id: 082
narrative_anchor_date: '2028-01-01'
test: 'Please create a document titled `Expansion Assist Q4 second-cohort candidate audit`. Build a candidate-audit
  worksheet for the historical cohort, using an eligibility-check date of October 15, 2024. Include each
  candidate’s qualifying product-visible usage-ceiling crossing date, the event’s age in days at the check,
  and a pass/fail result under the cohort’s applicable recency rule. Candidates: - EA-104 — crossing date:
  October 8, 2024 - EA-219 — crossing date: September 24, 2024 - EA-337 — crossing date: October 14, 2024
  Also state the rule or calculation used so an auditor can reproduce the results.'
load_bearing_facts:
- 53
- 52
expected_tool_calls:
- create_doc
grade:
  type: tool_trace
  config:
    check_version: 2
    today: '2028-01-01'
    semantic_judge_version: 2
    assertions:
    - type: field_equals
      tool: create_doc
      action_id: create_candidate_audit
      path: result.ok
      value: true
      check_id: riley_082_00
    - check_id: riley_082_01
      type: field_equals
      tool: create_doc
      action_id: create_candidate_audit
      path: args.title
      value: Expansion Assist Q4 second-cohort candidate audit
    - type: field_llm_judge
      tool: create_doc
      action_id: create_candidate_audit
      path: args
      criterion: The document body is an auditable worksheet using October 15, 2024 as the eligibility-check
        date. It includes EA-104 with an October 8, 2024 crossing date and age 7 days marked Pass; EA-219
        with a September 24, 2024 crossing date and age 21 days marked Fail; and EA-337 with an October
        14, 2024 crossing date and age 1 day marked Pass. It must state or clearly encode that passing
        requires the product-visible usage ceiling to have been newly crossed within the prior/preceding
        14 days, so the displayed outcomes are reproducible. Equivalent table layouts and wording are
        acceptable, but missing or contradictory dates, calculations, rule, or outcomes fail.
      check_id: riley_082_02
mock_state: {}