{
  "schemaVersion": 9,
  "generatedAt": "2026-09-28T16:49:14.280Z",
  "publicationScope": "completed-results",
  "activeExperiment": "test-05",
  "scopeNote": "Five distinct test scopes, including HTML-to-managed-app fidelity. Original, repeat, follow-up and HTML-reference cohorts remain separate. Compare settings against the same scope and demonstrated capabilities, not pooled scores or averages.",
  "app": "Device Procurement and Availability; Device Lifecycle Hub; Device Allocation and Policy Hub; Annual Leave Manager",
  "unitNote": "Observed Cowork task credits, not currency or a tenant invoice. Final cumulative readings are not added to initial readings.",
  "timingNote": "Observed build time per app sums the authoring turns: build, planned updates, clarifications and corrections. It excludes external review and screenshot-taking gaps between prompts, /cost commands and native self-review. Within-turn waits and observation uncertainty remain included; this is not pure model compute time. Missing authoring coverage is not estimated from elapsed test time.",
  "evidenceNote": "Exact prompts, visible transcripts, approval records, backend snapshots and raw cost evidence are retained privately. Test 5 separately publishes its authorized HTML reference and selected final app-only screenshots. Public data remains an explicit allowlisted projection, not a raw transcript export.",
  "experiments": [
    {
      "id": "test-03",
      "route": "test-03",
      "title": "Test 3: complex multi-prompt virtual-data apps",
      "design": "Five model/effort configurations once each, four identical ordered authoring prompts in each fresh app: build, connect the existing app to Office 365 Users, extend workflows, then change requirements. Independent external review.",
      "runs": [
        {
          "id": "run-01",
          "model": "Sonnet 5",
          "effort": "Medium",
          "replicate": 1,
          "block": null,
          "status": "reviewed",
          "comparisonEligible": true,
          "final": {
            "points": 90,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": [
              "CX14"
            ]
          },
          "firstCost": {
            "status": "observed",
            "value": 176,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 1326,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 1150,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": 2371.299,
              "lowerBoundSeconds": 2358.672,
              "observation": "interval-bounded",
              "authoringPrompts": 5,
              "observedPrompts": 5
            },
            "firstBuildResponseSeconds": 533.775,
            "firstBuildResponseLowerBoundSeconds": 531.625,
            "firstCompletionObservation": "interval-bounded",
            "finalReviewedResultSeconds": 7606.6,
            "timeToAcceptanceSeconds": 7606.6,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": 2.15,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 0,
          "stages": [
            {
              "id": "stage-01",
              "title": "Build the relational virtual-data app first",
              "kind": "build",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 176,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": null,
              "increaseSincePrecedingReading": null
            },
            {
              "id": "stage-02",
              "title": "Connect the existing managed app to Office 365 Users",
              "kind": "planned-change",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 367,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": {
                "status": "observed",
                "value": 251,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "increaseSincePrecedingReading": {
                "value": 116,
                "unit": "credits"
              }
            },
            {
              "id": "stage-03",
              "title": "Virtual approvals, asset lifecycle, themes and preferences",
              "kind": "planned-change",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 770,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": {
                "status": "observed",
                "value": 367,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "increaseSincePrecedingReading": {
                "value": 403,
                "unit": "credits"
              }
            },
            {
              "id": "stage-04",
              "title": "Approval policy change, revisions, cost centres and responsive refinement",
              "kind": "planned-change",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 1326,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": {
                "status": "observed",
                "value": 770,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "increaseSincePrecedingReading": {
                "value": 556,
                "unit": "credits"
              }
            }
          ],
          "firstFullTargetAssessmentStage": "stage-04",
          "plannedChanges": 3,
          "profileProof": {
            "stageId": "stage-02",
            "verifiedInApp": true,
            "reverifiedAfterFinalReopen": true,
            "errorRetryCoverage": "blocked"
          },
          "corrections": 1,
          "clarifications": 0,
          "approvals": 0,
          "screenshotFiles": 379,
          "archived": true,
          "promptSha256": "3f6463aff82d0b6a5bcfd20f0aae18e9cbe385b90c5f25f2b62c418055232ced",
          "checks": [
            {
              "id": "CX01",
              "capability": "Same private app opens with all seven navigable screens",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX02",
              "capability": "Exact six-device/four-request relational seed and derived initial metrics",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX03",
              "capability": "Catalogue search, combined category/stock filters and zero-stock protection",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX04",
              "capability": "Multi-line draft creation, line edits/removal and accurate GBP totals",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX05",
              "capability": "Required fields, dates, whole positive quantities, duplicates and stock validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX06",
              "capability": "Draft/persona isolation, edits and exact-once submission with stable IDs",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX07",
              "capability": "Controller-created draft, IDs, lines and timeline survive all three planned changes and reopen",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX08",
              "capability": "Virtual manager/finance role gates and approval transitions at the final threshold",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX09",
              "capability": "Rejection comments, append-only actor/timestamp history and terminal action guards",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX10",
              "capability": "Request changes, requester edit/resubmit and cancellation preserve identity/history",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX11",
              "capability": "Atomic fulfilment, stock revalidation, per-unit assets and duplicate-fulfilment protection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX12",
              "capability": "Good/damaged asset returns and duplicate-return protection",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX13",
              "capability": "Current-user Office 365 Users profile actually reads through supported app binding",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX14",
              "capability": "Profile loading/error/retry and genuinely missing optional fields are honest",
              "critical": false,
              "final": "blocked"
            },
            {
              "id": "CX15",
              "capability": "Explicit light/dark mode applies across all screens and persists on reopen",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX16",
              "capability": "Local preference Save/Cancel, density and default department work without directory writes",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX17",
              "capability": "Cost-centre migration, persistence and department mismatch validation",
              "critical": false,
              "final": "fail"
            },
            {
              "id": "CX18",
              "capability": "Derived status/asset/stock metrics, low-stock navigation and combined request filters",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX19",
              "capability": "All screens and key forms are usable at actual wide desktop and 390px mobile widths",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX20",
              "capability": "Labels, visible keyboard focus, readable contrast and reachable form errors",
              "critical": false,
              "final": "pass"
            }
          ]
        },
        {
          "id": "run-02",
          "model": "Opus 5",
          "effort": "High",
          "replicate": 1,
          "block": null,
          "status": "reviewed",
          "comparisonEligible": true,
          "final": {
            "points": 95,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": [
              "CX14"
            ]
          },
          "firstCost": {
            "status": "observed",
            "value": 561,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 3073,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 2512,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": 2763.457,
              "lowerBoundSeconds": 2752.088,
              "observation": "interval-bounded",
              "authoringPrompts": 5,
              "observedPrompts": 5
            },
            "firstBuildResponseSeconds": 881.455,
            "firstBuildResponseLowerBoundSeconds": 879.256,
            "firstCompletionObservation": "interval-bounded",
            "finalReviewedResultSeconds": 4851.881,
            "timeToAcceptanceSeconds": 4851.881,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": 2.199,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 0,
          "stages": [
            {
              "id": "stage-01",
              "title": "Build the relational virtual-data app first",
              "kind": "build",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 561,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": null,
              "increaseSincePrecedingReading": null
            },
            {
              "id": "stage-02",
              "title": "Connect the existing managed app to Office 365 Users",
              "kind": "planned-change",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 916,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": {
                "status": "observed",
                "value": 561,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "increaseSincePrecedingReading": {
                "value": 355,
                "unit": "credits"
              }
            },
            {
              "id": "stage-03",
              "title": "Virtual approvals, asset lifecycle, themes and preferences",
              "kind": "planned-change",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 1679,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": {
                "status": "observed",
                "value": 916,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "increaseSincePrecedingReading": {
                "value": 763,
                "unit": "credits"
              }
            },
            {
              "id": "stage-04",
              "title": "Approval policy change, revisions, cost centres and responsive refinement",
              "kind": "planned-change",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 2505,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": {
                "status": "observed",
                "value": 1679,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "increaseSincePrecedingReading": {
                "value": 826,
                "unit": "credits"
              }
            }
          ],
          "firstFullTargetAssessmentStage": "stage-04",
          "plannedChanges": 3,
          "profileProof": {
            "stageId": "stage-02",
            "verifiedInApp": true,
            "reverifiedAfterFinalReopen": true,
            "errorRetryCoverage": "blocked"
          },
          "corrections": 1,
          "clarifications": 0,
          "approvals": 0,
          "screenshotFiles": 340,
          "archived": true,
          "promptSha256": "3f6463aff82d0b6a5bcfd20f0aae18e9cbe385b90c5f25f2b62c418055232ced",
          "checks": [
            {
              "id": "CX01",
              "capability": "Same private app opens with all seven navigable screens",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX02",
              "capability": "Exact six-device/four-request relational seed and derived initial metrics",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX03",
              "capability": "Catalogue search, combined category/stock filters and zero-stock protection",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX04",
              "capability": "Multi-line draft creation, line edits/removal and accurate GBP totals",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX05",
              "capability": "Required fields, dates, whole positive quantities, duplicates and stock validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX06",
              "capability": "Draft/persona isolation, edits and exact-once submission with stable IDs",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX07",
              "capability": "Controller-created draft, IDs, lines and timeline survive all three planned changes and reopen",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX08",
              "capability": "Virtual manager/finance role gates and approval transitions at the final threshold",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX09",
              "capability": "Rejection comments, append-only actor/timestamp history and terminal action guards",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX10",
              "capability": "Request changes, requester edit/resubmit and cancellation preserve identity/history",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX11",
              "capability": "Atomic fulfilment, stock revalidation, per-unit assets and duplicate-fulfilment protection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX12",
              "capability": "Good/damaged asset returns and duplicate-return protection",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX13",
              "capability": "Current-user Office 365 Users profile actually reads through supported app binding",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX14",
              "capability": "Profile loading/error/retry and genuinely missing optional fields are honest",
              "critical": false,
              "final": "blocked"
            },
            {
              "id": "CX15",
              "capability": "Explicit light/dark mode applies across all screens and persists on reopen",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX16",
              "capability": "Local preference Save/Cancel, density and default department work without directory writes",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX17",
              "capability": "Cost-centre migration, persistence and department mismatch validation",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX18",
              "capability": "Derived status/asset/stock metrics, low-stock navigation and combined request filters",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX19",
              "capability": "All screens and key forms are usable at actual wide desktop and 390px mobile widths",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX20",
              "capability": "Labels, visible keyboard focus, readable contrast and reachable form errors",
              "critical": false,
              "final": "pass"
            }
          ]
        },
        {
          "id": "run-03",
          "model": "Opus 5",
          "effort": "Medium",
          "replicate": 1,
          "block": null,
          "status": "reviewed",
          "comparisonEligible": true,
          "final": {
            "points": 95,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": [
              "CX14"
            ]
          },
          "firstCost": {
            "status": "observed",
            "value": 505,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 2358,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 1853,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": 2111.675,
              "lowerBoundSeconds": 2097.083,
              "observation": "interval-bounded",
              "authoringPrompts": 6,
              "observedPrompts": 6
            },
            "firstBuildResponseSeconds": 644.642,
            "firstBuildResponseLowerBoundSeconds": 642.371,
            "firstCompletionObservation": "interval-bounded",
            "finalReviewedResultSeconds": 5105.311,
            "timeToAcceptanceSeconds": 5105.311,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": 2.271,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 0,
          "stages": [
            {
              "id": "stage-01",
              "title": "Build the relational virtual-data app first",
              "kind": "build",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 505,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": null,
              "increaseSincePrecedingReading": null
            },
            {
              "id": "stage-02",
              "title": "Connect the existing managed app to Office 365 Users",
              "kind": "planned-change",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 1124,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": {
                "status": "observed",
                "value": 726,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "increaseSincePrecedingReading": {
                "value": 398,
                "unit": "credits"
              }
            },
            {
              "id": "stage-03",
              "title": "Virtual approvals, asset lifecycle, themes and preferences",
              "kind": "planned-change",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 1720,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": {
                "status": "observed",
                "value": 1124,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "increaseSincePrecedingReading": {
                "value": 596,
                "unit": "credits"
              }
            },
            {
              "id": "stage-04",
              "title": "Approval policy change, revisions, cost centres and responsive refinement",
              "kind": "planned-change",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 2011,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": {
                "status": "observed",
                "value": 1720,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "increaseSincePrecedingReading": {
                "value": 291,
                "unit": "credits"
              }
            }
          ],
          "firstFullTargetAssessmentStage": "stage-04",
          "plannedChanges": 3,
          "profileProof": {
            "stageId": "stage-02",
            "verifiedInApp": true,
            "reverifiedAfterFinalReopen": true,
            "errorRetryCoverage": "blocked"
          },
          "corrections": 2,
          "clarifications": 0,
          "approvals": 0,
          "screenshotFiles": 441,
          "archived": true,
          "promptSha256": "3f6463aff82d0b6a5bcfd20f0aae18e9cbe385b90c5f25f2b62c418055232ced",
          "checks": [
            {
              "id": "CX01",
              "capability": "Same private app opens with all seven navigable screens",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX02",
              "capability": "Exact six-device/four-request relational seed and derived initial metrics",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX03",
              "capability": "Catalogue search, combined category/stock filters and zero-stock protection",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX04",
              "capability": "Multi-line draft creation, line edits/removal and accurate GBP totals",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX05",
              "capability": "Required fields, dates, whole positive quantities, duplicates and stock validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX06",
              "capability": "Draft/persona isolation, edits and exact-once submission with stable IDs",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX07",
              "capability": "Controller-created draft, IDs, lines and timeline survive all three planned changes and reopen",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX08",
              "capability": "Virtual manager/finance role gates and approval transitions at the final threshold",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX09",
              "capability": "Rejection comments, append-only actor/timestamp history and terminal action guards",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX10",
              "capability": "Request changes, requester edit/resubmit and cancellation preserve identity/history",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX11",
              "capability": "Atomic fulfilment, stock revalidation, per-unit assets and duplicate-fulfilment protection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX12",
              "capability": "Good/damaged asset returns and duplicate-return protection",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX13",
              "capability": "Current-user Office 365 Users profile actually reads through supported app binding",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX14",
              "capability": "Profile loading/error/retry and genuinely missing optional fields are honest",
              "critical": false,
              "final": "blocked"
            },
            {
              "id": "CX15",
              "capability": "Explicit light/dark mode applies across all screens and persists on reopen",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX16",
              "capability": "Local preference Save/Cancel, density and default department work without directory writes",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX17",
              "capability": "Cost-centre migration, persistence and department mismatch validation",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX18",
              "capability": "Derived status/asset/stock metrics, low-stock navigation and combined request filters",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX19",
              "capability": "All screens and key forms are usable at actual wide desktop and 390px mobile widths",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX20",
              "capability": "Labels, visible keyboard focus, readable contrast and reachable form errors",
              "critical": false,
              "final": "pass"
            }
          ]
        },
        {
          "id": "run-04",
          "model": "Sonnet 5",
          "effort": "High",
          "replicate": 1,
          "block": null,
          "status": "reviewed",
          "comparisonEligible": true,
          "final": {
            "points": 90,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": [
              "CX14",
              "CX16"
            ]
          },
          "firstCost": {
            "status": "observed",
            "value": 292,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 2289,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 1997,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": 3089.817,
              "lowerBoundSeconds": 3081.297,
              "observation": "interval-bounded",
              "authoringPrompts": 4,
              "observedPrompts": 4
            },
            "firstBuildResponseSeconds": 666.448,
            "firstBuildResponseLowerBoundSeconds": 664.215,
            "firstCompletionObservation": "interval-bounded",
            "finalReviewedResultSeconds": 5098.76,
            "timeToAcceptanceSeconds": 5098.76,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": 2.233,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 0,
          "stages": [
            {
              "id": "stage-01",
              "title": "Build the relational virtual-data app first",
              "kind": "build",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 292,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": null,
              "increaseSincePrecedingReading": null
            },
            {
              "id": "stage-02",
              "title": "Connect the existing managed app to Office 365 Users",
              "kind": "planned-change",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 550,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": {
                "status": "observed",
                "value": 292,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "increaseSincePrecedingReading": {
                "value": 258,
                "unit": "credits"
              }
            },
            {
              "id": "stage-03",
              "title": "Virtual approvals, asset lifecycle, themes and preferences",
              "kind": "planned-change",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 1507,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": {
                "status": "observed",
                "value": 550,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "increaseSincePrecedingReading": {
                "value": 957,
                "unit": "credits"
              }
            },
            {
              "id": "stage-04",
              "title": "Approval policy change, revisions, cost centres and responsive refinement",
              "kind": "planned-change",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 2289,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": {
                "status": "observed",
                "value": 1507,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "increaseSincePrecedingReading": {
                "value": 782,
                "unit": "credits"
              }
            }
          ],
          "firstFullTargetAssessmentStage": "stage-04",
          "plannedChanges": 3,
          "profileProof": {
            "stageId": "stage-02",
            "verifiedInApp": true,
            "reverifiedAfterFinalReopen": true,
            "errorRetryCoverage": "blocked"
          },
          "corrections": 0,
          "clarifications": 0,
          "approvals": 0,
          "screenshotFiles": 356,
          "archived": true,
          "promptSha256": "3f6463aff82d0b6a5bcfd20f0aae18e9cbe385b90c5f25f2b62c418055232ced",
          "checks": [
            {
              "id": "CX01",
              "capability": "Same private app opens with all seven navigable screens",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX02",
              "capability": "Exact six-device/four-request relational seed and derived initial metrics",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX03",
              "capability": "Catalogue search, combined category/stock filters and zero-stock protection",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX04",
              "capability": "Multi-line draft creation, line edits/removal and accurate GBP totals",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX05",
              "capability": "Required fields, dates, whole positive quantities, duplicates and stock validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX06",
              "capability": "Draft/persona isolation, edits and exact-once submission with stable IDs",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX07",
              "capability": "Controller-created draft, IDs, lines and timeline survive all three planned changes and reopen",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX08",
              "capability": "Virtual manager/finance role gates and approval transitions at the final threshold",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX09",
              "capability": "Rejection comments, append-only actor/timestamp history and terminal action guards",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX10",
              "capability": "Request changes, requester edit/resubmit and cancellation preserve identity/history",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX11",
              "capability": "Atomic fulfilment, stock revalidation, per-unit assets and duplicate-fulfilment protection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX12",
              "capability": "Good/damaged asset returns and duplicate-return protection",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX13",
              "capability": "Current-user Office 365 Users profile actually reads through supported app binding",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX14",
              "capability": "Profile loading/error/retry and genuinely missing optional fields are honest",
              "critical": false,
              "final": "blocked"
            },
            {
              "id": "CX15",
              "capability": "Explicit light/dark mode applies across all screens and persists on reopen",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX16",
              "capability": "Local preference Save/Cancel, density and default department work without directory writes",
              "critical": false,
              "final": "blocked"
            },
            {
              "id": "CX17",
              "capability": "Cost-centre migration, persistence and department mismatch validation",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX18",
              "capability": "Derived status/asset/stock metrics, low-stock navigation and combined request filters",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX19",
              "capability": "All screens and key forms are usable at actual wide desktop and 390px mobile widths",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX20",
              "capability": "Labels, visible keyboard focus, readable contrast and reachable form errors",
              "critical": false,
              "final": "pass"
            }
          ]
        },
        {
          "id": "run-05",
          "model": "GPT 5.5",
          "effort": "Medium",
          "replicate": 1,
          "block": null,
          "status": "reviewed",
          "comparisonEligible": true,
          "final": {
            "points": 90,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": [
              "CX14",
              "CX16"
            ]
          },
          "firstCost": {
            "status": "observed",
            "value": 381,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 944,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 563,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": 1303.39,
              "lowerBoundSeconds": 1291.207,
              "observation": "interval-bounded",
              "authoringPrompts": 5,
              "observedPrompts": 5
            },
            "firstBuildResponseSeconds": 161.688,
            "firstBuildResponseLowerBoundSeconds": 159.638,
            "firstCompletionObservation": "interval-bounded",
            "finalReviewedResultSeconds": 3486.634,
            "timeToAcceptanceSeconds": 3486.634,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": 2.05,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 0,
          "stages": [
            {
              "id": "stage-01",
              "title": "Build the relational virtual-data app first",
              "kind": "build",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 381,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": null,
              "increaseSincePrecedingReading": null
            },
            {
              "id": "stage-02",
              "title": "Connect the existing managed app to Office 365 Users",
              "kind": "planned-change",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 501,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": {
                "status": "observed",
                "value": 381,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "increaseSincePrecedingReading": {
                "value": 120,
                "unit": "credits"
              }
            },
            {
              "id": "stage-03",
              "title": "Virtual approvals, asset lifecycle, themes and preferences",
              "kind": "planned-change",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 736,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": {
                "status": "observed",
                "value": 501,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "increaseSincePrecedingReading": {
                "value": 235,
                "unit": "credits"
              }
            },
            {
              "id": "stage-04",
              "title": "Approval policy change, revisions, cost centres and responsive refinement",
              "kind": "planned-change",
              "submitted": true,
              "responseComplete": true,
              "cost": {
                "status": "observed",
                "value": 944,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "costBeforePrompt": {
                "status": "observed",
                "value": 736,
                "unit": "credits",
                "scope": "Cowork task"
              },
              "increaseSincePrecedingReading": {
                "value": 208,
                "unit": "credits"
              }
            }
          ],
          "firstFullTargetAssessmentStage": "stage-04",
          "plannedChanges": 3,
          "profileProof": {
            "stageId": "stage-02",
            "verifiedInApp": true,
            "reverifiedAfterFinalReopen": true,
            "errorRetryCoverage": "blocked"
          },
          "corrections": 0,
          "clarifications": 1,
          "approvals": 0,
          "screenshotFiles": 306,
          "archived": true,
          "promptSha256": "3f6463aff82d0b6a5bcfd20f0aae18e9cbe385b90c5f25f2b62c418055232ced",
          "checks": [
            {
              "id": "CX01",
              "capability": "Same private app opens with all seven navigable screens",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX02",
              "capability": "Exact six-device/four-request relational seed and derived initial metrics",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX03",
              "capability": "Catalogue search, combined category/stock filters and zero-stock protection",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX04",
              "capability": "Multi-line draft creation, line edits/removal and accurate GBP totals",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX05",
              "capability": "Required fields, dates, whole positive quantities, duplicates and stock validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX06",
              "capability": "Draft/persona isolation, edits and exact-once submission with stable IDs",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX07",
              "capability": "Controller-created draft, IDs, lines and timeline survive all three planned changes and reopen",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX08",
              "capability": "Virtual manager/finance role gates and approval transitions at the final threshold",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX09",
              "capability": "Rejection comments, append-only actor/timestamp history and terminal action guards",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX10",
              "capability": "Request changes, requester edit/resubmit and cancellation preserve identity/history",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX11",
              "capability": "Atomic fulfilment, stock revalidation, per-unit assets and duplicate-fulfilment protection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX12",
              "capability": "Good/damaged asset returns and duplicate-return protection",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX13",
              "capability": "Current-user Office 365 Users profile actually reads through supported app binding",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX14",
              "capability": "Profile loading/error/retry and genuinely missing optional fields are honest",
              "critical": false,
              "final": "blocked"
            },
            {
              "id": "CX15",
              "capability": "Explicit light/dark mode applies across all screens and persists on reopen",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX16",
              "capability": "Local preference Save/Cancel, density and default department work without directory writes",
              "critical": false,
              "final": "blocked"
            },
            {
              "id": "CX17",
              "capability": "Cost-centre migration, persistence and department mismatch validation",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX18",
              "capability": "Derived status/asset/stock metrics, low-stock navigation and combined request filters",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "CX19",
              "capability": "All screens and key forms are usable at actual wide desktop and 390px mobile widths",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "CX20",
              "capability": "Labels, visible keyboard focus, readable contrast and reachable form errors",
              "critical": false,
              "final": "pass"
            }
          ]
        }
      ],
      "comparisonEligible": true,
      "reviewMode": "external-only",
      "acceptanceCheckCount": 20,
      "stages": [
        {
          "id": "stage-01",
          "title": "Build the relational virtual-data app first",
          "kind": "build"
        },
        {
          "id": "stage-02",
          "title": "Connect the existing managed app to Office 365 Users",
          "kind": "planned-change"
        },
        {
          "id": "stage-03",
          "title": "Virtual approvals, asset lifecycle, themes and preferences",
          "kind": "planned-change"
        },
        {
          "id": "stage-04",
          "title": "Approval policy change, revisions, cost centres and responsive refinement",
          "kind": "planned-change"
        }
      ],
      "stageNote": "Build first; connect Office 365 Users through the existing app next. Planned changes are not repairs. First credits are the initial virtual build; the first full-capability assessment follows stage 4. Each stage card is cumulative. Increases use the last observed task reading before that prompt, including a preceding repair reading where present; they are not isolated connector or token prices.",
      "protocolSha256": "df5feaf15af92906b23619d11aafbfe6a0415eb9e57c75c819007b48471e61cf"
    },
    {
      "id": "test-02-sharepoint-clean",
      "route": "test-02",
      "title": "Test 2: 10 SharePoint-connected builds",
      "design": "Five model/effort configurations, each repeated twice in reverse-order blocks; ten fresh tasks and apps; controller-only review.",
      "runs": [
        {
          "id": "run-01",
          "model": "Sonnet 5",
          "effort": "Medium",
          "replicate": 1,
          "block": 1,
          "status": "reviewed",
          "comparisonEligible": true,
          "final": {
            "points": 90,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": [
              "BC09"
            ]
          },
          "firstCost": {
            "status": "observed",
            "value": 224,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 224,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 0,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": 600.06,
              "lowerBoundSeconds": 597.865,
              "observation": "interval-bounded",
              "authoringPrompts": 1,
              "observedPrompts": 1
            },
            "firstBuildResponseSeconds": 600.06,
            "firstBuildResponseLowerBoundSeconds": 597.865,
            "firstCompletionObservation": "interval-bounded",
            "finalReviewedResultSeconds": 1409.916,
            "timeToAcceptanceSeconds": 1409.916,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": 2.195,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 0,
          "corrections": 0,
          "clarifications": 0,
          "approvals": 0,
          "screenshotFiles": 87,
          "archived": true,
          "promptSha256": "71281644f6f8465bac8bcde4029fec509e09ea7c18cf2cd6dd6011b6586dad53",
          "checks": [
            {
              "id": "BC01",
              "capability": "Runnable private managed app",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC02",
              "capability": "Live catalogue connection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC03",
              "capability": "Live dashboard arithmetic",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC04",
              "capability": "Catalogue search and filtering",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC05",
              "capability": "Complete saved request",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC06",
              "capability": "Basic request validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC07",
              "capability": "Live request listing",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC08",
              "capability": "Reopen persistence",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC09",
              "capability": "Honest data and error states",
              "critical": false,
              "final": "blocked"
            },
            {
              "id": "BC10",
              "capability": "Usable desktop and mobile build",
              "critical": false,
              "final": "pass"
            }
          ]
        },
        {
          "id": "run-02",
          "model": "Opus 5",
          "effort": "High",
          "replicate": 1,
          "block": 1,
          "status": "reviewed",
          "comparisonEligible": true,
          "final": {
            "points": 90,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": [
              "BC09"
            ]
          },
          "firstCost": {
            "status": "observed",
            "value": 968,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 968,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 0,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": 1165.748,
              "lowerBoundSeconds": 1163.623,
              "observation": "interval-bounded",
              "authoringPrompts": 1,
              "observedPrompts": 1
            },
            "firstBuildResponseSeconds": 1165.748,
            "firstBuildResponseLowerBoundSeconds": 1163.623,
            "firstCompletionObservation": "interval-bounded",
            "finalReviewedResultSeconds": 1869.782,
            "timeToAcceptanceSeconds": 1869.782,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": 2.125,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 0,
          "corrections": 0,
          "clarifications": 0,
          "approvals": 0,
          "screenshotFiles": 112,
          "archived": true,
          "promptSha256": "71281644f6f8465bac8bcde4029fec509e09ea7c18cf2cd6dd6011b6586dad53",
          "checks": [
            {
              "id": "BC01",
              "capability": "Runnable private managed app",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC02",
              "capability": "Live catalogue connection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC03",
              "capability": "Live dashboard arithmetic",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC04",
              "capability": "Catalogue search and filtering",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC05",
              "capability": "Complete saved request",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC06",
              "capability": "Basic request validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC07",
              "capability": "Live request listing",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC08",
              "capability": "Reopen persistence",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC09",
              "capability": "Honest data and error states",
              "critical": false,
              "final": "blocked"
            },
            {
              "id": "BC10",
              "capability": "Usable desktop and mobile build",
              "critical": false,
              "final": "pass"
            }
          ]
        },
        {
          "id": "run-03",
          "model": "Opus 5",
          "effort": "Medium",
          "replicate": 1,
          "block": 1,
          "status": "reviewed",
          "comparisonEligible": true,
          "final": {
            "points": 90,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": [
              "BC09"
            ]
          },
          "firstCost": {
            "status": "observed",
            "value": 417,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 417,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 0,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": 528.115,
              "lowerBoundSeconds": 526.019,
              "observation": "interval-bounded",
              "authoringPrompts": 1,
              "observedPrompts": 1
            },
            "firstBuildResponseSeconds": 528.115,
            "firstBuildResponseLowerBoundSeconds": 526.019,
            "firstCompletionObservation": "interval-bounded",
            "finalReviewedResultSeconds": 1027.525,
            "timeToAcceptanceSeconds": 1027.525,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": 2.096,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 0,
          "corrections": 0,
          "clarifications": 0,
          "approvals": 0,
          "screenshotFiles": 83,
          "archived": true,
          "promptSha256": "71281644f6f8465bac8bcde4029fec509e09ea7c18cf2cd6dd6011b6586dad53",
          "checks": [
            {
              "id": "BC01",
              "capability": "Runnable private managed app",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC02",
              "capability": "Live catalogue connection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC03",
              "capability": "Live dashboard arithmetic",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC04",
              "capability": "Catalogue search and filtering",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC05",
              "capability": "Complete saved request",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC06",
              "capability": "Basic request validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC07",
              "capability": "Live request listing",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC08",
              "capability": "Reopen persistence",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC09",
              "capability": "Honest data and error states",
              "critical": false,
              "final": "blocked"
            },
            {
              "id": "BC10",
              "capability": "Usable desktop and mobile build",
              "critical": false,
              "final": "pass"
            }
          ]
        },
        {
          "id": "run-04",
          "model": "Sonnet 5",
          "effort": "High",
          "replicate": 1,
          "block": 1,
          "status": "reviewed",
          "comparisonEligible": true,
          "final": {
            "points": 90,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": [
              "BC09"
            ]
          },
          "firstCost": {
            "status": "observed",
            "value": 620,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 620,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 0,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": 1018.305,
              "lowerBoundSeconds": 1016.281,
              "observation": "interval-bounded",
              "authoringPrompts": 1,
              "observedPrompts": 1
            },
            "firstBuildResponseSeconds": 1018.305,
            "firstBuildResponseLowerBoundSeconds": 1016.281,
            "firstCompletionObservation": "interval-bounded",
            "finalReviewedResultSeconds": 1897.37,
            "timeToAcceptanceSeconds": 1897.37,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": 2.024,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 0,
          "corrections": 0,
          "clarifications": 0,
          "approvals": 0,
          "screenshotFiles": 111,
          "archived": true,
          "promptSha256": "71281644f6f8465bac8bcde4029fec509e09ea7c18cf2cd6dd6011b6586dad53",
          "checks": [
            {
              "id": "BC01",
              "capability": "Runnable private managed app",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC02",
              "capability": "Live catalogue connection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC03",
              "capability": "Live dashboard arithmetic",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC04",
              "capability": "Catalogue search and filtering",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC05",
              "capability": "Complete saved request",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC06",
              "capability": "Basic request validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC07",
              "capability": "Live request listing",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC08",
              "capability": "Reopen persistence",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC09",
              "capability": "Honest data and error states",
              "critical": false,
              "final": "blocked"
            },
            {
              "id": "BC10",
              "capability": "Usable desktop and mobile build",
              "critical": false,
              "final": "pass"
            }
          ]
        },
        {
          "id": "run-05",
          "model": "GPT 5.5",
          "effort": "Medium",
          "replicate": 1,
          "block": 1,
          "status": "reviewed",
          "comparisonEligible": true,
          "final": {
            "points": 90,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": [
              "BC09"
            ]
          },
          "firstCost": {
            "status": "observed",
            "value": 247,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 345,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 98,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": 631.581,
              "lowerBoundSeconds": 626.594,
              "observation": "interval-bounded",
              "authoringPrompts": 2,
              "observedPrompts": 2
            },
            "firstBuildResponseSeconds": 489.542,
            "firstBuildResponseLowerBoundSeconds": 486.655,
            "firstCompletionObservation": "interval-bounded",
            "finalReviewedResultSeconds": 1253.959,
            "timeToAcceptanceSeconds": 1253.959,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": 2.887,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 0,
          "corrections": 1,
          "clarifications": 0,
          "approvals": 0,
          "screenshotFiles": 121,
          "archived": true,
          "promptSha256": "71281644f6f8465bac8bcde4029fec509e09ea7c18cf2cd6dd6011b6586dad53",
          "checks": [
            {
              "id": "BC01",
              "capability": "Runnable private managed app",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC02",
              "capability": "Live catalogue connection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC03",
              "capability": "Live dashboard arithmetic",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC04",
              "capability": "Catalogue search and filtering",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC05",
              "capability": "Complete saved request",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC06",
              "capability": "Basic request validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC07",
              "capability": "Live request listing",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC08",
              "capability": "Reopen persistence",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC09",
              "capability": "Honest data and error states",
              "critical": false,
              "final": "blocked"
            },
            {
              "id": "BC10",
              "capability": "Usable desktop and mobile build",
              "critical": false,
              "final": "pass"
            }
          ]
        },
        {
          "id": "run-06",
          "model": "GPT 5.5",
          "effort": "Medium",
          "replicate": 2,
          "block": 2,
          "status": "reviewed",
          "comparisonEligible": true,
          "final": {
            "points": 90,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": [
              "BC09"
            ]
          },
          "firstCost": {
            "status": "observed",
            "value": 252,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 252,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 0,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": 471.478,
              "lowerBoundSeconds": 469.131,
              "observation": "interval-bounded",
              "authoringPrompts": 1,
              "observedPrompts": 1
            },
            "firstBuildResponseSeconds": 471.478,
            "firstBuildResponseLowerBoundSeconds": 469.131,
            "firstCompletionObservation": "interval-bounded",
            "finalReviewedResultSeconds": 1091.294,
            "timeToAcceptanceSeconds": 1091.294,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": 2.347,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 0,
          "corrections": 0,
          "clarifications": 0,
          "approvals": 0,
          "screenshotFiles": 79,
          "archived": true,
          "promptSha256": "71281644f6f8465bac8bcde4029fec509e09ea7c18cf2cd6dd6011b6586dad53",
          "checks": [
            {
              "id": "BC01",
              "capability": "Runnable private managed app",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC02",
              "capability": "Live catalogue connection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC03",
              "capability": "Live dashboard arithmetic",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC04",
              "capability": "Catalogue search and filtering",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC05",
              "capability": "Complete saved request",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC06",
              "capability": "Basic request validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC07",
              "capability": "Live request listing",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC08",
              "capability": "Reopen persistence",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC09",
              "capability": "Honest data and error states",
              "critical": false,
              "final": "blocked"
            },
            {
              "id": "BC10",
              "capability": "Usable desktop and mobile build",
              "critical": false,
              "final": "pass"
            }
          ]
        },
        {
          "id": "run-07",
          "model": "Sonnet 5",
          "effort": "High",
          "replicate": 2,
          "block": 2,
          "status": "reviewed",
          "comparisonEligible": true,
          "final": {
            "points": 90,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": [
              "BC09"
            ]
          },
          "firstCost": {
            "status": "observed",
            "value": 313,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 313,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 0,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": 686.485,
              "lowerBoundSeconds": 684.417,
              "observation": "interval-bounded",
              "authoringPrompts": 1,
              "observedPrompts": 1
            },
            "firstBuildResponseSeconds": 686.485,
            "firstBuildResponseLowerBoundSeconds": 684.417,
            "firstCompletionObservation": "interval-bounded",
            "finalReviewedResultSeconds": 1067.872,
            "timeToAcceptanceSeconds": 1067.872,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": 2.068,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 0,
          "corrections": 0,
          "clarifications": 0,
          "approvals": 0,
          "screenshotFiles": 95,
          "archived": true,
          "promptSha256": "71281644f6f8465bac8bcde4029fec509e09ea7c18cf2cd6dd6011b6586dad53",
          "checks": [
            {
              "id": "BC01",
              "capability": "Runnable private managed app",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC02",
              "capability": "Live catalogue connection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC03",
              "capability": "Live dashboard arithmetic",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC04",
              "capability": "Catalogue search and filtering",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC05",
              "capability": "Complete saved request",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC06",
              "capability": "Basic request validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC07",
              "capability": "Live request listing",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC08",
              "capability": "Reopen persistence",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC09",
              "capability": "Honest data and error states",
              "critical": false,
              "final": "blocked"
            },
            {
              "id": "BC10",
              "capability": "Usable desktop and mobile build",
              "critical": false,
              "final": "pass"
            }
          ]
        },
        {
          "id": "run-08",
          "model": "Opus 5",
          "effort": "Medium",
          "replicate": 2,
          "block": 2,
          "status": "reviewed",
          "comparisonEligible": true,
          "final": {
            "points": 90,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": [
              "BC09"
            ]
          },
          "firstCost": {
            "status": "observed",
            "value": 425,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 425,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 0,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": 624.126,
              "lowerBoundSeconds": 621.597,
              "observation": "interval-bounded",
              "authoringPrompts": 1,
              "observedPrompts": 1
            },
            "firstBuildResponseSeconds": 624.126,
            "firstBuildResponseLowerBoundSeconds": 621.597,
            "firstCompletionObservation": "interval-bounded",
            "finalReviewedResultSeconds": 910.55,
            "timeToAcceptanceSeconds": 910.55,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": 2.529,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 0,
          "corrections": 0,
          "clarifications": 0,
          "approvals": 0,
          "screenshotFiles": 88,
          "archived": true,
          "promptSha256": "71281644f6f8465bac8bcde4029fec509e09ea7c18cf2cd6dd6011b6586dad53",
          "checks": [
            {
              "id": "BC01",
              "capability": "Runnable private managed app",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC02",
              "capability": "Live catalogue connection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC03",
              "capability": "Live dashboard arithmetic",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC04",
              "capability": "Catalogue search and filtering",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC05",
              "capability": "Complete saved request",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC06",
              "capability": "Basic request validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC07",
              "capability": "Live request listing",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC08",
              "capability": "Reopen persistence",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC09",
              "capability": "Honest data and error states",
              "critical": false,
              "final": "blocked"
            },
            {
              "id": "BC10",
              "capability": "Usable desktop and mobile build",
              "critical": false,
              "final": "pass"
            }
          ]
        },
        {
          "id": "run-09",
          "model": "Opus 5",
          "effort": "High",
          "replicate": 2,
          "block": 2,
          "status": "reviewed",
          "comparisonEligible": true,
          "final": {
            "points": 90,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": [
              "BC09"
            ]
          },
          "firstCost": {
            "status": "observed",
            "value": 478,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 786,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 308,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": 928.287,
              "lowerBoundSeconds": 923.679,
              "observation": "interval-bounded",
              "authoringPrompts": 2,
              "observedPrompts": 2
            },
            "firstBuildResponseSeconds": 687.571,
            "firstBuildResponseLowerBoundSeconds": 685.371,
            "firstCompletionObservation": "interval-bounded",
            "finalReviewedResultSeconds": 1929.991,
            "timeToAcceptanceSeconds": 1929.991,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": 2.2,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 0,
          "corrections": 1,
          "clarifications": 0,
          "approvals": 0,
          "screenshotFiles": 142,
          "archived": true,
          "promptSha256": "71281644f6f8465bac8bcde4029fec509e09ea7c18cf2cd6dd6011b6586dad53",
          "checks": [
            {
              "id": "BC01",
              "capability": "Runnable private managed app",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC02",
              "capability": "Live catalogue connection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC03",
              "capability": "Live dashboard arithmetic",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC04",
              "capability": "Catalogue search and filtering",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC05",
              "capability": "Complete saved request",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC06",
              "capability": "Basic request validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC07",
              "capability": "Live request listing",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC08",
              "capability": "Reopen persistence",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC09",
              "capability": "Honest data and error states",
              "critical": false,
              "final": "blocked"
            },
            {
              "id": "BC10",
              "capability": "Usable desktop and mobile build",
              "critical": false,
              "final": "pass"
            }
          ]
        },
        {
          "id": "run-10",
          "model": "Sonnet 5",
          "effort": "Medium",
          "replicate": 2,
          "block": 2,
          "status": "reviewed",
          "comparisonEligible": true,
          "final": {
            "points": 90,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": [
              "BC09"
            ]
          },
          "firstCost": {
            "status": "observed",
            "value": 230,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 230,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 0,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": 526.239,
              "lowerBoundSeconds": 524.184,
              "observation": "interval-bounded",
              "authoringPrompts": 1,
              "observedPrompts": 1
            },
            "firstBuildResponseSeconds": 526.239,
            "firstBuildResponseLowerBoundSeconds": 524.184,
            "firstCompletionObservation": "interval-bounded",
            "finalReviewedResultSeconds": 857.488,
            "timeToAcceptanceSeconds": 857.488,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": 2.055,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 0,
          "corrections": 0,
          "clarifications": 0,
          "approvals": 0,
          "screenshotFiles": 88,
          "archived": true,
          "promptSha256": "71281644f6f8465bac8bcde4029fec509e09ea7c18cf2cd6dd6011b6586dad53",
          "checks": [
            {
              "id": "BC01",
              "capability": "Runnable private managed app",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC02",
              "capability": "Live catalogue connection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC03",
              "capability": "Live dashboard arithmetic",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC04",
              "capability": "Catalogue search and filtering",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC05",
              "capability": "Complete saved request",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC06",
              "capability": "Basic request validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC07",
              "capability": "Live request listing",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "BC08",
              "capability": "Reopen persistence",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "BC09",
              "capability": "Honest data and error states",
              "critical": false,
              "final": "blocked"
            },
            {
              "id": "BC10",
              "capability": "Usable desktop and mobile build",
              "critical": false,
              "final": "pass"
            }
          ]
        }
      ],
      "comparisonEligible": true,
      "reviewMode": "external-only",
      "acceptanceCheckCount": 10,
      "protocolSha256": "838a5c94f5e306fd30860a7e1b3fa4482a07fc31cecbc73a43269e4723e98acd"
    },
    {
      "id": "test-01",
      "route": "test-01",
      "title": "Test 1: virtual-data apps",
      "design": "Five virtual-data builds: two models at two reasoning levels, with a baseline repeat. Costs include build and native review.",
      "runs": [
        {
          "id": "run-01",
          "model": "Sonnet 5",
          "effort": "Medium",
          "replicate": 1,
          "block": null,
          "status": "reviewed",
          "comparisonEligible": false,
          "final": {
            "points": 95,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": []
          },
          "firstCost": {
            "status": "observed",
            "value": 276,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 317,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 41,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": null,
              "lowerBoundSeconds": null,
              "observation": "not-captured",
              "authoringPrompts": 1,
              "observedPrompts": 0
            },
            "firstBuildResponseSeconds": null,
            "firstBuildResponseLowerBoundSeconds": null,
            "firstCompletionObservation": "not-captured",
            "finalReviewedResultSeconds": null,
            "timeToAcceptanceSeconds": null,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": null,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 1,
          "corrections": 0,
          "clarifications": 0,
          "approvals": 1,
          "screenshotFiles": 126,
          "archived": true,
          "promptSha256": "d0466119660c3b0fd7d25117d94fdd3c7db60302635650e7c00ff51dac9187e9",
          "checks": [
            {
              "id": "AC01",
              "capability": "Exact catalogue",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC02",
              "capability": "Search and empty state",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC03",
              "capability": "Combined filtering",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC04",
              "capability": "Unavailable stock",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC05",
              "capability": "Dashboard arithmetic",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC06",
              "capability": "Valid request",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC07",
              "capability": "Required fields",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC08",
              "capability": "Quantity validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC09",
              "capability": "Date validation",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC10",
              "capability": "Approval",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC11",
              "capability": "Rejection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC12",
              "capability": "Oversell prevention",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC13",
              "capability": "Single state transition",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC14",
              "capability": "Managed persistence",
              "critical": false,
              "final": "not-required"
            },
            {
              "id": "AC15",
              "capability": "Requester scoping",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC16",
              "capability": "Safe repeatable reset",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC17",
              "capability": "Keyboard and labels",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC18",
              "capability": "Responsive layout",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC19",
              "capability": "Honest demo boundary",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC20",
              "capability": "Actual private managed app",
              "critical": true,
              "final": "pass"
            }
          ]
        },
        {
          "id": "run-02",
          "model": "Opus 5",
          "effort": "High",
          "replicate": 1,
          "block": null,
          "status": "reviewed",
          "comparisonEligible": false,
          "final": {
            "points": 95,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": []
          },
          "firstCost": {
            "status": "observed",
            "value": 621,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 892,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 271,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": null,
              "lowerBoundSeconds": null,
              "observation": "not-captured",
              "authoringPrompts": 2,
              "observedPrompts": 0
            },
            "firstBuildResponseSeconds": null,
            "firstBuildResponseLowerBoundSeconds": null,
            "firstCompletionObservation": "not-captured",
            "finalReviewedResultSeconds": null,
            "timeToAcceptanceSeconds": null,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": null,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 1,
          "corrections": 1,
          "clarifications": 0,
          "approvals": 0,
          "screenshotFiles": 151,
          "archived": true,
          "promptSha256": "d0466119660c3b0fd7d25117d94fdd3c7db60302635650e7c00ff51dac9187e9",
          "checks": [
            {
              "id": "AC01",
              "capability": "Exact catalogue",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC02",
              "capability": "Search and empty state",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC03",
              "capability": "Combined filtering",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC04",
              "capability": "Unavailable stock",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC05",
              "capability": "Dashboard arithmetic",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC06",
              "capability": "Valid request",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC07",
              "capability": "Required fields",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC08",
              "capability": "Quantity validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC09",
              "capability": "Date validation",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC10",
              "capability": "Approval",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC11",
              "capability": "Rejection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC12",
              "capability": "Oversell prevention",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC13",
              "capability": "Single state transition",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC14",
              "capability": "Managed persistence",
              "critical": false,
              "final": "not-required"
            },
            {
              "id": "AC15",
              "capability": "Requester scoping",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC16",
              "capability": "Safe repeatable reset",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC17",
              "capability": "Keyboard and labels",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC18",
              "capability": "Responsive layout",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC19",
              "capability": "Honest demo boundary",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC20",
              "capability": "Actual private managed app",
              "critical": true,
              "final": "pass"
            }
          ]
        },
        {
          "id": "run-03",
          "model": "Opus 5",
          "effort": "Medium",
          "replicate": 1,
          "block": null,
          "status": "reviewed",
          "comparisonEligible": false,
          "final": {
            "points": 95,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": []
          },
          "firstCost": {
            "status": "observed",
            "value": 342,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 1083,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 741,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": null,
              "lowerBoundSeconds": null,
              "observation": "not-captured",
              "authoringPrompts": 3,
              "observedPrompts": 0
            },
            "firstBuildResponseSeconds": null,
            "firstBuildResponseLowerBoundSeconds": null,
            "firstCompletionObservation": "not-captured",
            "finalReviewedResultSeconds": null,
            "timeToAcceptanceSeconds": null,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": null,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 1,
          "corrections": 2,
          "clarifications": 0,
          "approvals": 1,
          "screenshotFiles": 98,
          "archived": true,
          "promptSha256": "d0466119660c3b0fd7d25117d94fdd3c7db60302635650e7c00ff51dac9187e9",
          "checks": [
            {
              "id": "AC01",
              "capability": "Exact catalogue",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC02",
              "capability": "Search and empty state",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC03",
              "capability": "Combined filtering",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC04",
              "capability": "Unavailable stock",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC05",
              "capability": "Dashboard arithmetic",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC06",
              "capability": "Valid request",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC07",
              "capability": "Required fields",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC08",
              "capability": "Quantity validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC09",
              "capability": "Date validation",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC10",
              "capability": "Approval",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC11",
              "capability": "Rejection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC12",
              "capability": "Oversell prevention",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC13",
              "capability": "Single state transition",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC14",
              "capability": "Managed persistence",
              "critical": false,
              "final": "not-required"
            },
            {
              "id": "AC15",
              "capability": "Requester scoping",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC16",
              "capability": "Safe repeatable reset",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC17",
              "capability": "Keyboard and labels",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC18",
              "capability": "Responsive layout",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC19",
              "capability": "Honest demo boundary",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC20",
              "capability": "Actual private managed app",
              "critical": true,
              "final": "pass"
            }
          ]
        },
        {
          "id": "run-04",
          "model": "Sonnet 5",
          "effort": "High",
          "replicate": 1,
          "block": null,
          "status": "reviewed",
          "comparisonEligible": false,
          "final": {
            "points": 95,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": []
          },
          "firstCost": {
            "status": "observed",
            "value": 351,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 530,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 179,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": null,
              "lowerBoundSeconds": null,
              "observation": "not-captured",
              "authoringPrompts": 2,
              "observedPrompts": 0
            },
            "firstBuildResponseSeconds": null,
            "firstBuildResponseLowerBoundSeconds": null,
            "firstCompletionObservation": "not-captured",
            "finalReviewedResultSeconds": null,
            "timeToAcceptanceSeconds": null,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": null,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 1,
          "corrections": 1,
          "clarifications": 0,
          "approvals": 0,
          "screenshotFiles": 132,
          "archived": true,
          "promptSha256": "d0466119660c3b0fd7d25117d94fdd3c7db60302635650e7c00ff51dac9187e9",
          "checks": [
            {
              "id": "AC01",
              "capability": "Exact catalogue",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC02",
              "capability": "Search and empty state",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC03",
              "capability": "Combined filtering",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC04",
              "capability": "Unavailable stock",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC05",
              "capability": "Dashboard arithmetic",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC06",
              "capability": "Valid request",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC07",
              "capability": "Required fields",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC08",
              "capability": "Quantity validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC09",
              "capability": "Date validation",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC10",
              "capability": "Approval",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC11",
              "capability": "Rejection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC12",
              "capability": "Oversell prevention",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC13",
              "capability": "Single state transition",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC14",
              "capability": "Managed persistence",
              "critical": false,
              "final": "not-required"
            },
            {
              "id": "AC15",
              "capability": "Requester scoping",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC16",
              "capability": "Safe repeatable reset",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC17",
              "capability": "Keyboard and labels",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC18",
              "capability": "Responsive layout",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC19",
              "capability": "Honest demo boundary",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC20",
              "capability": "Actual private managed app",
              "critical": true,
              "final": "pass"
            }
          ]
        },
        {
          "id": "run-05",
          "model": "Sonnet 5",
          "effort": "Medium",
          "replicate": 2,
          "block": null,
          "status": "reviewed",
          "comparisonEligible": false,
          "final": {
            "points": 95,
            "maximum": 100,
            "accepted": true,
            "criticalFailures": [],
            "incomplete": []
          },
          "firstCost": {
            "status": "observed",
            "value": 174,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "finalCost": {
            "status": "observed",
            "value": 397,
            "unit": "credits",
            "scope": "Cowork task"
          },
          "increment": {
            "value": 223,
            "unit": "credits"
          },
          "timing": {
            "build": {
              "seconds": null,
              "lowerBoundSeconds": null,
              "observation": "not-captured",
              "authoringPrompts": 2,
              "observedPrompts": 0
            },
            "firstBuildResponseSeconds": null,
            "firstBuildResponseLowerBoundSeconds": null,
            "firstCompletionObservation": "not-captured",
            "finalReviewedResultSeconds": null,
            "timeToAcceptanceSeconds": null,
            "observedApprovalWaitSeconds": 0,
            "recordedInterruptionSeconds": 0,
            "approvalWaitStillOpen": false,
            "interruptionStillOpen": false,
            "firstCompletionDetectionWindowSeconds": null,
            "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
          },
          "reviews": 1,
          "corrections": 1,
          "clarifications": 0,
          "approvals": 0,
          "screenshotFiles": 147,
          "archived": true,
          "promptSha256": "d0466119660c3b0fd7d25117d94fdd3c7db60302635650e7c00ff51dac9187e9",
          "checks": [
            {
              "id": "AC01",
              "capability": "Exact catalogue",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC02",
              "capability": "Search and empty state",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC03",
              "capability": "Combined filtering",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC04",
              "capability": "Unavailable stock",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC05",
              "capability": "Dashboard arithmetic",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC06",
              "capability": "Valid request",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC07",
              "capability": "Required fields",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC08",
              "capability": "Quantity validation",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC09",
              "capability": "Date validation",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC10",
              "capability": "Approval",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC11",
              "capability": "Rejection",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC12",
              "capability": "Oversell prevention",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC13",
              "capability": "Single state transition",
              "critical": true,
              "final": "pass"
            },
            {
              "id": "AC14",
              "capability": "Managed persistence",
              "critical": false,
              "final": "not-required"
            },
            {
              "id": "AC15",
              "capability": "Requester scoping",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC16",
              "capability": "Safe repeatable reset",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC17",
              "capability": "Keyboard and labels",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC18",
              "capability": "Responsive layout",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC19",
              "capability": "Honest demo boundary",
              "critical": false,
              "final": "pass"
            },
            {
              "id": "AC20",
              "capability": "Actual private managed app",
              "critical": true,
              "final": "pass"
            }
          ]
        }
      ],
      "comparisonEligible": false,
      "reviewMode": "build-and-native-review",
      "acceptanceCheckCount": 20,
      "acceptanceScope": {
        "virtualDataAllowed": true,
        "notRequired": [
          "AC14"
        ],
        "note": "Virtual/local data is allowed. Managed-store persistence is not required; recorded scores are unchanged."
      },
      "protocolSha256": "b4a373aa7348f0a09964a1313dbb779dc033e74ec33fb62f070ba6ac1100e460"
    }
  ],
  "validation": {
    "title": "Additional external-review measurements",
    "note": "Original cohorts are preserved, not pooled with these observations. The four common settings provide a like-for-like model mix, but virtual approvals and SharePoint request tracking are still different app scopes.",
    "promptNote": "The virtual sample spans two prompt versions: the same business specification, with private owner-only publication explicitly permitted in version 2. All five additional SharePoint builds use one identical amended prompt. Publication permission and review method differ from the original cohorts; the differences do not isolate connector pricing.",
    "cohorts": [
      {
        "test": "test-01",
        "reviewMode": "external-only",
        "runs": [
          {
            "id": "run-01",
            "model": "Sonnet 5",
            "effort": "Medium",
            "firstCost": {
              "status": "observed",
              "value": 215,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "finalCost": {
              "status": "observed",
              "value": 215,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "final": {
              "points": 100,
              "maximum": 100,
              "accepted": true,
              "criticalFailures": [],
              "incomplete": []
            },
            "reviews": 0,
            "corrections": 0,
            "clarifications": 0,
            "timing": {
              "build": {
                "seconds": 591.786,
                "lowerBoundSeconds": 583.64,
                "observation": "interval-bounded",
                "authoringPrompts": 1,
                "observedPrompts": 1
              },
              "firstBuildResponseSeconds": 591.786,
              "firstBuildResponseLowerBoundSeconds": 583.64,
              "firstCompletionObservation": "interval-bounded",
              "finalReviewedResultSeconds": 1269.205,
              "timeToAcceptanceSeconds": 1269.205,
              "observedApprovalWaitSeconds": 0,
              "recordedInterruptionSeconds": 0,
              "approvalWaitStillOpen": false,
              "interruptionStillOpen": false,
              "firstCompletionDetectionWindowSeconds": 8.146,
              "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
            },
            "promptSha256": "371fc49875e5c19170df827535fd11dc075b8c6080660f5b4d4cac898c21e506",
            "promptVersion": "virtual-v1",
            "screenshotFiles": 94,
            "archived": true,
            "checks": [
              {
                "id": "AC01",
                "capability": "Exact catalogue",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC02",
                "capability": "Search and empty state",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC03",
                "capability": "Combined filtering",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC04",
                "capability": "Unavailable stock",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC05",
                "capability": "Dashboard arithmetic",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC06",
                "capability": "Valid request",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC07",
                "capability": "Required fields",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC08",
                "capability": "Quantity validation",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC09",
                "capability": "Date validation",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC10",
                "capability": "Approval",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC11",
                "capability": "Rejection",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC12",
                "capability": "Oversell prevention",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC13",
                "capability": "Single state transition",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC14",
                "capability": "Virtual/local reopen persistence",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC15",
                "capability": "Requester scoping",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC16",
                "capability": "Safe repeatable reset",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC17",
                "capability": "Keyboard and labels",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC18",
                "capability": "Responsive layout",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC19",
                "capability": "Honest demo boundary",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC20",
                "capability": "Actual private managed app",
                "critical": true,
                "final": "pass"
              }
            ]
          },
          {
            "id": "run-02",
            "model": "Opus 5",
            "effort": "High",
            "firstCost": {
              "status": "observed",
              "value": 611,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "finalCost": {
              "status": "observed",
              "value": 826,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "final": {
              "points": 100,
              "maximum": 100,
              "accepted": true,
              "criticalFailures": [],
              "incomplete": []
            },
            "reviews": 0,
            "corrections": 1,
            "clarifications": 0,
            "timing": {
              "build": {
                "seconds": 901.309,
                "lowerBoundSeconds": 896.781,
                "observation": "interval-bounded",
                "authoringPrompts": 2,
                "observedPrompts": 2
              },
              "firstBuildResponseSeconds": 731.296,
              "firstBuildResponseLowerBoundSeconds": 728.986,
              "firstCompletionObservation": "interval-bounded",
              "finalReviewedResultSeconds": 1684.242,
              "timeToAcceptanceSeconds": null,
              "observedApprovalWaitSeconds": 0,
              "recordedInterruptionSeconds": 0,
              "approvalWaitStillOpen": false,
              "interruptionStillOpen": false,
              "firstCompletionDetectionWindowSeconds": 2.31,
              "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
            },
            "promptSha256": "371fc49875e5c19170df827535fd11dc075b8c6080660f5b4d4cac898c21e506",
            "promptVersion": "virtual-v1",
            "screenshotFiles": 137,
            "archived": true,
            "checks": [
              {
                "id": "AC01",
                "capability": "Exact catalogue",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC02",
                "capability": "Search and empty state",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC03",
                "capability": "Combined filtering",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC04",
                "capability": "Unavailable stock",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC05",
                "capability": "Dashboard arithmetic",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC06",
                "capability": "Valid request",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC07",
                "capability": "Required fields",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC08",
                "capability": "Quantity validation",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC09",
                "capability": "Date validation",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC10",
                "capability": "Approval",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC11",
                "capability": "Rejection",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC12",
                "capability": "Oversell prevention",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC13",
                "capability": "Single state transition",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC14",
                "capability": "Virtual/local reopen persistence",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC15",
                "capability": "Requester scoping",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC16",
                "capability": "Safe repeatable reset",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC17",
                "capability": "Keyboard and labels",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC18",
                "capability": "Responsive layout",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC19",
                "capability": "Honest demo boundary",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC20",
                "capability": "Actual private managed app",
                "critical": true,
                "final": "pass"
              }
            ]
          },
          {
            "id": "run-04",
            "model": "Sonnet 5",
            "effort": "High",
            "firstCost": {
              "status": "observed",
              "value": 220,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "finalCost": {
              "status": "observed",
              "value": 220,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "final": {
              "points": 100,
              "maximum": 100,
              "accepted": true,
              "criticalFailures": [],
              "incomplete": []
            },
            "reviews": 0,
            "corrections": 0,
            "clarifications": 0,
            "timing": {
              "build": {
                "seconds": 673.43,
                "lowerBoundSeconds": 671.255,
                "observation": "interval-bounded",
                "authoringPrompts": 1,
                "observedPrompts": 1
              },
              "firstBuildResponseSeconds": 673.43,
              "firstBuildResponseLowerBoundSeconds": 671.255,
              "firstCompletionObservation": "interval-bounded",
              "finalReviewedResultSeconds": 841.612,
              "timeToAcceptanceSeconds": 841.612,
              "observedApprovalWaitSeconds": 0,
              "recordedInterruptionSeconds": 0,
              "approvalWaitStillOpen": false,
              "interruptionStillOpen": false,
              "firstCompletionDetectionWindowSeconds": 2.175,
              "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
            },
            "promptSha256": "15e2426b3b9e06c35706e48a928e70a454859b3fcf6e4d2dec8d6255b01a9e4a",
            "promptVersion": "virtual-v2",
            "screenshotFiles": 110,
            "archived": true,
            "checks": [
              {
                "id": "AC01",
                "capability": "Exact catalogue",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC02",
                "capability": "Search and empty state",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC03",
                "capability": "Combined filtering",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC04",
                "capability": "Unavailable stock",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC05",
                "capability": "Dashboard arithmetic",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC06",
                "capability": "Valid request",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC07",
                "capability": "Required fields",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC08",
                "capability": "Quantity validation",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC09",
                "capability": "Date validation",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC10",
                "capability": "Approval",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC11",
                "capability": "Rejection",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC12",
                "capability": "Oversell prevention",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC13",
                "capability": "Single state transition",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC14",
                "capability": "Virtual/local reopen persistence",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC15",
                "capability": "Requester scoping",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC16",
                "capability": "Safe repeatable reset",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC17",
                "capability": "Keyboard and labels",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC18",
                "capability": "Responsive layout",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC19",
                "capability": "Honest demo boundary",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC20",
                "capability": "Actual private managed app",
                "critical": true,
                "final": "pass"
              }
            ]
          },
          {
            "id": "run-05",
            "model": "GPT 5.5",
            "effort": "Medium",
            "firstCost": {
              "status": "observed",
              "value": 199,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "finalCost": {
              "status": "observed",
              "value": 269,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "final": {
              "points": 100,
              "maximum": 100,
              "accepted": true,
              "criticalFailures": [],
              "incomplete": []
            },
            "reviews": 0,
            "corrections": 1,
            "clarifications": 0,
            "timing": {
              "build": {
                "seconds": 545.105,
                "lowerBoundSeconds": 539.408,
                "observation": "interval-bounded",
                "authoringPrompts": 2,
                "observedPrompts": 2
              },
              "firstBuildResponseSeconds": 366.222,
              "firstBuildResponseLowerBoundSeconds": 362.637,
              "firstCompletionObservation": "interval-bounded",
              "finalReviewedResultSeconds": 1154.629,
              "timeToAcceptanceSeconds": 1154.629,
              "observedApprovalWaitSeconds": 0,
              "recordedInterruptionSeconds": 0,
              "approvalWaitStillOpen": false,
              "interruptionStillOpen": false,
              "firstCompletionDetectionWindowSeconds": 3.585,
              "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
            },
            "promptSha256": "15e2426b3b9e06c35706e48a928e70a454859b3fcf6e4d2dec8d6255b01a9e4a",
            "promptVersion": "virtual-v2",
            "screenshotFiles": 120,
            "archived": true,
            "checks": [
              {
                "id": "AC01",
                "capability": "Exact catalogue",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC02",
                "capability": "Search and empty state",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC03",
                "capability": "Combined filtering",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC04",
                "capability": "Unavailable stock",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC05",
                "capability": "Dashboard arithmetic",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC06",
                "capability": "Valid request",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC07",
                "capability": "Required fields",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC08",
                "capability": "Quantity validation",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC09",
                "capability": "Date validation",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC10",
                "capability": "Approval",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC11",
                "capability": "Rejection",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC12",
                "capability": "Oversell prevention",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC13",
                "capability": "Single state transition",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC14",
                "capability": "Virtual/local reopen persistence",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC15",
                "capability": "Requester scoping",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC16",
                "capability": "Safe repeatable reset",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC17",
                "capability": "Keyboard and labels",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC18",
                "capability": "Responsive layout",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC19",
                "capability": "Honest demo boundary",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC20",
                "capability": "Actual private managed app",
                "critical": true,
                "final": "pass"
              }
            ]
          }
        ],
        "meanFinalCredits": 382.5,
        "note": "Four reviewed virtual-data observations, one per shown setting. Browser-local persistence is explicitly allowed and independently verified; no native self-review."
      },
      {
        "test": "test-02",
        "reviewMode": "external-only",
        "runs": [
          {
            "id": "run-01",
            "model": "Sonnet 5",
            "effort": "Medium",
            "firstCost": {
              "status": "observed",
              "value": 301,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "finalCost": {
              "status": "observed",
              "value": 301,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "final": {
              "points": 90,
              "maximum": 100,
              "accepted": true,
              "criticalFailures": [],
              "incomplete": [
                "BC09"
              ]
            },
            "reviews": 0,
            "corrections": 0,
            "clarifications": 0,
            "timing": {
              "build": {
                "seconds": 715.385,
                "lowerBoundSeconds": 712.917,
                "observation": "interval-bounded",
                "authoringPrompts": 1,
                "observedPrompts": 1
              },
              "firstBuildResponseSeconds": 715.385,
              "firstBuildResponseLowerBoundSeconds": 712.917,
              "firstCompletionObservation": "interval-bounded",
              "finalReviewedResultSeconds": 1084.126,
              "timeToAcceptanceSeconds": 1084.126,
              "observedApprovalWaitSeconds": 0,
              "recordedInterruptionSeconds": 0,
              "approvalWaitStillOpen": false,
              "interruptionStillOpen": false,
              "firstCompletionDetectionWindowSeconds": 2.468,
              "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
            },
            "promptSha256": "77f27f229570d44ae0c3114b5ca3efc644cd03718bf2b03a07711468592f31a8",
            "promptVersion": "sharepoint-v2",
            "screenshotFiles": 92,
            "archived": true,
            "checks": [
              {
                "id": "BC01",
                "capability": "Runnable private managed app",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC02",
                "capability": "Live catalogue connection",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC03",
                "capability": "Live dashboard arithmetic",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC04",
                "capability": "Catalogue search and filtering",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC05",
                "capability": "Complete saved request",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC06",
                "capability": "Basic request validation",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC07",
                "capability": "Live request listing",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC08",
                "capability": "Reopen persistence",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC09",
                "capability": "Honest data and error states",
                "critical": false,
                "final": "blocked"
              },
              {
                "id": "BC10",
                "capability": "Usable desktop and mobile build",
                "critical": false,
                "final": "pass"
              }
            ]
          },
          {
            "id": "run-02",
            "model": "Opus 5",
            "effort": "High",
            "firstCost": {
              "status": "observed",
              "value": 868,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "finalCost": {
              "status": "observed",
              "value": 1130,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "final": {
              "points": 90,
              "maximum": 100,
              "accepted": true,
              "criticalFailures": [],
              "incomplete": [
                "BC09"
              ]
            },
            "reviews": 0,
            "corrections": 1,
            "clarifications": 0,
            "timing": {
              "build": {
                "seconds": 912.876,
                "lowerBoundSeconds": 908.157,
                "observation": "interval-bounded",
                "authoringPrompts": 2,
                "observedPrompts": 2
              },
              "firstBuildResponseSeconds": 759.285,
              "firstBuildResponseLowerBoundSeconds": 757.064,
              "firstCompletionObservation": "interval-bounded",
              "finalReviewedResultSeconds": 1179.952,
              "timeToAcceptanceSeconds": 1179.952,
              "observedApprovalWaitSeconds": 0,
              "recordedInterruptionSeconds": 0,
              "approvalWaitStillOpen": false,
              "interruptionStillOpen": false,
              "firstCompletionDetectionWindowSeconds": 2.221,
              "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
            },
            "promptSha256": "77f27f229570d44ae0c3114b5ca3efc644cd03718bf2b03a07711468592f31a8",
            "promptVersion": "sharepoint-v2",
            "screenshotFiles": 124,
            "archived": true,
            "checks": [
              {
                "id": "BC01",
                "capability": "Runnable private managed app",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC02",
                "capability": "Live catalogue connection",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC03",
                "capability": "Live dashboard arithmetic",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC04",
                "capability": "Catalogue search and filtering",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC05",
                "capability": "Complete saved request",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC06",
                "capability": "Basic request validation",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC07",
                "capability": "Live request listing",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC08",
                "capability": "Reopen persistence",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC09",
                "capability": "Honest data and error states",
                "critical": false,
                "final": "blocked"
              },
              {
                "id": "BC10",
                "capability": "Usable desktop and mobile build",
                "critical": false,
                "final": "pass"
              }
            ]
          },
          {
            "id": "run-03",
            "model": "Opus 5",
            "effort": "Medium",
            "firstCost": {
              "status": "observed",
              "value": 544,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "finalCost": {
              "status": "observed",
              "value": 842,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "final": {
              "points": 90,
              "maximum": 100,
              "accepted": true,
              "criticalFailures": [],
              "incomplete": [
                "BC09"
              ]
            },
            "reviews": 0,
            "corrections": 1,
            "clarifications": 0,
            "timing": {
              "build": {
                "seconds": 875.964,
                "lowerBoundSeconds": 871.543,
                "observation": "interval-bounded",
                "authoringPrompts": 2,
                "observedPrompts": 2
              },
              "firstBuildResponseSeconds": 598.961,
              "firstBuildResponseLowerBoundSeconds": 596.819,
              "firstCompletionObservation": "interval-bounded",
              "finalReviewedResultSeconds": 1208.886,
              "timeToAcceptanceSeconds": 1208.886,
              "observedApprovalWaitSeconds": 0,
              "recordedInterruptionSeconds": 0,
              "approvalWaitStillOpen": false,
              "interruptionStillOpen": false,
              "firstCompletionDetectionWindowSeconds": 2.142,
              "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
            },
            "promptSha256": "77f27f229570d44ae0c3114b5ca3efc644cd03718bf2b03a07711468592f31a8",
            "promptVersion": "sharepoint-v2",
            "screenshotFiles": 123,
            "archived": true,
            "checks": [
              {
                "id": "BC01",
                "capability": "Runnable private managed app",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC02",
                "capability": "Live catalogue connection",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC03",
                "capability": "Live dashboard arithmetic",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC04",
                "capability": "Catalogue search and filtering",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC05",
                "capability": "Complete saved request",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC06",
                "capability": "Basic request validation",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC07",
                "capability": "Live request listing",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC08",
                "capability": "Reopen persistence",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC09",
                "capability": "Honest data and error states",
                "critical": false,
                "final": "blocked"
              },
              {
                "id": "BC10",
                "capability": "Usable desktop and mobile build",
                "critical": false,
                "final": "pass"
              }
            ]
          },
          {
            "id": "run-04",
            "model": "Sonnet 5",
            "effort": "High",
            "firstCost": {
              "status": "observed",
              "value": 586,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "finalCost": {
              "status": "observed",
              "value": 1025,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "final": {
              "points": 90,
              "maximum": 100,
              "accepted": true,
              "criticalFailures": [],
              "incomplete": [
                "BC09"
              ]
            },
            "reviews": 0,
            "corrections": 2,
            "clarifications": 0,
            "timing": {
              "build": {
                "seconds": 1656.779,
                "lowerBoundSeconds": 1645.937,
                "observation": "interval-bounded",
                "authoringPrompts": 3,
                "observedPrompts": 3
              },
              "firstBuildResponseSeconds": 941.877,
              "firstBuildResponseLowerBoundSeconds": 939.669,
              "firstCompletionObservation": "interval-bounded",
              "finalReviewedResultSeconds": 2678.46,
              "timeToAcceptanceSeconds": 2678.46,
              "observedApprovalWaitSeconds": 0,
              "recordedInterruptionSeconds": 0,
              "approvalWaitStillOpen": false,
              "interruptionStillOpen": false,
              "firstCompletionDetectionWindowSeconds": 2.208,
              "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
            },
            "promptSha256": "77f27f229570d44ae0c3114b5ca3efc644cd03718bf2b03a07711468592f31a8",
            "promptVersion": "sharepoint-v2",
            "screenshotFiles": 176,
            "archived": true,
            "checks": [
              {
                "id": "BC01",
                "capability": "Runnable private managed app",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC02",
                "capability": "Live catalogue connection",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC03",
                "capability": "Live dashboard arithmetic",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC04",
                "capability": "Catalogue search and filtering",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC05",
                "capability": "Complete saved request",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC06",
                "capability": "Basic request validation",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC07",
                "capability": "Live request listing",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC08",
                "capability": "Reopen persistence",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC09",
                "capability": "Honest data and error states",
                "critical": false,
                "final": "blocked"
              },
              {
                "id": "BC10",
                "capability": "Usable desktop and mobile build",
                "critical": false,
                "final": "pass"
              }
            ]
          },
          {
            "id": "run-05",
            "model": "GPT 5.5",
            "effort": "Medium",
            "firstCost": {
              "status": "observed",
              "value": 314,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "finalCost": {
              "status": "observed",
              "value": 314,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "final": {
              "points": 90,
              "maximum": 100,
              "accepted": true,
              "criticalFailures": [],
              "incomplete": [
                "BC09"
              ]
            },
            "reviews": 0,
            "corrections": 0,
            "clarifications": 0,
            "timing": {
              "build": {
                "seconds": 468.877,
                "lowerBoundSeconds": 466.578,
                "observation": "interval-bounded",
                "authoringPrompts": 1,
                "observedPrompts": 1
              },
              "firstBuildResponseSeconds": 468.877,
              "firstBuildResponseLowerBoundSeconds": 466.578,
              "firstCompletionObservation": "interval-bounded",
              "finalReviewedResultSeconds": 1335.266,
              "timeToAcceptanceSeconds": 1335.266,
              "observedApprovalWaitSeconds": 0,
              "recordedInterruptionSeconds": 0,
              "approvalWaitStillOpen": false,
              "interruptionStillOpen": false,
              "firstCompletionDetectionWindowSeconds": 2.299,
              "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
            },
            "promptSha256": "77f27f229570d44ae0c3114b5ca3efc644cd03718bf2b03a07711468592f31a8",
            "promptVersion": "sharepoint-v2",
            "screenshotFiles": 109,
            "archived": true,
            "checks": [
              {
                "id": "BC01",
                "capability": "Runnable private managed app",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC02",
                "capability": "Live catalogue connection",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC03",
                "capability": "Live dashboard arithmetic",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC04",
                "capability": "Catalogue search and filtering",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC05",
                "capability": "Complete saved request",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC06",
                "capability": "Basic request validation",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC07",
                "capability": "Live request listing",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC08",
                "capability": "Reopen persistence",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC09",
                "capability": "Honest data and error states",
                "critical": false,
                "final": "blocked"
              },
              {
                "id": "BC10",
                "capability": "Usable desktop and mobile build",
                "critical": false,
                "final": "pass"
              }
            ]
          }
        ],
        "meanFinalCredits": 722.4,
        "note": "Five additional SharePoint observations, one per setting, using the same four-view business scope, owned fixture and external-only review."
      }
    ],
    "matchedConfigurations": [
      "Sonnet 5 / Medium",
      "Opus 5 / High",
      "Sonnet 5 / High",
      "GPT 5.5 / Medium"
    ],
    "matchedMeans": {
      "test-01": 382.5,
      "test-02": 692.5
    }
  },
  "complexValidation": {
    "test": "test-03",
    "title": "Additional complex-app measurements",
    "reviewMode": "external-only",
    "promptPath": "benchmark/test-03-repeat/stage-01-build.txt",
    "runs": [
      {
        "id": "run-06",
        "model": "GPT 5.5",
        "effort": "Medium",
        "replicate": 2,
        "firstCost": {
          "status": "observed",
          "value": 183,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "finalCost": {
          "status": "observed",
          "value": 573,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "final": {
          "points": 85,
          "maximum": 100,
          "accepted": true,
          "criticalFailures": [],
          "incomplete": [
            "CX14",
            "CX16"
          ]
        },
        "corrections": 1,
        "clarifications": 0,
        "reviews": 0,
        "plannedChanges": 3,
        "timing": {
          "build": {
            "seconds": 1268.978,
            "lowerBoundSeconds": 1253.919,
            "observation": "interval-bounded",
            "authoringPrompts": 5,
            "observedPrompts": 5
          },
          "firstBuildResponseSeconds": 339.929,
          "firstBuildResponseLowerBoundSeconds": 337.225,
          "firstCompletionObservation": "interval-bounded",
          "finalReviewedResultSeconds": 5143.882,
          "timeToAcceptanceSeconds": 5143.882,
          "observedApprovalWaitSeconds": 0,
          "recordedInterruptionSeconds": 0,
          "approvalWaitStillOpen": false,
          "interruptionStillOpen": false,
          "firstCompletionDetectionWindowSeconds": 2.704,
          "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
        },
        "stages": [
          {
            "id": "stage-01",
            "title": "Build the relational virtual-data app first",
            "cost": {
              "status": "observed",
              "value": 183,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "costBeforePrompt": null,
            "increaseSincePrecedingReading": null
          },
          {
            "id": "stage-02",
            "title": "Connect the existing managed app to Office 365 Users",
            "cost": {
              "status": "observed",
              "value": 276,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "costBeforePrompt": {
              "status": "observed",
              "value": 183,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "increaseSincePrecedingReading": {
              "value": 93,
              "unit": "credits"
            }
          },
          {
            "id": "stage-03",
            "title": "Virtual approvals, asset lifecycle, themes and preferences",
            "cost": {
              "status": "observed",
              "value": 397,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "costBeforePrompt": {
              "status": "observed",
              "value": 276,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "increaseSincePrecedingReading": {
              "value": 121,
              "unit": "credits"
            }
          },
          {
            "id": "stage-04",
            "title": "Approval policy change, revisions, cost centres and responsive refinement",
            "cost": {
              "status": "observed",
              "value": 573,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "costBeforePrompt": {
              "status": "observed",
              "value": 442,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "increaseSincePrecedingReading": {
              "value": 131,
              "unit": "credits"
            }
          }
        ],
        "promptSha256": "00e76f1c01c59953a44e6199dedb1bad741dbb4ec4f5051c18e150d32f9a175c",
        "promptVersion": "complex-private-v2",
        "archived": true,
        "screenshotFiles": 378,
        "profileProof": {
          "verifiedInApp": true,
          "reverifiedAfterFinalReopen": true,
          "directoryWritesObserved": 0
        },
        "preferences": {
          "saveCancelAndPersistence": "verified",
          "requestRowDensity": "not-demonstrated",
          "rowHeightDifferenceCssPixels": 0,
          "gapDifferenceCssPixels": 0,
          "defaultDepartmentConsumption": "not-demonstrated"
        },
        "checks": [
          {
            "id": "CX01",
            "capability": "Same private app opens with all seven navigable screens",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX02",
            "capability": "Exact six-device/four-request relational seed and derived initial metrics",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX03",
            "capability": "Catalogue search, combined category/stock filters and zero-stock protection",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX04",
            "capability": "Multi-line draft creation, line edits/removal and accurate GBP totals",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX05",
            "capability": "Required fields, dates, whole positive quantities, duplicates and stock validation",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX06",
            "capability": "Draft/persona isolation, edits and exact-once submission with stable IDs",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX07",
            "capability": "Controller-created draft, IDs, lines and timeline survive all three planned changes and reopen",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX08",
            "capability": "Virtual manager/finance role gates and approval transitions at the final threshold",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX09",
            "capability": "Rejection comments, append-only actor/timestamp history and terminal action guards",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX10",
            "capability": "Request changes, requester edit/resubmit and cancellation preserve identity/history",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX11",
            "capability": "Atomic fulfilment, stock revalidation, per-unit assets and duplicate-fulfilment protection",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX12",
            "capability": "Good/damaged asset returns and duplicate-return protection",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX13",
            "capability": "Current-user Office 365 Users profile actually reads through supported app binding",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX14",
            "capability": "Profile loading/error/retry and genuinely missing optional fields are honest",
            "critical": false,
            "final": "blocked"
          },
          {
            "id": "CX15",
            "capability": "Explicit light/dark mode applies across all screens and persists on reopen",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX16",
            "capability": "Local preference Save/Cancel, density and default department work without directory writes",
            "critical": false,
            "final": "blocked"
          },
          {
            "id": "CX17",
            "capability": "Cost-centre migration, persistence and department mismatch validation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX18",
            "capability": "Derived status/asset/stock metrics, low-stock navigation and combined request filters",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX19",
            "capability": "All screens and key forms are usable at actual wide desktop and 390px mobile widths",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX20",
            "capability": "Labels, visible keyboard focus, readable contrast and reachable form errors",
            "critical": false,
            "final": "fail"
          }
        ]
      },
      {
        "id": "run-07",
        "model": "Sonnet 5",
        "effort": "High",
        "replicate": 2,
        "firstCost": {
          "status": "observed",
          "value": 433,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "finalCost": {
          "status": "observed",
          "value": 2134,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "final": {
          "points": 85,
          "maximum": 100,
          "accepted": true,
          "criticalFailures": [],
          "incomplete": [
            "CX14",
            "CX16"
          ]
        },
        "corrections": 1,
        "clarifications": 0,
        "reviews": 0,
        "plannedChanges": 3,
        "timing": {
          "build": {
            "seconds": 2950.698,
            "lowerBoundSeconds": 2923.852,
            "observation": "interval-bounded",
            "authoringPrompts": 5,
            "observedPrompts": 5
          },
          "firstBuildResponseSeconds": 1037.248,
          "firstBuildResponseLowerBoundSeconds": 1033.201,
          "firstCompletionObservation": "interval-bounded",
          "finalReviewedResultSeconds": 5135.595,
          "timeToAcceptanceSeconds": 5135.595,
          "observedApprovalWaitSeconds": 0,
          "recordedInterruptionSeconds": 0,
          "approvalWaitStillOpen": false,
          "interruptionStillOpen": false,
          "firstCompletionDetectionWindowSeconds": 4.047,
          "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
        },
        "stages": [
          {
            "id": "stage-01",
            "title": "Build the relational virtual-data app first",
            "cost": {
              "status": "observed",
              "value": 433,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "costBeforePrompt": null,
            "increaseSincePrecedingReading": null
          },
          {
            "id": "stage-02",
            "title": "Connect the existing managed app to Office 365 Users",
            "cost": {
              "status": "observed",
              "value": 654,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "costBeforePrompt": {
              "status": "observed",
              "value": 497,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "increaseSincePrecedingReading": {
              "value": 157,
              "unit": "credits"
            }
          },
          {
            "id": "stage-03",
            "title": "Virtual approvals, asset lifecycle, themes and preferences",
            "cost": {
              "status": "observed",
              "value": 1232,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "costBeforePrompt": {
              "status": "observed",
              "value": 654,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "increaseSincePrecedingReading": {
              "value": 578,
              "unit": "credits"
            }
          },
          {
            "id": "stage-04",
            "title": "Approval policy change, revisions, cost centres and responsive refinement",
            "cost": {
              "status": "observed",
              "value": 2134,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "costBeforePrompt": {
              "status": "observed",
              "value": 1232,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "increaseSincePrecedingReading": {
              "value": 902,
              "unit": "credits"
            }
          }
        ],
        "promptSha256": "00e76f1c01c59953a44e6199dedb1bad741dbb4ec4f5051c18e150d32f9a175c",
        "promptVersion": "complex-private-v2",
        "archived": true,
        "screenshotFiles": 401,
        "profileProof": {
          "verifiedInApp": true,
          "reverifiedAfterFinalReopen": true,
          "directoryWritesObserved": 0
        },
        "preferences": {
          "saveCancelAndPersistence": "verified",
          "requestRowDensity": "verified",
          "rowHeightDifferenceCssPixels": 4,
          "gapDifferenceCssPixels": 0,
          "defaultDepartmentConsumption": "not-demonstrated"
        },
        "checks": [
          {
            "id": "CX01",
            "capability": "Same private app opens with all seven navigable screens",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX02",
            "capability": "Exact six-device/four-request relational seed and derived initial metrics",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX03",
            "capability": "Catalogue search, combined category/stock filters and zero-stock protection",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX04",
            "capability": "Multi-line draft creation, line edits/removal and accurate GBP totals",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX05",
            "capability": "Required fields, dates, whole positive quantities, duplicates and stock validation",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX06",
            "capability": "Draft/persona isolation, edits and exact-once submission with stable IDs",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX07",
            "capability": "Controller-created draft, IDs, lines and timeline survive all three planned changes and reopen",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX08",
            "capability": "Virtual manager/finance role gates and approval transitions at the final threshold",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX09",
            "capability": "Rejection comments, append-only actor/timestamp history and terminal action guards",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX10",
            "capability": "Request changes, requester edit/resubmit and cancellation preserve identity/history",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX11",
            "capability": "Atomic fulfilment, stock revalidation, per-unit assets and duplicate-fulfilment protection",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX12",
            "capability": "Good/damaged asset returns and duplicate-return protection",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX13",
            "capability": "Current-user Office 365 Users profile actually reads through supported app binding",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX14",
            "capability": "Profile loading/error/retry and genuinely missing optional fields are honest",
            "critical": false,
            "final": "blocked"
          },
          {
            "id": "CX15",
            "capability": "Explicit light/dark mode applies across all screens and persists on reopen",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX16",
            "capability": "Local preference Save/Cancel, density and default department work without directory writes",
            "critical": false,
            "final": "blocked"
          },
          {
            "id": "CX17",
            "capability": "Cost-centre migration, persistence and department mismatch validation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX18",
            "capability": "Derived status/asset/stock metrics, low-stock navigation and combined request filters",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX19",
            "capability": "All screens and key forms are usable at actual wide desktop and 390px mobile widths",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX20",
            "capability": "Labels, visible keyboard focus, readable contrast and reachable form errors",
            "critical": false,
            "final": "fail"
          }
        ]
      },
      {
        "id": "run-09",
        "model": "Opus 5",
        "effort": "High",
        "replicate": 2,
        "firstCost": {
          "status": "observed",
          "value": 521,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "finalCost": {
          "status": "observed",
          "value": 2860,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "final": {
          "points": 95,
          "maximum": 100,
          "accepted": true,
          "criticalFailures": [],
          "incomplete": [
            "CX14"
          ]
        },
        "corrections": 2,
        "clarifications": 0,
        "reviews": 0,
        "plannedChanges": 3,
        "timing": {
          "build": {
            "seconds": 2565.492,
            "lowerBoundSeconds": 2541.55,
            "observation": "interval-bounded",
            "authoringPrompts": 6,
            "observedPrompts": 6
          },
          "firstBuildResponseSeconds": 824.791,
          "firstBuildResponseLowerBoundSeconds": 821.099,
          "firstCompletionObservation": "interval-bounded",
          "finalReviewedResultSeconds": 26070.834,
          "timeToAcceptanceSeconds": 26070.834,
          "observedApprovalWaitSeconds": 0,
          "recordedInterruptionSeconds": 0,
          "approvalWaitStillOpen": false,
          "interruptionStillOpen": false,
          "firstCompletionDetectionWindowSeconds": 3.692,
          "basis": "Observed wall-clock time, including waits; not isolated model execution. A completion first seen after it happened is an upper bound. Missing endpoints stay null."
        },
        "stages": [
          {
            "id": "stage-01",
            "title": "Build the relational virtual-data app first",
            "cost": {
              "status": "observed",
              "value": 521,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "costBeforePrompt": null,
            "increaseSincePrecedingReading": null
          },
          {
            "id": "stage-02",
            "title": "Connect the existing managed app to Office 365 Users",
            "cost": {
              "status": "observed",
              "value": 1219,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "costBeforePrompt": {
              "status": "observed",
              "value": 832,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "increaseSincePrecedingReading": {
              "value": 387,
              "unit": "credits"
            }
          },
          {
            "id": "stage-03",
            "title": "Virtual approvals, asset lifecycle, themes and preferences",
            "cost": {
              "status": "observed",
              "value": 1860,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "costBeforePrompt": {
              "status": "observed",
              "value": 1219,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "increaseSincePrecedingReading": {
              "value": 641,
              "unit": "credits"
            }
          },
          {
            "id": "stage-04",
            "title": "Approval policy change, revisions, cost centres and responsive refinement",
            "cost": {
              "status": "observed",
              "value": 2431,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "costBeforePrompt": {
              "status": "observed",
              "value": 1860,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "increaseSincePrecedingReading": {
              "value": 571,
              "unit": "credits"
            }
          }
        ],
        "promptSha256": "00e76f1c01c59953a44e6199dedb1bad741dbb4ec4f5051c18e150d32f9a175c",
        "promptVersion": "complex-private-v2",
        "archived": true,
        "screenshotFiles": 465,
        "profileProof": {
          "verifiedInApp": true,
          "reverifiedAfterFinalReopen": true,
          "directoryWritesObserved": 0
        },
        "preferences": {
          "saveCancelAndPersistence": "verified",
          "requestRowDensity": "verified",
          "rowHeightDifferenceCssPixels": 8,
          "gapDifferenceCssPixels": 0,
          "defaultDepartmentConsumption": "verified",
          "defaultDepartmentSurfaces": [
            "Approvals department filter",
            "All-requests department filter"
          ]
        },
        "checks": [
          {
            "id": "CX01",
            "capability": "Same private app opens with all seven navigable screens",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX02",
            "capability": "Exact six-device/four-request relational seed and derived initial metrics",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX03",
            "capability": "Catalogue search, combined category/stock filters and zero-stock protection",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX04",
            "capability": "Multi-line draft creation, line edits/removal and accurate GBP totals",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX05",
            "capability": "Required fields, dates, whole positive quantities, duplicates and stock validation",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX06",
            "capability": "Draft/persona isolation, edits and exact-once submission with stable IDs",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX07",
            "capability": "Controller-created draft, IDs, lines and timeline survive all three planned changes and reopen",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX08",
            "capability": "Virtual manager/finance role gates and approval transitions at the final threshold",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX09",
            "capability": "Rejection comments, append-only actor/timestamp history and terminal action guards",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX10",
            "capability": "Request changes, requester edit/resubmit and cancellation preserve identity/history",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX11",
            "capability": "Atomic fulfilment, stock revalidation, per-unit assets and duplicate-fulfilment protection",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX12",
            "capability": "Good/damaged asset returns and duplicate-return protection",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX13",
            "capability": "Current-user Office 365 Users profile actually reads through supported app binding",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX14",
            "capability": "Profile loading/error/retry and genuinely missing optional fields are honest",
            "critical": false,
            "final": "blocked"
          },
          {
            "id": "CX15",
            "capability": "Explicit light/dark mode applies across all screens and persists on reopen",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX16",
            "capability": "Local preference Save/Cancel, density and default department work without directory writes",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX17",
            "capability": "Cost-centre migration, persistence and department mismatch validation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX18",
            "capability": "Derived status/asset/stock metrics, low-stock navigation and combined request filters",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "CX19",
            "capability": "All screens and key forms are usable at actual wide desktop and 390px mobile widths",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "CX20",
            "capability": "Labels, visible keyboard focus, readable contrast and reachable form errors",
            "critical": false,
            "final": "pass"
          }
        ]
      }
    ],
    "pairs": [
      {
        "model": "GPT 5.5",
        "effort": "Medium",
        "observations": 2,
        "originalCredits": 944,
        "additionalCredits": 573,
        "meanFinalCredits": 758.5,
        "originalCorrections": 0,
        "additionalCorrections": 1
      },
      {
        "model": "Sonnet 5",
        "effort": "High",
        "observations": 2,
        "originalCredits": 2289,
        "additionalCredits": 2134,
        "meanFinalCredits": 2211.5,
        "originalCorrections": 0,
        "additionalCorrections": 1
      },
      {
        "model": "Opus 5",
        "effort": "High",
        "observations": 2,
        "originalCredits": 3073,
        "additionalCredits": 2860,
        "meanFinalCredits": 2966.5,
        "originalCorrections": 1,
        "additionalCorrections": 2
      }
    ],
    "matchedConfigurationCount": 3,
    "originalMeanFinalCredits": 1998,
    "meanFinalCredits": 1855.6666666666667,
    "matchedOriginalMeanFinalCredits": 2102,
    "matchedAdditionalMeanFinalCredits": 1855.6666666666667,
    "matchedCombinedMeanFinalCredits": 1978.8333333333333,
    "note": "Completed externally reviewed applications only. Original results remain unchanged. Compare original and additional observations within the same model/effort settings shown.",
    "comparisonNote": "The paired means use exactly the same settings on both sides. The original overall average covers all five settings; do not attribute a different model mix to a model-price change.",
    "promptNote": "All four business stages and the acceptance rubric are unchanged. The additional foundation prompt permits private owner-only publication. This scope difference and two observations per paired setting prevent causal pricing or reliability claims.",
    "stageNote": "Stage readings are cumulative and must not be added together. An increase uses the last reading before that prompt, including any earlier repair; it is not an isolated connector or token price."
  },
  "stressTest": {
    "test": "test-04",
    "route": "test-04",
    "title": "Test 4: capability stress test",
    "reviewMode": "external-only",
    "runs": [
      {
        "id": "run-03",
        "model": "GPT 5.5",
        "effort": "Medium",
        "replicate": 1,
        "firstCompleteTargetCost": {
          "status": "observed",
          "value": 1095,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "finalCost": {
          "status": "observed",
          "value": 1573,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "firstCore": {
          "points": 100,
          "maximum": 100,
          "accepted": true,
          "criticalFailures": [],
          "incomplete": []
        },
        "final": {
          "points": 100,
          "maximum": 100,
          "accepted": true,
          "criticalFailures": [],
          "incomplete": []
        },
        "firstAdvancedPoints": 80,
        "finalAdvancedPoints": 90,
        "incrementalGuidanceCredits": 478,
        "earlyCorrections": 0,
        "laterCorrections": 1,
        "corrections": 1,
        "clarifications": 0,
        "nativeReviews": 0,
        "plannedChanges": 3,
        "timing": {
          "build": {
            "seconds": 2092.087,
            "lowerBoundSeconds": 2078.532,
            "observation": "interval-bounded",
            "authoringPrompts": 5,
            "observedPrompts": 5
          }
        },
        "coreChecks": [
          {
            "id": "TB01",
            "capability": "Same private managed app opens with all eight navigable areas",
            "critical": true,
            "first": "pass",
            "final": "pass"
          },
          {
            "id": "TB02",
            "capability": "App-isolated persistence and exact four-device/two-request seed with correct derived metrics",
            "critical": true,
            "first": "pass",
            "final": "pass"
          },
          {
            "id": "TB03",
            "capability": "Multi-line draft creation, save, quantity/removal/edit and stable unchanged IDs with accurate simple totals",
            "critical": true,
            "first": "pass",
            "final": "pass"
          },
          {
            "id": "TB04",
            "capability": "Required fields, dates, whole quantities, duplicate lines and useful field errors",
            "critical": false,
            "first": "pass",
            "final": "pass"
          },
          {
            "id": "TB05",
            "capability": "Requester draft isolation and an ordinary submission enters the correct manager queue",
            "critical": true,
            "first": "pass",
            "final": "pass"
          },
          {
            "id": "TB06",
            "capability": "Ordinary manager/finance approval and storekeeper fulfilment complete with matching visible stock",
            "critical": true,
            "first": "pass",
            "final": "pass"
          },
          {
            "id": "TB07",
            "capability": "Required rejection reason, correct role actions and visible decision/audit history",
            "critical": false,
            "first": "pass",
            "final": "pass"
          },
          {
            "id": "TB08",
            "capability": "Controller-created draft, unchanged line IDs and simple values survive all stages and native reopen",
            "critical": true,
            "first": "pass",
            "final": "pass"
          },
          {
            "id": "TB09",
            "capability": "Light/dark modes apply consistently and persist without altering business records",
            "critical": false,
            "first": "pass",
            "final": "pass"
          },
          {
            "id": "TB10",
            "capability": "All eight areas and core forms usable at wide desktop and actual 390px without page-wide overflow",
            "critical": true,
            "first": "pass",
            "final": "pass"
          }
        ],
        "advancedChecks": [
          {
            "id": "TA01",
            "capability": "Historical price snapshots and exact integer discount/VAT rounding, including a half-penny tie",
            "critical": false,
            "first": "pass",
            "final": "pass",
            "recheckedAfterCorrection": false
          },
          {
            "id": "TA02",
            "capability": "Pending and approved V1 decisions survive V2; revised/new/boundary requests follow the correct version",
            "critical": false,
            "first": "pass",
            "final": "pass",
            "recheckedAfterCorrection": false
          },
          {
            "id": "TA03",
            "capability": "Self-approval and same-person manager/finance decisions are blocked across synthetic dual roles",
            "critical": false,
            "first": "pass",
            "final": "pass",
            "recheckedAfterCorrection": false
          },
          {
            "id": "TA04",
            "capability": "Budget reservation is atomic at the insufficient-funds boundary and releases correctly on revision/cancellation",
            "critical": false,
            "first": "pass",
            "final": "pass",
            "recheckedAfterCorrection": false
          },
          {
            "id": "TA05",
            "capability": "Partial fulfilment and closing the remainder conserve stock, quantities, reserved/spent value and every penny",
            "critical": false,
            "first": "pass",
            "final": "pass",
            "recheckedAfterCorrection": false
          },
          {
            "id": "TA06",
            "capability": "Multi-line fulfilment with one insufficient-stock line makes no partial business-state changes",
            "critical": false,
            "first": "pass",
            "final": "pass",
            "recheckedAfterCorrection": false
          },
          {
            "id": "TA07",
            "capability": "Double confirmation and replay from a stale tab do not duplicate or overwrite a request operation",
            "critical": false,
            "first": "fail",
            "final": "pass",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TA08",
            "capability": "Two stale request views competing for one remaining department budget cannot both reserve it",
            "critical": false,
            "first": "pass",
            "final": "pass",
            "recheckedAfterCorrection": false
          },
          {
            "id": "TA09",
            "capability": "Price-import preview, cancellation, all-row validation, atomic apply and stale-preview rejection work",
            "critical": false,
            "first": "pass",
            "final": "pass",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TA10",
            "capability": "Keyboard dialogs/focus/errors and sampled control/error contrast meet the specified desktop/mobile targets in both themes",
            "critical": false,
            "first": "fail",
            "final": "untested",
            "recheckedAfterCorrection": true
          }
        ],
        "archived": true,
        "reviewSurface": "Native owner preview of the same privately published app; later edits are not claimed republished.",
        "stages": [
          {
            "id": "stage-01",
            "title": "Isolated relational foundation and stable request editing",
            "cost": {
              "status": "observed",
              "value": 210,
              "unit": "credits",
              "scope": "Cowork task"
            }
          },
          {
            "id": "stage-02",
            "title": "Versioned approvals, budget reservation and partial fulfilment",
            "cost": {
              "status": "observed",
              "value": 431,
              "unit": "credits",
              "scope": "Cowork task"
            }
          },
          {
            "id": "stage-03",
            "title": "Change policy and money rules while preserving existing business state",
            "cost": {
              "status": "observed",
              "value": 677,
              "unit": "credits",
              "scope": "Cowork task"
            }
          },
          {
            "id": "stage-04",
            "title": "Competing updates, atomic imports and accessible interaction",
            "cost": {
              "status": "observed",
              "value": 1095,
              "unit": "credits",
              "scope": "Cowork task"
            }
          }
        ]
      },
      {
        "id": "run-08",
        "model": "Opus 5",
        "effort": "Medium",
        "replicate": 2,
        "firstCompleteTargetCost": {
          "status": "observed",
          "value": 2426,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "finalCost": {
          "status": "observed",
          "value": 2938,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "firstCore": {
          "points": 70,
          "maximum": 100,
          "accepted": false,
          "criticalFailures": [
            "TB05",
            "TB06"
          ],
          "incomplete": []
        },
        "final": {
          "points": 100,
          "maximum": 100,
          "accepted": true,
          "criticalFailures": [],
          "incomplete": []
        },
        "firstAdvancedPoints": 0,
        "finalAdvancedPoints": 70,
        "incrementalGuidanceCredits": 512,
        "earlyCorrections": 1,
        "laterCorrections": 1,
        "corrections": 2,
        "clarifications": 0,
        "nativeReviews": 0,
        "plannedChanges": 3,
        "timing": {
          "build": {
            "seconds": 3098.346,
            "lowerBoundSeconds": 3013.07,
            "observation": "interval-bounded",
            "authoringPrompts": 6,
            "observedPrompts": 6
          }
        },
        "coreChecks": [
          {
            "id": "TB01",
            "capability": "Same private managed app opens with all eight navigable areas",
            "critical": true,
            "first": "pass",
            "final": "pass",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TB02",
            "capability": "App-isolated persistence and exact four-device/two-request seed with correct derived metrics",
            "critical": true,
            "first": "pass",
            "final": "pass",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TB03",
            "capability": "Multi-line draft creation, save, quantity/removal/edit and stable unchanged IDs with accurate simple totals",
            "critical": true,
            "first": "pass",
            "final": "pass",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TB04",
            "capability": "Required fields, dates, whole quantities, duplicate lines and useful field errors",
            "critical": false,
            "first": "pass",
            "final": "pass",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TB05",
            "capability": "Requester draft isolation and an ordinary submission enters the correct manager queue",
            "critical": true,
            "first": "fail",
            "final": "pass",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TB06",
            "capability": "Ordinary manager/finance approval and storekeeper fulfilment complete with matching visible stock",
            "critical": true,
            "first": "fail",
            "final": "pass",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TB07",
            "capability": "Required rejection reason, correct role actions and visible decision/audit history",
            "critical": false,
            "first": "fail",
            "final": "pass",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TB08",
            "capability": "Controller-created draft, unchanged line IDs and simple values survive all stages and native reopen",
            "critical": true,
            "first": "pass",
            "final": "pass",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TB09",
            "capability": "Light/dark modes apply consistently and persist without altering business records",
            "critical": false,
            "first": "pass",
            "final": "pass",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TB10",
            "capability": "All eight areas and core forms usable at wide desktop and actual 390px without page-wide overflow",
            "critical": true,
            "first": "pass",
            "final": "pass",
            "recheckedAfterCorrection": true
          }
        ],
        "advancedChecks": [
          {
            "id": "TA01",
            "capability": "Historical price snapshots and exact integer discount/VAT rounding, including a half-penny tie",
            "critical": false,
            "first": "fail",
            "final": "pass",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TA02",
            "capability": "Pending and approved V1 decisions survive V2; revised/new/boundary requests follow the correct version",
            "critical": false,
            "first": "fail",
            "final": "pass",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TA03",
            "capability": "Self-approval and same-person manager/finance decisions are blocked across synthetic dual roles",
            "critical": false,
            "first": "blocked",
            "final": "pass",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TA04",
            "capability": "Budget reservation is atomic at the insufficient-funds boundary and releases correctly on revision/cancellation",
            "critical": false,
            "first": "blocked",
            "final": "pass",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TA05",
            "capability": "Partial fulfilment and closing the remainder conserve stock, quantities, reserved/spent value and every penny",
            "critical": false,
            "first": "blocked",
            "final": "pass",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TA06",
            "capability": "Multi-line fulfilment with one insufficient-stock line makes no partial business-state changes",
            "critical": false,
            "first": "blocked",
            "final": "pass",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TA07",
            "capability": "Double confirmation and replay from a stale tab do not duplicate or overwrite a request operation",
            "critical": false,
            "first": "fail",
            "final": "blocked",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TA08",
            "capability": "Two stale request views competing for one remaining department budget cannot both reserve it",
            "critical": false,
            "first": "blocked",
            "final": "blocked",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TA09",
            "capability": "Price-import preview, cancellation, all-row validation, atomic apply and stale-preview rejection work",
            "critical": false,
            "first": "fail",
            "final": "pass",
            "recheckedAfterCorrection": true
          },
          {
            "id": "TA10",
            "capability": "Keyboard dialogs/focus/errors and sampled control/error contrast meet the specified desktop/mobile targets in both themes",
            "critical": false,
            "first": "fail",
            "final": "fail",
            "recheckedAfterCorrection": true
          }
        ],
        "archived": true,
        "reviewSurface": "Native owner preview of the same privately published app; later edits are not claimed republished.",
        "stages": [
          {
            "id": "stage-01",
            "title": "Isolated relational foundation and stable request editing",
            "cost": {
              "status": "observed",
              "value": 585,
              "unit": "credits",
              "scope": "Cowork task"
            }
          },
          {
            "id": "stage-02",
            "title": "Versioned approvals, budget reservation and partial fulfilment",
            "cost": {
              "status": "observed",
              "value": 1392,
              "unit": "credits",
              "scope": "Cowork task"
            }
          },
          {
            "id": "stage-03",
            "title": "Change policy and money rules while preserving existing business state",
            "cost": {
              "status": "observed",
              "value": 1928,
              "unit": "credits",
              "scope": "Cowork task"
            }
          },
          {
            "id": "stage-04",
            "title": "Competing updates, atomic imports and accessible interaction",
            "cost": {
              "status": "observed",
              "value": 2426,
              "unit": "credits",
              "scope": "Cowork task"
            }
          }
        ]
      }
    ],
    "meanFinalCredits": 2255.5,
    "note": "All five model/effort settings have actual final task readings. Only accepted, independently reviewed completed apps enter the accepted-app average and capability matrix. Spend-only readings are actual task costs, not accepted-app prices or zero capability scores.",
    "scope": "Eight areas; multi-line virtual requests; versioned approvals; budget reservations; partial fulfilment; exact-pence pricing; stale-view guards; atomic price imports; light/dark desktop and mobile interaction. No live business or profile connector.",
    "scoreNote": "Core acceptance and ten advanced challenges are separate. Each check is worth ten points; every frozen subcase must be verified for an advanced pass. Fail, blocked and untested each earn zero. No combined quality-per-credit score.",
    "guidanceNote": "Four identical planned prompts per app. Early corrections are included in first-complete-target cost; it is not necessarily a no-guidance first pass. Later guidance is final minus first-complete-target credits. Cumulative readings must not be added.",
    "coverageNote": "Every frozen subcase is required for a pass. Automatic cross-tab refresh can prevent an intentionally stale form from remaining available; an unexercised stale prerequisite is a coverage gap, not a demonstrated race defect. The keyboard scenario specifies request search and a rejection-reason dialog, whereas the authoring prompts allow other interfaces. A different interface is not automatically a requested-feature defect. Earlier historical evidence and post-correction rechecks remain distinct; no source or data is altered to force a test precondition.",
    "limitation": "Controlled same-context local simulations and sampled accessibility do not establish production concurrency, real authorization or accessibility certification. The selected Cowork model does not prove every internal app-building worker used that model. Compare only within this test, not against Tests 1-3 scores or averages.",
    "registeredSettings": [
      {
        "runId": "run-03",
        "model": "GPT 5.5",
        "effort": "Medium",
        "cohort": "Original",
        "status": "accepted",
        "finalSpend": {
          "status": "observed",
          "value": 1573,
          "unit": "credits",
          "scope": "Cowork task"
        }
      },
      {
        "runId": "run-06",
        "model": "Sonnet 5",
        "effort": "Medium",
        "status": "not-accepted",
        "finalSpend": {
          "status": "observed",
          "value": 1992,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "cohort": "Follow-up"
      },
      {
        "runId": "run-07",
        "model": "Opus 5",
        "effort": "High",
        "status": "not-accepted",
        "finalSpend": {
          "status": "observed",
          "value": 8144,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "cohort": "Follow-up"
      },
      {
        "runId": "run-08",
        "model": "Opus 5",
        "effort": "Medium",
        "status": "accepted",
        "finalSpend": {
          "status": "observed",
          "value": 2938,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "cohort": "Follow-up"
      },
      {
        "runId": "run-09",
        "model": "Sonnet 5",
        "effort": "High",
        "status": "not-accepted",
        "finalSpend": {
          "status": "observed",
          "value": 4229,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "cohort": "Follow-up"
      }
    ],
    "followupNote": "The original GPT Medium observation is retained alongside four separately registered follow-ups using the identical original prompts, fixtures, rubrics and authoring/correction limits. Original outcomes are not overwritten. All four follow-ups are closed. This is a disclosed follow-up comparison, not five original first attempts; unfinished or unaccepted spend is excluded from accepted-app averages."
  },
  "htmlReferenceTest": {
    "test": "test-05",
    "route": "test-05",
    "cohort": "original-html-reference",
    "title": "Test 5: recreate an HTML reference",
    "reviewMode": "external-only",
    "protocolSha256": "f93b8aaf313977049c80c523ce9cf0ddd75b5b9921bb2073eaece9fe6f99183c",
    "buildPromptSha256": "c3c9e4e9de95e488fde6ecd99c7ba443be63302664b587c86ffc2bb6f1ef17ff",
    "referenceSha256": "b04d8b8bb71e18b2429a4ddf995b698fcaebfbeac1f9926c804f437d256b0d28",
    "referenceBytes": 188501,
    "costMode": "complete-app-only",
    "runs": [
      {
        "id": "run-01",
        "model": "Sonnet 5",
        "effort": "Medium",
        "replicate": 1,
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 2517,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 8,
        "corrections": 7,
        "clarifications": 0,
        "freeCostCommands": 2,
        "final": {
          "accepted": true,
          "functionalPassed": 17,
          "functionalTotal": 20,
          "visualPassed": 8,
          "visualTotal": 10,
          "allCriticalPassed": true
        },
        "functionalChecks": [
          {
            "id": "HF01",
            "capability": "Working managed app, complete role-aware scope and reference fixture",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF02",
            "capability": "Derived dashboards and balances",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF03",
            "capability": "My Leave filters, sorting and empty states",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF04",
            "capability": "Request detail and retained history",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF05",
            "capability": "Four-step request wizard",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF06",
            "capability": "Working-day and half-day arithmetic",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF07",
            "capability": "Blocking validation",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF08",
            "capability": "Warnings and policy-dependent validation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF09",
            "capability": "Draft, edit and duplicate lifecycle",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF10",
            "capability": "Submission, approval routing and notifications",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF11",
            "capability": "Withdraw and cancel with balance reconciliation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF12",
            "capability": "Manager decisions and reasons",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF13",
            "capability": "HR second-step and no-manager routing",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF14",
            "capability": "Bulk decisions, delegation and escalation",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HF15",
            "capability": "Team calendars",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF16",
            "capability": "Scoped reports and actual CSV exports",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HF17",
            "capability": "Live policy and pro-rata calculator",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF18",
            "capability": "Six working HR settings tabs",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF19",
            "capability": "Search, notifications, themes and keyboard interaction",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HF20",
            "capability": "Faithful demo lifecycle and isolation",
            "critical": true,
            "final": "pass"
          }
        ],
        "visualChecks": [
          {
            "id": "HV01",
            "capability": "Reference identity and shell",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HV02",
            "capability": "Dashboard composition",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV03",
            "capability": "Wizard and form hierarchy",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV04",
            "capability": "Tables, filters and status treatment",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV05",
            "capability": "Calendar visual encoding",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV06",
            "capability": "Reports, charts and settings layouts",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HV07",
            "capability": "Light/dark palette and typography",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV08",
            "capability": "Cards, spacing and content density",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HV09",
            "capability": "Dialogs, drawers and feedback",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV10",
            "capability": "Responsive correspondence",
            "critical": true,
            "final": "pass"
          }
        ],
        "timing": {
          "build": {
            "seconds": 5529.797,
            "lowerBoundSeconds": 5453.061,
            "observation": "interval-bounded",
            "authoringPrompts": 8,
            "observedPrompts": 8
          }
        },
        "archived": true
      },
      {
        "id": "run-02",
        "model": "Opus 5",
        "effort": "High",
        "replicate": 1,
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 3197,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 3,
        "corrections": 2,
        "clarifications": 0,
        "freeCostCommands": 1,
        "final": {
          "accepted": true,
          "functionalPassed": 19,
          "functionalTotal": 20,
          "visualPassed": 8,
          "visualTotal": 10,
          "allCriticalPassed": true
        },
        "functionalChecks": [
          {
            "id": "HF01",
            "capability": "Working managed app, complete role-aware scope and reference fixture",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF02",
            "capability": "Derived dashboards and balances",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF03",
            "capability": "My Leave filters, sorting and empty states",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF04",
            "capability": "Request detail and retained history",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF05",
            "capability": "Four-step request wizard",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF06",
            "capability": "Working-day and half-day arithmetic",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF07",
            "capability": "Blocking validation",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF08",
            "capability": "Warnings and policy-dependent validation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF09",
            "capability": "Draft, edit and duplicate lifecycle",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF10",
            "capability": "Submission, approval routing and notifications",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF11",
            "capability": "Withdraw and cancel with balance reconciliation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF12",
            "capability": "Manager decisions and reasons",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF13",
            "capability": "HR second-step and no-manager routing",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF14",
            "capability": "Bulk decisions, delegation and escalation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF15",
            "capability": "Team calendars",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF16",
            "capability": "Scoped reports and actual CSV exports",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HF17",
            "capability": "Live policy and pro-rata calculator",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF18",
            "capability": "Six working HR settings tabs",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF19",
            "capability": "Search, notifications, themes and keyboard interaction",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF20",
            "capability": "Faithful demo lifecycle and isolation",
            "critical": true,
            "final": "pass"
          }
        ],
        "visualChecks": [
          {
            "id": "HV01",
            "capability": "Reference identity and shell",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HV02",
            "capability": "Dashboard composition",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV03",
            "capability": "Wizard and form hierarchy",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV04",
            "capability": "Tables, filters and status treatment",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV05",
            "capability": "Calendar visual encoding",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV06",
            "capability": "Reports, charts and settings layouts",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV07",
            "capability": "Light/dark palette and typography",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HV08",
            "capability": "Cards, spacing and content density",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HV09",
            "capability": "Dialogs, drawers and feedback",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV10",
            "capability": "Responsive correspondence",
            "critical": true,
            "final": "pass"
          }
        ],
        "timing": {
          "build": {
            "seconds": 3179,
            "lowerBoundSeconds": 3140.938,
            "observation": "interval-bounded",
            "authoringPrompts": 3,
            "observedPrompts": 3
          }
        },
        "archived": true
      },
      {
        "id": "run-03",
        "model": "GPT 5.5",
        "effort": "Medium",
        "replicate": 1,
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 659,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 7,
        "corrections": 6,
        "clarifications": 0,
        "freeCostCommands": 1,
        "final": {
          "accepted": true,
          "functionalPassed": 17,
          "functionalTotal": 20,
          "visualPassed": 8,
          "visualTotal": 10,
          "allCriticalPassed": true
        },
        "functionalChecks": [
          {
            "id": "HF01",
            "capability": "Working managed app, complete role-aware scope and reference fixture",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF02",
            "capability": "Derived dashboards and balances",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF03",
            "capability": "My Leave filters, sorting and empty states",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF04",
            "capability": "Request detail and retained history",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF05",
            "capability": "Four-step request wizard",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF06",
            "capability": "Working-day and half-day arithmetic",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF07",
            "capability": "Blocking validation",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF08",
            "capability": "Warnings and policy-dependent validation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF09",
            "capability": "Draft, edit and duplicate lifecycle",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HF10",
            "capability": "Submission, approval routing and notifications",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF11",
            "capability": "Withdraw and cancel with balance reconciliation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF12",
            "capability": "Manager decisions and reasons",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF13",
            "capability": "HR second-step and no-manager routing",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF14",
            "capability": "Bulk decisions, delegation and escalation",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HF15",
            "capability": "Team calendars",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF16",
            "capability": "Scoped reports and actual CSV exports",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF17",
            "capability": "Live policy and pro-rata calculator",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF18",
            "capability": "Six working HR settings tabs",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF19",
            "capability": "Search, notifications, themes and keyboard interaction",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HF20",
            "capability": "Faithful demo lifecycle and isolation",
            "critical": true,
            "final": "pass"
          }
        ],
        "visualChecks": [
          {
            "id": "HV01",
            "capability": "Reference identity and shell",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HV02",
            "capability": "Dashboard composition",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV03",
            "capability": "Wizard and form hierarchy",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV04",
            "capability": "Tables, filters and status treatment",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV05",
            "capability": "Calendar visual encoding",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV06",
            "capability": "Reports, charts and settings layouts",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV07",
            "capability": "Light/dark palette and typography",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HV08",
            "capability": "Cards, spacing and content density",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HV09",
            "capability": "Dialogs, drawers and feedback",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV10",
            "capability": "Responsive correspondence",
            "critical": true,
            "final": "pass"
          }
        ],
        "timing": {
          "build": {
            "seconds": 2909.171,
            "lowerBoundSeconds": 2820.289,
            "observation": "interval-bounded",
            "authoringPrompts": 7,
            "observedPrompts": 7
          }
        },
        "archived": true
      },
      {
        "id": "run-04",
        "model": "Opus 5",
        "effort": "Medium",
        "replicate": 1,
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 3220,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 2,
        "corrections": 1,
        "clarifications": 0,
        "freeCostCommands": 1,
        "final": {
          "accepted": true,
          "functionalPassed": 20,
          "functionalTotal": 20,
          "visualPassed": 8,
          "visualTotal": 10,
          "allCriticalPassed": true
        },
        "functionalChecks": [
          {
            "id": "HF01",
            "capability": "Working managed app, complete role-aware scope and reference fixture",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF02",
            "capability": "Derived dashboards and balances",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF03",
            "capability": "My Leave filters, sorting and empty states",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF04",
            "capability": "Request detail and retained history",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF05",
            "capability": "Four-step request wizard",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF06",
            "capability": "Working-day and half-day arithmetic",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF07",
            "capability": "Blocking validation",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF08",
            "capability": "Warnings and policy-dependent validation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF09",
            "capability": "Draft, edit and duplicate lifecycle",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF10",
            "capability": "Submission, approval routing and notifications",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF11",
            "capability": "Withdraw and cancel with balance reconciliation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF12",
            "capability": "Manager decisions and reasons",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF13",
            "capability": "HR second-step and no-manager routing",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF14",
            "capability": "Bulk decisions, delegation and escalation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF15",
            "capability": "Team calendars",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF16",
            "capability": "Scoped reports and actual CSV exports",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF17",
            "capability": "Live policy and pro-rata calculator",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF18",
            "capability": "Six working HR settings tabs",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF19",
            "capability": "Search, notifications, themes and keyboard interaction",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF20",
            "capability": "Faithful demo lifecycle and isolation",
            "critical": true,
            "final": "pass"
          }
        ],
        "visualChecks": [
          {
            "id": "HV01",
            "capability": "Reference identity and shell",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HV02",
            "capability": "Dashboard composition",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV03",
            "capability": "Wizard and form hierarchy",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV04",
            "capability": "Tables, filters and status treatment",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV05",
            "capability": "Calendar visual encoding",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV06",
            "capability": "Reports, charts and settings layouts",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HV07",
            "capability": "Light/dark palette and typography",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HV08",
            "capability": "Cards, spacing and content density",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV09",
            "capability": "Dialogs, drawers and feedback",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV10",
            "capability": "Responsive correspondence",
            "critical": true,
            "final": "pass"
          }
        ],
        "timing": {
          "build": {
            "seconds": 3060.46,
            "lowerBoundSeconds": 3021.219,
            "observation": "interval-bounded",
            "authoringPrompts": 2,
            "observedPrompts": 2
          }
        },
        "archived": true
      },
      {
        "id": "run-05",
        "model": "Sonnet 5",
        "effort": "High",
        "replicate": 1,
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 2523,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 3,
        "corrections": 2,
        "clarifications": 0,
        "freeCostCommands": 1,
        "final": {
          "accepted": true,
          "functionalPassed": 18,
          "functionalTotal": 20,
          "visualPassed": 8,
          "visualTotal": 10,
          "allCriticalPassed": true
        },
        "functionalChecks": [
          {
            "id": "HF01",
            "capability": "Working managed app, complete role-aware scope and reference fixture",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF02",
            "capability": "Derived dashboards and balances",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF03",
            "capability": "My Leave filters, sorting and empty states",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF04",
            "capability": "Request detail and retained history",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF05",
            "capability": "Four-step request wizard",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF06",
            "capability": "Working-day and half-day arithmetic",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF07",
            "capability": "Blocking validation",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF08",
            "capability": "Warnings and policy-dependent validation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF09",
            "capability": "Draft, edit and duplicate lifecycle",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF10",
            "capability": "Submission, approval routing and notifications",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF11",
            "capability": "Withdraw and cancel with balance reconciliation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF12",
            "capability": "Manager decisions and reasons",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF13",
            "capability": "HR second-step and no-manager routing",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF14",
            "capability": "Bulk decisions, delegation and escalation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF15",
            "capability": "Team calendars",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HF16",
            "capability": "Scoped reports and actual CSV exports",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF17",
            "capability": "Live policy and pro-rata calculator",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF18",
            "capability": "Six working HR settings tabs",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF19",
            "capability": "Search, notifications, themes and keyboard interaction",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HF20",
            "capability": "Faithful demo lifecycle and isolation",
            "critical": true,
            "final": "pass"
          }
        ],
        "visualChecks": [
          {
            "id": "HV01",
            "capability": "Reference identity and shell",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HV02",
            "capability": "Dashboard composition",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV03",
            "capability": "Wizard and form hierarchy",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV04",
            "capability": "Tables, filters and status treatment",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV05",
            "capability": "Calendar visual encoding",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HV06",
            "capability": "Reports, charts and settings layouts",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV07",
            "capability": "Light/dark palette and typography",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HV08",
            "capability": "Cards, spacing and content density",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV09",
            "capability": "Dialogs, drawers and feedback",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV10",
            "capability": "Responsive correspondence",
            "critical": true,
            "final": "pass"
          }
        ],
        "timing": {
          "build": {
            "seconds": 3651.981,
            "lowerBoundSeconds": 3632.183,
            "observation": "interval-bounded",
            "authoringPrompts": 3,
            "observedPrompts": 3
          }
        },
        "archived": true
      }
    ],
    "registeredSettings": [
      {
        "runId": "run-01",
        "model": "Sonnet 5",
        "effort": "Medium",
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 2517,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 8
      },
      {
        "runId": "run-02",
        "model": "Opus 5",
        "effort": "High",
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 3197,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 3
      },
      {
        "runId": "run-03",
        "model": "GPT 5.5",
        "effort": "Medium",
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 659,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 7
      },
      {
        "runId": "run-04",
        "model": "Opus 5",
        "effort": "Medium",
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 3220,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 2
      },
      {
        "runId": "run-05",
        "model": "Sonnet 5",
        "effort": "High",
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 2523,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 3
      }
    ],
    "meanFinalCredits": 2423.2,
    "complete": true,
    "scope": "Annual Leave Manager: role-aware leave requests, approvals, calendars, reports and six HR settings tabs; fictional in-memory data only.",
    "method": "Identical initial prompt and frozen HTML bytes in five fresh tasks. Independent native functional and visual review; targeted repairs in the same app. No arbitrary authoring-time or credit cutoff and no paid self-review.",
    "acceptanceNote": "The frozen minimum requires 17 of 20 functional groups and 8 of 10 visual groups, including every critical group. Passing that minimum does not mean complete parity; all remaining failed groups are shown.",
    "costNote": "Only cumulative task credits after full independent review and joint acceptance. No interim, repair-only or initial-output credit figures. Free cost commands do not count as authoring prompts.",
    "limitation": "One original observation per setting; selected model and effort, not a proven internal worker model or a causal ranking. Guided replay is not included in this original cohort.",
    "privacy": "The exact HTML reference and selected final app-only screenshots are available in the authorized Test 5 showcase. The fixture is fictional. Transcripts, generated source, account/task/resource identifiers and other raw evidence remain private."
  },
  "htmlReferenceRepeats": {
    "test": "test-05",
    "route": "test-05",
    "cohort": "matched-adaptive-repeat",
    "title": "Test 5: matched adaptive repeats",
    "reviewMode": "external-only",
    "protocolSha256": "8ca487b03edfacf06a51444cb384af9488135dc2cc5a7332002c6ba41d3cd018",
    "buildPromptSha256": "c3c9e4e9de95e488fde6ecd99c7ba443be63302664b587c86ffc2bb6f1ef17ff",
    "referenceSha256": "b04d8b8bb71e18b2429a4ddf995b698fcaebfbeac1f9926c804f437d256b0d28",
    "referenceBytes": 188501,
    "costMode": "complete-app-only",
    "runs": [
      {
        "id": "run-06",
        "model": "Sonnet 5",
        "effort": "High",
        "replicate": 2,
        "matchedOriginalRunId": "run-05",
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 3153,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 6,
        "corrections": 5,
        "clarifications": 0,
        "freeCostCommands": 1,
        "final": {
          "accepted": true,
          "functionalPassed": 17,
          "functionalTotal": 20,
          "visualPassed": 8,
          "visualTotal": 10,
          "allCriticalPassed": true
        },
        "functionalChecks": [
          {
            "id": "HF01",
            "capability": "Working managed app, complete role-aware scope and reference fixture",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF02",
            "capability": "Derived dashboards and balances",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF03",
            "capability": "My Leave filters, sorting and empty states",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF04",
            "capability": "Request detail and retained history",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HF05",
            "capability": "Four-step request wizard",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF06",
            "capability": "Working-day and half-day arithmetic",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF07",
            "capability": "Blocking validation",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF08",
            "capability": "Warnings and policy-dependent validation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF09",
            "capability": "Draft, edit and duplicate lifecycle",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HF10",
            "capability": "Submission, approval routing and notifications",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF11",
            "capability": "Withdraw and cancel with balance reconciliation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF12",
            "capability": "Manager decisions and reasons",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF13",
            "capability": "HR second-step and no-manager routing",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF14",
            "capability": "Bulk decisions, delegation and escalation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF15",
            "capability": "Team calendars",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF16",
            "capability": "Scoped reports and actual CSV exports",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF17",
            "capability": "Live policy and pro-rata calculator",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF18",
            "capability": "Six working HR settings tabs",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF19",
            "capability": "Search, notifications, themes and keyboard interaction",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HF20",
            "capability": "Faithful demo lifecycle and isolation",
            "critical": true,
            "final": "pass"
          }
        ],
        "visualChecks": [
          {
            "id": "HV01",
            "capability": "Reference identity and shell",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HV02",
            "capability": "Dashboard composition",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV03",
            "capability": "Wizard and form hierarchy",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV04",
            "capability": "Tables, filters and status treatment",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV05",
            "capability": "Calendar visual encoding",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV06",
            "capability": "Reports, charts and settings layouts",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV07",
            "capability": "Light/dark palette and typography",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HV08",
            "capability": "Cards, spacing and content density",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HV09",
            "capability": "Dialogs, drawers and feedback",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV10",
            "capability": "Responsive correspondence",
            "critical": true,
            "final": "pass"
          }
        ],
        "timing": {
          "build": {
            "seconds": 4536.348,
            "lowerBoundSeconds": null,
            "observation": "upper-bound-only",
            "authoringPrompts": 6,
            "observedPrompts": 6
          }
        },
        "archived": true
      },
      {
        "id": "run-08",
        "model": "GPT 5.5",
        "effort": "Medium",
        "replicate": 2,
        "matchedOriginalRunId": "run-03",
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 7464,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 16,
        "corrections": 8,
        "clarifications": 7,
        "freeCostCommands": 1,
        "final": {
          "accepted": true,
          "functionalPassed": 20,
          "functionalTotal": 20,
          "visualPassed": 8,
          "visualTotal": 10,
          "allCriticalPassed": true
        },
        "functionalChecks": [
          {
            "id": "HF01",
            "capability": "Working managed app, complete role-aware scope and reference fixture",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF02",
            "capability": "Derived dashboards and balances",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF03",
            "capability": "My Leave filters, sorting and empty states",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF04",
            "capability": "Request detail and retained history",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF05",
            "capability": "Four-step request wizard",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF06",
            "capability": "Working-day and half-day arithmetic",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF07",
            "capability": "Blocking validation",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF08",
            "capability": "Warnings and policy-dependent validation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF09",
            "capability": "Draft, edit and duplicate lifecycle",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF10",
            "capability": "Submission, approval routing and notifications",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF11",
            "capability": "Withdraw and cancel with balance reconciliation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF12",
            "capability": "Manager decisions and reasons",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF13",
            "capability": "HR second-step and no-manager routing",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF14",
            "capability": "Bulk decisions, delegation and escalation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF15",
            "capability": "Team calendars",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF16",
            "capability": "Scoped reports and actual CSV exports",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF17",
            "capability": "Live policy and pro-rata calculator",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF18",
            "capability": "Six working HR settings tabs",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF19",
            "capability": "Search, notifications, themes and keyboard interaction",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF20",
            "capability": "Faithful demo lifecycle and isolation",
            "critical": true,
            "final": "pass"
          }
        ],
        "visualChecks": [
          {
            "id": "HV01",
            "capability": "Reference identity and shell",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HV02",
            "capability": "Dashboard composition",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV03",
            "capability": "Wizard and form hierarchy",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV04",
            "capability": "Tables, filters and status treatment",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV05",
            "capability": "Calendar visual encoding",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV06",
            "capability": "Reports, charts and settings layouts",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HV07",
            "capability": "Light/dark palette and typography",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV08",
            "capability": "Cards, spacing and content density",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HV09",
            "capability": "Dialogs, drawers and feedback",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV10",
            "capability": "Responsive correspondence",
            "critical": true,
            "final": "pass"
          }
        ],
        "timing": {
          "build": {
            "seconds": 8266.903,
            "lowerBoundSeconds": null,
            "observation": "upper-bound-only",
            "authoringPrompts": 16,
            "observedPrompts": 16
          }
        },
        "archived": true
      },
      {
        "id": "run-10",
        "model": "Sonnet 5",
        "effort": "Medium",
        "replicate": 2,
        "matchedOriginalRunId": "run-01",
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 3309,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 3,
        "corrections": 2,
        "clarifications": 0,
        "freeCostCommands": 1,
        "final": {
          "accepted": true,
          "functionalPassed": 18,
          "functionalTotal": 20,
          "visualPassed": 8,
          "visualTotal": 10,
          "allCriticalPassed": true
        },
        "functionalChecks": [
          {
            "id": "HF01",
            "capability": "Working managed app, complete role-aware scope and reference fixture",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF02",
            "capability": "Derived dashboards and balances",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF03",
            "capability": "My Leave filters, sorting and empty states",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF04",
            "capability": "Request detail and retained history",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF05",
            "capability": "Four-step request wizard",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF06",
            "capability": "Working-day and half-day arithmetic",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF07",
            "capability": "Blocking validation",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF08",
            "capability": "Warnings and policy-dependent validation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF09",
            "capability": "Draft, edit and duplicate lifecycle",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF10",
            "capability": "Submission, approval routing and notifications",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF11",
            "capability": "Withdraw and cancel with balance reconciliation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF12",
            "capability": "Manager decisions and reasons",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF13",
            "capability": "HR second-step and no-manager routing",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF14",
            "capability": "Bulk decisions, delegation and escalation",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF15",
            "capability": "Team calendars",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF16",
            "capability": "Scoped reports and actual CSV exports",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HF17",
            "capability": "Live policy and pro-rata calculator",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HF18",
            "capability": "Six working HR settings tabs",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HF19",
            "capability": "Search, notifications, themes and keyboard interaction",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HF20",
            "capability": "Faithful demo lifecycle and isolation",
            "critical": true,
            "final": "pass"
          }
        ],
        "visualChecks": [
          {
            "id": "HV01",
            "capability": "Reference identity and shell",
            "critical": true,
            "final": "pass"
          },
          {
            "id": "HV02",
            "capability": "Dashboard composition",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV03",
            "capability": "Wizard and form hierarchy",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV04",
            "capability": "Tables, filters and status treatment",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV05",
            "capability": "Calendar visual encoding",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV06",
            "capability": "Reports, charts and settings layouts",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV07",
            "capability": "Light/dark palette and typography",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HV08",
            "capability": "Cards, spacing and content density",
            "critical": false,
            "final": "fail"
          },
          {
            "id": "HV09",
            "capability": "Dialogs, drawers and feedback",
            "critical": false,
            "final": "pass"
          },
          {
            "id": "HV10",
            "capability": "Responsive correspondence",
            "critical": true,
            "final": "pass"
          }
        ],
        "timing": {
          "build": {
            "seconds": 4992.702,
            "lowerBoundSeconds": 4962.343,
            "observation": "interval-bounded",
            "authoringPrompts": 3,
            "observedPrompts": 3
          }
        },
        "archived": true
      }
    ],
    "registeredSettings": [
      {
        "runId": "run-06",
        "model": "Sonnet 5",
        "effort": "High",
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 3153,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 6
      },
      {
        "runId": "run-07",
        "model": "Opus 5",
        "effort": "Medium",
        "status": "pending",
        "finalCost": null,
        "authoringPrompts": null
      },
      {
        "runId": "run-08",
        "model": "GPT 5.5",
        "effort": "Medium",
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 7464,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 16
      },
      {
        "runId": "run-09",
        "model": "Opus 5",
        "effort": "High",
        "status": "pending",
        "finalCost": null,
        "authoringPrompts": null
      },
      {
        "runId": "run-10",
        "model": "Sonnet 5",
        "effort": "Medium",
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 3309,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 3
      }
    ],
    "meanFinalCredits": 4642,
    "complete": false,
    "scope": "Annual Leave Manager: role-aware leave requests, approvals, calendars, reports and six HR settings tabs; fictional in-memory data only.",
    "method": "Identical initial prompt and frozen HTML bytes in five fresh tasks. Independent native functional and visual review; targeted repairs in the same app. No arbitrary authoring-time or credit cutoff and no paid self-review.",
    "acceptanceNote": "The frozen minimum requires 17 of 20 functional groups and 8 of 10 visual groups, including every critical group. Passing that minimum does not mean complete parity; all remaining failed groups are shown.",
    "costNote": "Only cumulative task credits after full independent review and joint acceptance. No interim, repair-only or initial-output credit figures. Free cost commands do not count as authoring prompts.",
    "limitation": "One additional adaptive observation per original setting, in reverse order. Two observations remain exploratory; compare both delivered capability results. The separately guided replay is not pooled with these runs.",
    "privacy": "Only allowlisted measurements and capability outcomes are public. All repeat screenshots, transcripts, generated source and identifiers remain private. The original showcase permission is not extended.",
    "comparison": {
      "method": "matched-adaptive-repeat",
      "completePairs": 3,
      "complete": false,
      "pairs": [
        {
          "model": "Sonnet 5",
          "effort": "Medium",
          "originalRunId": "run-01",
          "repeatRunId": "run-10",
          "observations": 2,
          "completePair": true,
          "credits": {
            "original": 2517,
            "repeat": 3309,
            "mean": 2913,
            "minimum": 2517,
            "maximum": 3309
          },
          "authoringPrompts": {
            "original": 8,
            "repeat": 3,
            "mean": 5.5,
            "minimum": 3,
            "maximum": 8
          },
          "capabilityChanges": {
            "functional": [
              {
                "id": "HF14",
                "original": "fail",
                "repeat": "pass"
              }
            ],
            "visual": [
              {
                "id": "HV06",
                "original": "fail",
                "repeat": "pass"
              },
              {
                "id": "HV07",
                "original": "pass",
                "repeat": "fail"
              }
            ]
          }
        },
        {
          "model": "Opus 5",
          "effort": "High",
          "originalRunId": "run-02",
          "repeatRunId": null,
          "observations": 1,
          "completePair": false,
          "credits": {
            "original": 3197,
            "repeat": null,
            "mean": null,
            "minimum": null,
            "maximum": null
          },
          "authoringPrompts": {
            "original": 3,
            "repeat": null,
            "mean": null,
            "minimum": null,
            "maximum": null
          },
          "capabilityChanges": {
            "functional": null,
            "visual": null
          }
        },
        {
          "model": "GPT 5.5",
          "effort": "Medium",
          "originalRunId": "run-03",
          "repeatRunId": "run-08",
          "observations": 2,
          "completePair": true,
          "credits": {
            "original": 659,
            "repeat": 7464,
            "mean": 4061.5,
            "minimum": 659,
            "maximum": 7464
          },
          "authoringPrompts": {
            "original": 7,
            "repeat": 16,
            "mean": 11.5,
            "minimum": 7,
            "maximum": 16
          },
          "capabilityChanges": {
            "functional": [
              {
                "id": "HF09",
                "original": "fail",
                "repeat": "pass"
              },
              {
                "id": "HF14",
                "original": "fail",
                "repeat": "pass"
              },
              {
                "id": "HF19",
                "original": "fail",
                "repeat": "pass"
              }
            ],
            "visual": [
              {
                "id": "HV06",
                "original": "pass",
                "repeat": "fail"
              },
              {
                "id": "HV07",
                "original": "fail",
                "repeat": "pass"
              }
            ]
          }
        },
        {
          "model": "Opus 5",
          "effort": "Medium",
          "originalRunId": "run-04",
          "repeatRunId": null,
          "observations": 1,
          "completePair": false,
          "credits": {
            "original": 3220,
            "repeat": null,
            "mean": null,
            "minimum": null,
            "maximum": null
          },
          "authoringPrompts": {
            "original": 2,
            "repeat": null,
            "mean": null,
            "minimum": null,
            "maximum": null
          },
          "capabilityChanges": {
            "functional": null,
            "visual": null
          }
        },
        {
          "model": "Sonnet 5",
          "effort": "High",
          "originalRunId": "run-05",
          "repeatRunId": "run-06",
          "observations": 2,
          "completePair": true,
          "credits": {
            "original": 2523,
            "repeat": 3153,
            "mean": 2838,
            "minimum": 2523,
            "maximum": 3153
          },
          "authoringPrompts": {
            "original": 3,
            "repeat": 6,
            "mean": 4.5,
            "minimum": 3,
            "maximum": 6
          },
          "capabilityChanges": {
            "functional": [
              {
                "id": "HF04",
                "original": "pass",
                "repeat": "fail"
              },
              {
                "id": "HF09",
                "original": "pass",
                "repeat": "fail"
              },
              {
                "id": "HF15",
                "original": "fail",
                "repeat": "pass"
              }
            ],
            "visual": [
              {
                "id": "HV05",
                "original": "fail",
                "repeat": "pass"
              },
              {
                "id": "HV08",
                "original": "pass",
                "repeat": "fail"
              }
            ]
          }
        }
      ],
      "limitation": "Two observations per setting are exploratory, not a stable ranking or guaranteed budget. The acceptance target is fixed, but passed capability groups can differ; inspect both results. Guided replay is excluded."
    }
  },
  "newModelSuite": {
    "id": "opus55-suite-20260922",
    "cohort": "new-model-adaptive",
    "model": "Opus 5.5",
    "effort": "Medium",
    "reviewMode": "external-only",
    "costMode": "complete-app-only",
    "registeredBuilds": 6,
    "reportedBuilds": 5,
    "reportingScope": {
      "includedTests": [
        "test-05",
        "test-01",
        "test-02",
        "test-03"
      ],
      "omittedIncompleteTests": [
        "test-04"
      ]
    },
    "completedBuilds": 5,
    "registeredSettings": [
      {
        "test": "test-05",
        "runId": "run-01",
        "model": "Opus 5.5",
        "effort": "Medium",
        "replicate": 1,
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 365,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 2
      },
      {
        "test": "test-01",
        "runId": "run-02",
        "model": "Opus 5.5",
        "effort": "Medium",
        "replicate": 1,
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 159,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 1
      },
      {
        "test": "test-02",
        "runId": "run-03",
        "model": "Opus 5.5",
        "effort": "Medium",
        "replicate": 1,
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 610,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 3
      },
      {
        "test": "test-02",
        "runId": "run-04",
        "model": "Opus 5.5",
        "effort": "Medium",
        "replicate": 2,
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 419,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 2
      },
      {
        "test": "test-03",
        "runId": "run-05",
        "model": "Opus 5.5",
        "effort": "Medium",
        "replicate": 1,
        "status": "accepted",
        "finalCost": {
          "status": "observed",
          "value": 920,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 4
      }
    ],
    "cohorts": [
      {
        "test": "test-05",
        "title": "HTML reference recreation",
        "runs": [
          {
            "id": "run-01",
            "model": "Opus 5.5",
            "effort": "Medium",
            "replicate": 1,
            "status": "accepted",
            "finalCost": {
              "status": "observed",
              "value": 365,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "authoringPrompts": 2,
            "corrections": 1,
            "clarifications": 0,
            "freeCostCommands": 1,
            "final": {
              "accepted": true,
              "functionalPassed": 20,
              "functionalTotal": 20,
              "visualPassed": 10,
              "visualTotal": 10,
              "allCriticalPassed": true
            },
            "functionalChecks": [
              {
                "id": "HF01",
                "capability": "Working managed app, complete role-aware scope and reference fixture",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "HF02",
                "capability": "Derived dashboards and balances",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "HF03",
                "capability": "My Leave filters, sorting and empty states",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "HF04",
                "capability": "Request detail and retained history",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "HF05",
                "capability": "Four-step request wizard",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "HF06",
                "capability": "Working-day and half-day arithmetic",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "HF07",
                "capability": "Blocking validation",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "HF08",
                "capability": "Warnings and policy-dependent validation",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "HF09",
                "capability": "Draft, edit and duplicate lifecycle",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "HF10",
                "capability": "Submission, approval routing and notifications",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "HF11",
                "capability": "Withdraw and cancel with balance reconciliation",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "HF12",
                "capability": "Manager decisions and reasons",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "HF13",
                "capability": "HR second-step and no-manager routing",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "HF14",
                "capability": "Bulk decisions, delegation and escalation",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "HF15",
                "capability": "Team calendars",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "HF16",
                "capability": "Scoped reports and actual CSV exports",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "HF17",
                "capability": "Live policy and pro-rata calculator",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "HF18",
                "capability": "Six working HR settings tabs",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "HF19",
                "capability": "Search, notifications, themes and keyboard interaction",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "HF20",
                "capability": "Faithful demo lifecycle and isolation",
                "critical": true,
                "final": "pass"
              }
            ],
            "visualChecks": [
              {
                "id": "HV01",
                "capability": "Reference identity and shell",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "HV02",
                "capability": "Dashboard composition",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "HV03",
                "capability": "Wizard and form hierarchy",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "HV04",
                "capability": "Tables, filters and status treatment",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "HV05",
                "capability": "Calendar visual encoding",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "HV06",
                "capability": "Reports, charts and settings layouts",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "HV07",
                "capability": "Light/dark palette and typography",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "HV08",
                "capability": "Cards, spacing and content density",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "HV09",
                "capability": "Dialogs, drawers and feedback",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "HV10",
                "capability": "Responsive correspondence",
                "critical": true,
                "final": "pass"
              }
            ],
            "timing": {
              "build": {
                "seconds": 633.225,
                "lowerBoundSeconds": 608.26,
                "observation": "interval-bounded",
                "authoringPrompts": 2,
                "observedPrompts": 2
              }
            },
            "archived": true
          }
        ],
        "registeredBuilds": 1,
        "complete": true,
        "meanFinalCredits": 365,
        "strategy": "Reused the supplied reference screens and business logic. One targeted correction fixed inherited theme styles; final review verified all functional and visual groups.",
        "strategyCaveat": "Reuse was observed, not separately randomized. This does not isolate a model-only or reuse-only saving."
      },
      {
        "test": "test-01",
        "title": "Virtual procurement",
        "runs": [
          {
            "id": "run-02",
            "model": "Opus 5.5",
            "effort": "Medium",
            "replicate": 1,
            "status": "accepted",
            "finalCost": {
              "status": "observed",
              "value": 159,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "authoringPrompts": 1,
            "corrections": 0,
            "clarifications": 0,
            "final": {
              "accepted": true,
              "functionalPassed": 20,
              "functionalTotal": 20,
              "allCriticalPassed": true
            },
            "functionalChecks": [
              {
                "id": "AC01",
                "capability": "Exact catalogue",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC02",
                "capability": "Search and empty state",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC03",
                "capability": "Combined filtering",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC04",
                "capability": "Unavailable stock",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC05",
                "capability": "Dashboard arithmetic",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC06",
                "capability": "Valid request",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC07",
                "capability": "Required fields",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC08",
                "capability": "Quantity validation",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC09",
                "capability": "Date validation",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC10",
                "capability": "Approval",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC11",
                "capability": "Rejection",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC12",
                "capability": "Oversell prevention",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC13",
                "capability": "Single state transition",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC14",
                "capability": "Virtual/local reopen persistence",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "AC15",
                "capability": "Requester scoping",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC16",
                "capability": "Safe repeatable reset",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC17",
                "capability": "Keyboard and labels",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC18",
                "capability": "Responsive layout",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC19",
                "capability": "Honest demo boundary",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "AC20",
                "capability": "Actual private managed app",
                "critical": true,
                "final": "pass"
              }
            ],
            "timing": {
              "build": {
                "seconds": 361.767,
                "lowerBoundSeconds": 350.923,
                "observation": "interval-bounded",
                "authoringPrompts": 1,
                "observedPrompts": 1
              }
            },
            "archived": true,
            "reviewedRevisionScope": "owner-preview"
          }
        ],
        "registeredBuilds": 1,
        "complete": true,
        "meanFinalCredits": 159
      },
      {
        "test": "test-02",
        "title": "Live SharePoint requests",
        "runs": [
          {
            "id": "run-03",
            "model": "Opus 5.5",
            "effort": "Medium",
            "replicate": 1,
            "status": "accepted",
            "finalCost": {
              "status": "observed",
              "value": 610,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "authoringPrompts": 3,
            "corrections": 1,
            "clarifications": 1,
            "final": {
              "accepted": true,
              "functionalPassed": 10,
              "functionalTotal": 10,
              "allCriticalPassed": true
            },
            "functionalChecks": [
              {
                "id": "BC01",
                "capability": "Runnable private managed app",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC02",
                "capability": "Live catalogue connection",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC03",
                "capability": "Live dashboard arithmetic",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC04",
                "capability": "Catalogue search and filtering",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC05",
                "capability": "Complete saved request",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC06",
                "capability": "Basic request validation",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC07",
                "capability": "Live request listing",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC08",
                "capability": "Reopen persistence",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC09",
                "capability": "Honest data and error states",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC10",
                "capability": "Usable desktop and mobile build",
                "critical": false,
                "final": "pass"
              }
            ],
            "timing": {
              "build": {
                "seconds": null,
                "status": "observer-recovery",
                "note": "Completion confirmation was delayed by the external observer; no model-speed comparison is reported."
              }
            },
            "archived": true,
            "reviewedRevisionScope": "owner-preview"
          },
          {
            "id": "run-04",
            "model": "Opus 5.5",
            "effort": "Medium",
            "replicate": 2,
            "status": "accepted",
            "finalCost": {
              "status": "observed",
              "value": 419,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "authoringPrompts": 2,
            "corrections": 1,
            "clarifications": 0,
            "final": {
              "accepted": true,
              "functionalPassed": 10,
              "functionalTotal": 10,
              "allCriticalPassed": true
            },
            "functionalChecks": [
              {
                "id": "BC01",
                "capability": "Runnable private managed app",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC02",
                "capability": "Live catalogue connection",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC03",
                "capability": "Live dashboard arithmetic",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC04",
                "capability": "Catalogue search and filtering",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC05",
                "capability": "Complete saved request",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC06",
                "capability": "Basic request validation",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC07",
                "capability": "Live request listing",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC08",
                "capability": "Reopen persistence",
                "critical": true,
                "final": "pass"
              },
              {
                "id": "BC09",
                "capability": "Honest data and error states",
                "critical": false,
                "final": "pass"
              },
              {
                "id": "BC10",
                "capability": "Usable desktop and mobile build",
                "critical": false,
                "final": "pass"
              }
            ],
            "timing": {
              "build": {
                "seconds": 683.114,
                "lowerBoundSeconds": 656.306,
                "observation": "interval-bounded",
                "authoringPrompts": 2,
                "observedPrompts": 2
              }
            },
            "archived": true,
            "reviewedRevisionScope": "owner-preview"
          }
        ],
        "registeredBuilds": 2,
        "complete": true,
        "meanFinalCredits": 514.5
      },
      {
        "test": "test-03",
        "title": "Complex lifecycle",
        "runs": [
          {
            "id": "run-05",
            "model": "Opus 5.5",
            "effort": "Medium",
            "replicate": 1,
            "status": "accepted",
            "finalCost": {
              "status": "observed",
              "value": 920,
              "unit": "credits",
              "scope": "Cowork task"
            },
            "authoringPrompts": 4,
            "corrections": 0,
            "clarifications": 0,
            "plannedStages": 4,
            "plannedChanges": 3,
            "firstCompleteTarget": {
              "authoringPrompts": 4,
              "functionalPassed": 20,
              "functionalTotal": 20,
              "cost": null
            },
            "final": {
              "accepted": true,
              "functionalPassed": 20,
              "functionalTotal": 20,
              "allCriticalPassed": true
            },
            "functionalChecks": [
              {
                "id": "CX01",
                "capability": "Same private app opens with all seven navigable screens",
                "critical": true,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX02",
                "capability": "Exact six-device/four-request relational seed and derived initial metrics",
                "critical": true,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX03",
                "capability": "Catalogue search, combined category/stock filters and zero-stock protection",
                "critical": false,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX04",
                "capability": "Multi-line draft creation, line edits/removal and accurate GBP totals",
                "critical": true,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX05",
                "capability": "Required fields, dates, whole positive quantities, duplicates and stock validation",
                "critical": true,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX06",
                "capability": "Draft/persona isolation, edits and exact-once submission with stable IDs",
                "critical": true,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX07",
                "capability": "Controller-created draft, IDs, lines and timeline survive all three planned changes and reopen",
                "critical": true,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX08",
                "capability": "Virtual manager/finance role gates and approval transitions at the final threshold",
                "critical": true,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX09",
                "capability": "Rejection comments, append-only actor/timestamp history and terminal action guards",
                "critical": false,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX10",
                "capability": "Request changes, requester edit/resubmit and cancellation preserve identity/history",
                "critical": false,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX11",
                "capability": "Atomic fulfilment, stock revalidation, per-unit assets and duplicate-fulfilment protection",
                "critical": true,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX12",
                "capability": "Good/damaged asset returns and duplicate-return protection",
                "critical": false,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX13",
                "capability": "Current-user Office 365 Users profile actually reads through supported app binding",
                "critical": true,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX14",
                "capability": "Profile loading/error/retry and genuinely missing optional fields are honest",
                "critical": false,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX15",
                "capability": "Explicit light/dark mode applies across all screens and persists on reopen",
                "critical": true,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX16",
                "capability": "Local preference Save/Cancel, density and default department work without directory writes",
                "critical": false,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX17",
                "capability": "Cost-centre migration, persistence and department mismatch validation",
                "critical": false,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX18",
                "capability": "Derived status/asset/stock metrics, low-stock navigation and combined request filters",
                "critical": false,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX19",
                "capability": "All screens and key forms are usable at actual wide desktop and 390px mobile widths",
                "critical": true,
                "first": "pass",
                "final": "pass"
              },
              {
                "id": "CX20",
                "capability": "Labels, visible keyboard focus, readable contrast and reachable form errors",
                "critical": false,
                "first": "pass",
                "final": "pass"
              }
            ],
            "timing": {
              "build": {
                "seconds": 1473.303,
                "lowerBoundSeconds": 1379.479,
                "observation": "interval-bounded",
                "authoringPrompts": 4,
                "observedPrompts": 4
              }
            },
            "archived": true,
            "reviewedRevisionScope": "owner-preview"
          }
        ],
        "registeredBuilds": 1,
        "complete": true,
        "meanFinalCredits": 920,
        "reviewCaveat": "All optional profile values were present, so genuinely missing-value behavior was not exercised. One native low-stock review justification differed only by a hyphen; all numeric, date, persona and workflow inputs matched. The frozen authoring prompts were unchanged."
      }
    ],
    "method": "Six fresh tasks were originally registered: Test 5 first, then Tests 1–4, with two SharePoint builds. The original frozen business inputs and scope-specific criteria were retained. Medium was the observed native default. Independent review, adaptive repairs in the same app and one actual completed-app cost. Reporting is limited to the requested scopes; omitted incomplete builds contribute no cost, prompt total or capability result.",
    "limitation": "New model version, not an Opus 5 repeat. No pooled cross-version or cross-scope average. One observation per scope, or two SharePoint observations when complete, is exploratory rather than a stable price or causal ranking.",
    "privacy": "Only allowlisted measurements and capability verdicts are public. New-version screenshots, transcripts, generated source and account/resource identifiers remain private."
  },
  "testOneRepeatPair": {
    "test": "test-01",
    "cohort": "same-tenant-adaptive-pair",
    "model": "GPT 5.5",
    "effort": "Medium",
    "date": "2026-09-28",
    "sampleSize": 2,
    "complete": true,
    "reviewMode": "external-only",
    "costMode": "complete-app-only",
    "tenantScope": "One demo tenant",
    "runs": [
      {
        "id": "run-01",
        "model": "GPT 5.5",
        "effort": "Medium",
        "replicate": 1,
        "status": "accepted",
        "final": {
          "points": 100,
          "accepted": true,
          "allCriticalPassed": true,
          "passed": 20,
          "total": 20
        },
        "finalCost": {
          "status": "observed",
          "value": 281,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 2,
        "corrections": 1,
        "clarifications": 0,
        "reviews": 0,
        "archived": true,
        "timing": {
          "build": {
            "seconds": 866.943,
            "lowerBoundSeconds": 832.146,
            "observation": "interval-bounded",
            "authoringPrompts": 2,
            "observedPrompts": 2
          }
        },
        "responsiveLayout": {
          "status": "pass",
          "overflowScenes": []
        },
        "firstOutput": {
          "cost": null,
          "passed": 0,
          "incomplete": 19
        },
        "checks": [
          {
            "id": "AC01",
            "capability": "Exact catalogue",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC02",
            "capability": "Search and empty state",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC03",
            "capability": "Combined filtering",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC04",
            "capability": "Unavailable stock",
            "critical": true,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC05",
            "capability": "Dashboard arithmetic",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC06",
            "capability": "Valid request",
            "critical": true,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC07",
            "capability": "Required fields",
            "critical": true,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC08",
            "capability": "Quantity validation",
            "critical": true,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC09",
            "capability": "Date validation",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC10",
            "capability": "Approval",
            "critical": true,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC11",
            "capability": "Rejection",
            "critical": true,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC12",
            "capability": "Oversell prevention",
            "critical": true,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC13",
            "capability": "Single state transition",
            "critical": true,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC14",
            "capability": "Virtual/local reopen persistence",
            "critical": true,
            "first": "fail",
            "final": "pass"
          },
          {
            "id": "AC15",
            "capability": "Requester scoping",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC16",
            "capability": "Safe repeatable reset",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC17",
            "capability": "Keyboard and labels",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC18",
            "capability": "Responsive layout",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC19",
            "capability": "Honest demo boundary",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC20",
            "capability": "Actual private managed app",
            "critical": true,
            "first": "blocked",
            "final": "pass"
          }
        ],
        "reviewedRevisionScope": "owner-preview",
        "latestRevisionRepublished": false
      },
      {
        "id": "run-02",
        "model": "GPT 5.5",
        "effort": "Medium",
        "replicate": 2,
        "status": "accepted",
        "final": {
          "points": 95,
          "accepted": true,
          "allCriticalPassed": true,
          "passed": 19,
          "total": 20
        },
        "finalCost": {
          "status": "observed",
          "value": 221,
          "unit": "credits",
          "scope": "Cowork task"
        },
        "authoringPrompts": 2,
        "corrections": 1,
        "clarifications": 0,
        "reviews": 0,
        "archived": true,
        "timing": {
          "build": {
            "seconds": 513.598,
            "lowerBoundSeconds": 466.893,
            "observation": "interval-bounded",
            "authoringPrompts": 2,
            "observedPrompts": 2
          }
        },
        "responsiveLayout": {
          "status": "fail",
          "overflowScenes": [
            {
              "view": "Dashboard",
              "viewportWidth": 390,
              "contentWidth": 669
            }
          ]
        },
        "firstOutput": {
          "cost": null,
          "passed": 0,
          "incomplete": 19
        },
        "checks": [
          {
            "id": "AC01",
            "capability": "Exact catalogue",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC02",
            "capability": "Search and empty state",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC03",
            "capability": "Combined filtering",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC04",
            "capability": "Unavailable stock",
            "critical": true,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC05",
            "capability": "Dashboard arithmetic",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC06",
            "capability": "Valid request",
            "critical": true,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC07",
            "capability": "Required fields",
            "critical": true,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC08",
            "capability": "Quantity validation",
            "critical": true,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC09",
            "capability": "Date validation",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC10",
            "capability": "Approval",
            "critical": true,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC11",
            "capability": "Rejection",
            "critical": true,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC12",
            "capability": "Oversell prevention",
            "critical": true,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC13",
            "capability": "Single state transition",
            "critical": true,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC14",
            "capability": "Virtual/local reopen persistence",
            "critical": true,
            "first": "fail",
            "final": "pass"
          },
          {
            "id": "AC15",
            "capability": "Requester scoping",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC16",
            "capability": "Safe repeatable reset",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC17",
            "capability": "Keyboard and labels",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC18",
            "capability": "Responsive layout",
            "critical": false,
            "first": "blocked",
            "final": "fail"
          },
          {
            "id": "AC19",
            "capability": "Honest demo boundary",
            "critical": false,
            "first": "blocked",
            "final": "pass"
          },
          {
            "id": "AC20",
            "capability": "Actual private managed app",
            "critical": true,
            "first": "blocked",
            "final": "pass"
          }
        ],
        "reviewedRevisionScope": "owner-preview",
        "latestRevisionRepublished": false
      }
    ],
    "meanFinalCredits": 251,
    "minimumCredits": 221,
    "maximumCredits": 281,
    "rangeCredits": 60,
    "note": "Two fresh apps with the identical Test 1 business prompt and setting. Each final cumulative cost includes all guidance; no paid self-review or interim cost.",
    "limitation": "Separate from historical cohorts: review and persistence scope differ from original Test 1. Two observations do not establish stable pricing, a confidence interval or model superiority. Owner preview was reviewed; no separately verified deployment is claimed."
  }
}