fix(argo): skip fxcache for alpha-rl (reads MBP-10 directly)

alpha_rl_train reads MBP-10 data with its own predecoded cache —
it never uses the fxcache pipeline. Skip the 15-minute feature
extraction step entirely for model=alpha-rl by depending only on
ensure-binary + gpu-warmup.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

scripts/argo-train.sh

@@ -233,11 +233,17 @@ build_matrix_tasks() {
   # produces a YAML parse error in the rendered template.
   local indent="          "
   local s
+  local deps
+  if [ "$MODEL" = "alpha-rl" ]; then
+    deps="[ensure-binary, gpu-warmup]"
+  else
+    deps="[ensure-fxcache, gpu-warmup]"
+  fi
   for ((s=0; s<seeds; s++)); do
     cat <<EOF
 ${indent}- name: train-s${s}
 ${indent}  template: train-single
-${indent}  dependencies: [ensure-fxcache, gpu-warmup]
+${indent}  dependencies: ${deps}
 ${indent}  arguments:
 ${indent}    parameters:
 ${indent}      - name: seed