Summary

The model receives the prompt (and optional system message). The run uses scorer contains_all with the JSON configuration below. Pass/fail and partial credit are determined entirely by that scorer against the model output; no human grading.

Summarize in exactly 3 short bullet points: Frontend delivery got faster after introducing component libraries, visual regression tests, and bundle-size budgets in CI. Teams tracked web vitals each sprint.

{
  "expected_contains": [
    "component libraries",
    "visual regression",
    "bundle-size"
  ]
}