ci(robotwin): add metrics parsing and upload steps

Match the libero/metaworld CI pattern: parse eval metrics with
parse_eval_metrics.py and upload both video and metrics.json artifacts.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: pkooij

.github/workflows/benchmark_tests.yml

@@ -279,10 +279,27 @@ jobs:
           docker cp robotwin-eval:/tmp/eval-artifacts/. /tmp/robotwin-artifacts/ 2>/dev/null || true
           docker rm -f robotwin-eval || true
 
+      - name: Parse RoboTwin eval metrics
+        if: always()
+        run: |
+          python3 scripts/ci/parse_eval_metrics.py \
+            --artifacts-dir /tmp/robotwin-artifacts \
+            --env robotwin \
+            --task beat_block_hammer \
+            --policy pepijn223/smolvla_robotwin
+
       - name: Upload RoboTwin rollout video
         if: always()
         uses: actions/upload-artifact@v4
         with:
           name: robotwin-rollout-video
           path: /tmp/robotwin-artifacts/videos/
           if-no-files-found: warn
+
+      - name: Upload RoboTwin eval metrics
+        if: always()
+        uses: actions/upload-artifact@v4
+        with:
+          name: robotwin-metrics
+          path: /tmp/robotwin-artifacts/metrics.json
+          if-no-files-found: warn