+```bash
+rsync -av /path/to/local/checkpoint <cluster-login>:<cluster-workspace>/checkpoints/
+```
 
-If a checkpoint was produced on a different cluster or workstation, copy it to the target cluster's accessible storage before submitting jobs. NEL and SLURM do NOT sync checkpoints automatically.
+Use the `workspace` path from your cluster config as the destination. Compute nodes on a given cluster share the same storage as its login node, so once staged, the path works everywhere on that cluster.
 
 See `.claude/clusters.yaml.example` for a fully annotated example with multiple cluster types.
 

.markdownlint-cli2.yaml

@@ -10,3 +10,9 @@ config:
   MD036: false # no-emphasis-as-heading - allow **bold** as section markers
   MD041: false # first-line-heading
   MD059: false # no-hard-tabs
+
+# Vendored upstream skills — kept byte-identical to upstream via
+# .claude/scripts/sync-upstream-skills.sh; do not reformat.
+ignores:
+  - ".claude/skills/launching-evals/**"
+  - ".claude/skills/accessing-mlflow/**"