NodeRoom/docs/eval/proofloop-prod-browser-adapters.json at main · HomenShum/NodeRoom · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
{
  "schema": "proofloop-prod-browser-adapter-ledger-v1",
  "generatedAt": "2026-07-05T08:13:18.817Z",
  "harnessVersion": "prod-browser-adapters-2026-07-05.4",
  "matrixModelCount": 4,
  "summary": {
    "adaptersTracked": 6,
    "contractScaffolded": 0,
    "browserScenarioMissing": 0,
    "taskTargetsCoveredByContracts": 1251,
    "modelTaskAttemptsCoveredByContracts": 5004
  },
  "adapters": [
    {
      "id": "spreadsheetbench-v1-official-workbook-prod-browser",
      "familyId": "spreadsheetbench-v1-full-912",
      "version": "0.2.0",
      "status": "browser_scenario_ready",
      "browserScenarioStatus": "ready",
      "taskCount": 912,
      "attemptCountAtCurrentMatrix": 3648,
      "sourceBenchmark": "SpreadsheetBench V1 full 912",
      "taskLoader": ".tmp/official-benchmarks/staged-v1-912/tasks/<taskId>/agent/task.json",
      "browserScenario": "e2e/benchmark-ui-spreadsheetbench-generic.spec.ts",
      "commandShape": "SPREADSHEETBENCH_TASK_ID=<taskId> BENCH_AGENT_MODEL_POLICY=<modelId> npm run proofloop:live:spreadsheetbench-v1",
      "existingSmoke": "e2e/benchmark-ui-spreadsheetbench.spec.ts proves one fixed nb-01 style browser path; e2e/benchmark-ui-spreadsheetbench-generic.spec.ts is the generic staged-task adapter.",
      "requiredArtifacts": [
        "live-user-contract.json",
        "node-trace-v2.json",
        "node-eval.json",
        "scorecard.md",
        "cost-ledger.json",
        "verifier-receipt.json",
        "visual-proof.png",
        "spreadsheetbench-v1-official-score.json",
        "exported-workbook.xlsx"
      ],
      "scorerBoundary": "Official workbook scorer/golden metadata must remain evaluator-only until after NodeRoom exports candidate output.",
      "blockers": [
        "No passing prod receipts are recorded for every staged V1 task/model yet; run the long-run queue to replace this with score evidence."
      ],
      "changelog": [
        "2026-07-05.1: contract created; existing one-task smoke is explicitly not promoted to the full 912-task adapter.",
        "2026-07-05.2: generic staged-task browser scenario added; tasks are runnable but not scored until receipts exist."
      ]
    },
    {
      "id": "spreadsheetbench-v2-workflow-chart-prod-browser",
      "familyId": "spreadsheetbench-v2-full-321",
      "version": "0.2.0",
      "status": "browser_scenario_ready",
      "browserScenarioStatus": "ready",
      "taskCount": 321,
      "attemptCountAtCurrentMatrix": 1284,
      "sourceBenchmark": "SpreadsheetBench V2 full 321",
      "taskLoader": ".tmp/official-benchmarks/staged-v2-full/tasks/<taskId>/agent/task.json",
      "browserScenario": "e2e/benchmark-ui-spreadsheetbench-generic.spec.ts",
      "commandShape": "SPREADSHEETBENCH_TASK_ID=<taskId> BENCH_AGENT_MODEL_POLICY=<modelId> npm run proofloop:live:spreadsheetbench-v2",
      "existingSmoke": "e2e/benchmark-ui-spreadsheetbench-v2.spec.ts proves one synthetic debugging path; e2e/benchmark-ui-spreadsheetbench-generic.spec.ts is the generic staged-task adapter.",
      "requiredArtifacts": [
        "live-user-contract.json",
        "node-trace-v2.json",
        "node-eval.json",
        "scorecard.md",
        "cost-ledger.json",
        "verifier-receipt.json",
        "visual-proof.png",
        "spreadsheetbench-v2-official-score.json",
        "exported-workbook.xlsx",
        "chart-visual-grade.json"
      ],
      "scorerBoundary": "Workbook/static scorer and rendered/VLM chart grading must run after candidate export; chart rubrics stay evaluator-only.",
      "blockers": [
        "No passing prod receipts are recorded for every staged V2 task/model yet; run the long-run queue to replace this with score evidence."
      ],
      "changelog": [
        "2026-07-05.1: contract created; existing synthetic V2 smoke remains only an adapter seed.",
        "2026-07-05.2: generic staged-task browser scenario added; tasks are runnable but not scored until receipts exist."
      ]
    },
    {
      "id": "accounting-live-config-to-prod-browser-room",
      "familyId": "accounting-live-proofloop",
      "version": "0.2.0",
      "status": "browser_scenario_ready",
      "browserScenarioStatus": "ready",
      "taskCount": 4,
      "attemptCountAtCurrentMatrix": 16,
      "sourceBenchmark": "Accounting ProofLoop product suite",
      "taskLoader": "proofloop/accounting/live.accounting.config.json",
      "browserScenario": "proofloop/live-browser-proof.spec.ts",
      "commandShape": "npm run proofloop:live:accounting:browser -- --prod --task-id <taskId> --model <modelId> --real-user",
      "requiredArtifacts": [
        "live-user-contract.json",
        "node-trace-v2.json",
        "node-eval.json",
        "scorecard.md",
        "cost-ledger.json",
        "verifier-receipt.json",
        "visual-proof.png",
        "accounting-live-config-to-prod-browser-room-scorecard.json"
      ],
      "scorerBoundary": "Product proof-loop suite; no official accounting benchmark score is claimed unless an upstream scorer is imported.",
      "blockers": [
        "No passing prod receipts are recorded for every accounting task/model yet; run the long-run queue to replace this with score evidence."
      ],
      "changelog": [
        "2026-07-05.1: contract created; current non-browser runner remains evidence but not prod real-user proof.",
        "2026-07-05.4: proofloop/live-browser-proof.spec.ts promoted as the real-user browser scenario with benchmark runtime profile disabled."
      ]
    },
    {
      "id": "notion-live-config-to-prod-browser-room",
      "familyId": "notion-live-proofloop",
      "version": "0.2.0",
      "status": "browser_scenario_ready",
      "browserScenarioStatus": "ready",
      "taskCount": 4,
      "attemptCountAtCurrentMatrix": 16,
      "sourceBenchmark": "Notion SDR/BDR ProofLoop product suite",
      "taskLoader": "proofloop/notion/live.notion.config.json",
      "browserScenario": "proofloop/live-browser-proof.spec.ts",
      "commandShape": "npm run proofloop:live:notion:browser -- --prod --task-id <taskId> --model <modelId> --real-user",
      "requiredArtifacts": [
        "live-user-contract.json",
        "node-trace-v2.json",
        "node-eval.json",
        "scorecard.md",
        "cost-ledger.json",
        "verifier-receipt.json",
        "visual-proof.png",
        "notion-live-config-to-prod-browser-room-scorecard.json"
      ],
      "scorerBoundary": "Product workflow suite; no official public benchmark score is claimed.",
      "blockers": [
        "No passing prod receipts are recorded for every Notion task/model yet; run the long-run queue to replace this with score evidence."
      ],
      "changelog": [
        "2026-07-05.1: contract created; current non-browser runner remains evidence but not prod real-user proof.",
        "2026-07-05.4: proofloop/live-browser-proof.spec.ts promoted as the real-user browser scenario with benchmark runtime profile disabled."
      ]
    },
    {
      "id": "proximitty-underwriting-prod-browser-room",
      "familyId": "proximitty-underwriting-pr0",
      "version": "0.2.0",
      "status": "browser_scenario_ready",
      "browserScenarioStatus": "ready",
      "taskCount": 4,
      "attemptCountAtCurrentMatrix": 16,
      "sourceBenchmark": "Synthetic Proximitty underwriting ProofLoop suite",
      "taskLoader": "proofloop/scenarios/proximitty-*.spec.ts",
      "browserScenario": "proofloop/benchmarks/proximitty/live-room-scenario.spec.ts",
      "commandShape": "npm run proofloop:proximitty:browser -- --prod --scenario <taskId> --model <modelId> --real-user",
      "requiredArtifacts": [
        "live-user-contract.json",
        "node-trace-v2.json",
        "node-eval.json",
        "scorecard.md",
        "cost-ledger.json",
        "verifier-receipt.json",
        "visual-proof.png",
        "proximitty-underwriting-prod-browser-room-scorecard.json"
      ],
      "scorerBoundary": "Synthetic underwriting evaluation only; not a real lending, insurance, legal, or financial decision score.",
      "blockers": [
        "No passing prod receipts are recorded for every Proximitty task/model yet; run the long-run queue to replace this with score evidence."
      ],
      "changelog": [
        "2026-07-05.1: contract created; current non-browser runner remains evidence but not prod real-user proof.",
        "2026-07-05.4: proofloop/benchmarks/proximitty/live-room-scenario.spec.ts promoted as the real-user browser scenario with benchmark runtime profile disabled."
      ]
    },
    {
      "id": "noderoom-multi-user-conflict-prod-browser-room",
      "familyId": "noderoom-multi-user-conflict",
      "version": "0.2.0",
      "status": "browser_scenario_ready",
      "browserScenarioStatus": "ready",
      "taskCount": 6,
      "attemptCountAtCurrentMatrix": 24,
      "sourceBenchmark": "NodeRoom internal multi-user coordination suite",
      "taskLoader": "tests/multiUserCoordinationProof.test.ts",
      "browserScenario": "proofloop/benchmarks/noderoom-multi-user/live-room-scenario.spec.ts",
      "commandShape": "npm run proofloop:live:multi-user-conflict -- --prod --task-id <taskId> --model <modelId> --real-user",
      "requiredArtifacts": [
        "live-user-contract.json",
        "node-trace-v2.json",
        "node-eval.json",
        "scorecard.md",
        "cost-ledger.json",
        "verifier-receipt.json",
        "visual-proof.png",
        "noderoom-multi-user-conflict-prod-browser-room-scorecard.json"
      ],
      "scorerBoundary": "Internal product coordination benchmark; no external official score is claimed.",
      "blockers": [
        "No passing prod receipts are recorded for every multi-user conflict task/model yet; run the long-run queue to replace this with score evidence."
      ],
      "changelog": [
        "2026-07-05.1: contract created; current non-browser runner remains evidence but not prod real-user proof.",
        "2026-07-05.4: proofloop/benchmarks/noderoom-multi-user/live-room-scenario.spec.ts promoted as the real-user browser scenario with benchmark runtime profile disabled."
      ]
    }
  ]
}