feat: inline HF model validation at submit time (replaces dead background thread)

Single HTTP HEAD to huggingface.co/api/models/{repo} before sbatch.
Shows green checkmark or yellow warning. Takes <1s, never blocks on failure.

Author: alec-flowers

src/srtctl/cli/submit.py

@@ -391,6 +391,16 @@ def submit_with_orchestrator(
         f.write(script_content)
     os.chmod(script_path, 0o755)
 
+    # Pre-submit identity validation (inline, <1s)
+    if hasattr(config, "identity") and config.identity and config.identity.model and config.identity.model.repo:
+        from srtctl.core.validation import validate_hf_model
+
+        hf_result = validate_hf_model(config.identity.model.repo, config.identity.model.revision)
+        if hf_result.ok:
+            console.print(f"[green]✓[/] HF model: {hf_result.message}")
+        else:
+            console.print(f"[yellow]⚠ HF model: {hf_result.message}[/]")
+
     console.print(f"[bold cyan]🚀 Submitting:[/] {config.name}")
     logging.debug(f"Script: {script_path}")