@tangle-network/agent-runtime 0.101.0 → 0.102.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (53) hide show
  1. package/README.md +1 -1
  2. package/dist/agent.js +4 -3
  3. package/dist/agent.js.map +1 -1
  4. package/dist/candidate-execution/index.d.ts +13 -2
  5. package/dist/candidate-execution/index.js +7 -3
  6. package/dist/{chunk-LV6WCT43.js → chunk-23Z7N5IG.js} +5 -5
  7. package/dist/chunk-2VZYRB55.js +834 -0
  8. package/dist/chunk-2VZYRB55.js.map +1 -0
  9. package/dist/{chunk-VOPHC4LB.js → chunk-4AMAFNUD.js} +32 -192
  10. package/dist/chunk-4AMAFNUD.js.map +1 -0
  11. package/dist/{chunk-BXZ7GPL4.js → chunk-AIBIRAJ6.js} +2 -2
  12. package/dist/{chunk-Y3SRWZMP.js → chunk-AYU35OTU.js} +2 -2
  13. package/dist/{chunk-U4PRAATQ.js → chunk-DA7YL6I4.js} +3 -3
  14. package/dist/{chunk-HZDEXTSL.js → chunk-IKGCKPDX.js} +2 -833
  15. package/dist/chunk-IKGCKPDX.js.map +1 -0
  16. package/dist/{chunk-WYC2XJF2.js → chunk-KSBUSEUA.js} +2 -2
  17. package/dist/{chunk-UQ6PNNXM.js → chunk-KVJQV342.js} +5 -3
  18. package/dist/{chunk-UQ6PNNXM.js.map → chunk-KVJQV342.js.map} +1 -1
  19. package/dist/chunk-LKKIM5QD.js +200 -0
  20. package/dist/chunk-LKKIM5QD.js.map +1 -0
  21. package/dist/{chunk-FXF2OL34.js → chunk-MAONWP4R.js} +2 -2
  22. package/dist/{chunk-SGQ4YIQW.js → chunk-N2CEE5Q3.js} +3 -3
  23. package/dist/chunk-SPCD4IQQ.js +578 -0
  24. package/dist/chunk-SPCD4IQQ.js.map +1 -0
  25. package/dist/{chunk-HGSHPVJ6.js → chunk-TGDHHHH4.js} +189 -189
  26. package/dist/{chunk-HGSHPVJ6.js.map → chunk-TGDHHHH4.js.map} +1 -1
  27. package/dist/{improve-g75IE2Cx.d.ts → improve-CIml6hT3.d.ts} +7 -0
  28. package/dist/index.d.ts +4 -4
  29. package/dist/index.js +16 -10
  30. package/dist/index.js.map +1 -1
  31. package/dist/intelligence.d.ts +60 -59
  32. package/dist/intelligence.js +65 -473
  33. package/dist/intelligence.js.map +1 -1
  34. package/dist/knowledge.js +8 -6
  35. package/dist/loop-runner-bin.js +5 -4
  36. package/dist/loops.js +4 -3
  37. package/dist/mcp/bin.js +4 -3
  38. package/dist/mcp/bin.js.map +1 -1
  39. package/dist/mcp/index.js +6 -5
  40. package/dist/mcp/index.js.map +1 -1
  41. package/dist/testing.d.ts +6 -0
  42. package/dist/testing.js +2638 -0
  43. package/dist/testing.js.map +1 -0
  44. package/package.json +11 -4
  45. package/dist/chunk-HZDEXTSL.js.map +0 -1
  46. package/dist/chunk-VOPHC4LB.js.map +0 -1
  47. /package/dist/{chunk-LV6WCT43.js.map → chunk-23Z7N5IG.js.map} +0 -0
  48. /package/dist/{chunk-BXZ7GPL4.js.map → chunk-AIBIRAJ6.js.map} +0 -0
  49. /package/dist/{chunk-Y3SRWZMP.js.map → chunk-AYU35OTU.js.map} +0 -0
  50. /package/dist/{chunk-U4PRAATQ.js.map → chunk-DA7YL6I4.js.map} +0 -0
  51. /package/dist/{chunk-WYC2XJF2.js.map → chunk-KSBUSEUA.js.map} +0 -0
  52. /package/dist/{chunk-FXF2OL34.js.map → chunk-MAONWP4R.js.map} +0 -0
  53. /package/dist/{chunk-SGQ4YIQW.js.map → chunk-N2CEE5Q3.js.map} +0 -0
@@ -0,0 +1,2638 @@
1
+ import {
2
+ verifyAgentImprovementProposal
3
+ } from "./chunk-4AMAFNUD.js";
4
+ import "./chunk-TGDHHHH4.js";
5
+ import "./chunk-IKGCKPDX.js";
6
+ import "./chunk-IMSNJSXH.js";
7
+ import "./chunk-3MDZX7YU.js";
8
+ import "./chunk-FD2MBMOH.js";
9
+ import "./chunk-M22PL4GH.js";
10
+ import "./chunk-YEJR7IXO.js";
11
+
12
+ // src/testing/fixtures/agent-improvement-proposal.json
13
+ var agent_improvement_proposal_default = {
14
+ changedSurfaces: ["prompt"],
15
+ digest: "sha256:b1a7049775c37741f37c3b5279a879748b7d33d114fc4b8f498fa2daa1fce1e3",
16
+ evaluation: {
17
+ decision: {
18
+ contributingChecks: [
19
+ {
20
+ name: "paired-significance",
21
+ passed: true
22
+ },
23
+ {
24
+ name: "statistical-power",
25
+ passed: true
26
+ },
27
+ {
28
+ name: "all-runs-completed",
29
+ passed: true
30
+ },
31
+ {
32
+ name: "candidate-task-pass",
33
+ passed: true
34
+ },
35
+ {
36
+ name: "critical-dimensions",
37
+ passed: true
38
+ },
39
+ {
40
+ name: "budget",
41
+ passed: true
42
+ }
43
+ ],
44
+ outcome: "ship",
45
+ reasons: ["all measured checks passed"]
46
+ },
47
+ diff: '--- baseline/profile (sha256:e53e1c8f9632bcb5db9033e32f33a90b9bd99abecef0bb52ff9069f07509e2f9)\n+++ candidate/profile (sha256:8b1e2f0f7b832958381c29baebe4a961b151af97a0ce9d611a15ad0741aeaa75)\n{\n "baseline": {\n "name": "candidate",\n "prompt": {\n "instructions": [\n "Inspect the repository, implement the fix, and run tests."\n ]\n },\n "model": {\n "default": "provider/model",\n "reasoningEffort": "high"\n },\n "harness": "codex",\n "resources": {\n "failOnError": true\n }\n },\n "candidate": {\n "name": "candidate",\n "prompt": {\n "systemPrompt": "Return the exact measured answer.",\n "instructions": [\n "Inspect the repository, implement the fix, and run tests."\n ]\n },\n "model": {\n "default": "provider/model",\n "reasoningEffort": "high"\n },\n "harness": "codex",\n "resources": {\n "failOnError": true\n }\n }\n}',
48
+ evaluation: {
49
+ durationMs: 300,
50
+ executionCostUsd: 0,
51
+ executionDurationMs: 300,
52
+ generationsExplored: 0,
53
+ searchCostUsd: 0,
54
+ searchDurationMs: 0,
55
+ totalCostUsd: 0
56
+ },
57
+ experiment: {
58
+ baseline: {
59
+ code: {
60
+ kind: "disabled"
61
+ },
62
+ digest: "sha256:5c21ee53e513fc604cb09754e21c392b24a424da0ef37dbf8f1ee4a8a0b08f09",
63
+ digestAlgorithm: "rfc8785-sha256",
64
+ execution: {
65
+ cwd: {
66
+ path: ".",
67
+ workspace: "task"
68
+ },
69
+ env: {
70
+ PATH: {
71
+ kind: "public",
72
+ value: "/usr/local/bin:/usr/bin:/bin"
73
+ }
74
+ },
75
+ environment: {
76
+ kind: "evaluator-task-container"
77
+ },
78
+ harness: "codex",
79
+ harnessVersion: "1.2.3",
80
+ instructionDelivery: {
81
+ kind: "stdin-utf8"
82
+ },
83
+ isolation: {
84
+ candidateSecrets: "disabled",
85
+ network: "disabled",
86
+ remoteIntegrations: "disabled"
87
+ },
88
+ launch: {
89
+ executable: "codex",
90
+ kind: "container-command"
91
+ }
92
+ },
93
+ kind: "agent-candidate-bundle",
94
+ memory: {
95
+ mode: "disabled"
96
+ },
97
+ profile: {
98
+ harness: "codex",
99
+ model: {
100
+ default: "provider/model",
101
+ reasoningEffort: "high"
102
+ },
103
+ name: "candidate",
104
+ prompt: {
105
+ instructions: ["Inspect the repository, implement the fix, and run tests."]
106
+ },
107
+ resources: {
108
+ failOnError: true
109
+ }
110
+ }
111
+ },
112
+ benchmark: {
113
+ suite: {
114
+ digest: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5",
115
+ digestAlgorithm: "rfc8785-sha256",
116
+ kind: "agent-candidate-benchmark-suite",
117
+ reps: 3,
118
+ seeds: [101, 102, 103],
119
+ taskDigests: ["sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277"]
120
+ },
121
+ tasks: [
122
+ {
123
+ attempt: {
124
+ maxAttempts: 1,
125
+ retryPolicy: "none"
126
+ },
127
+ benchmark: {
128
+ name: "repository-disjoint-smoke",
129
+ splitDigest: "sha256:dddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddd",
130
+ version: "1"
131
+ },
132
+ digest: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277",
133
+ digestAlgorithm: "rfc8785-sha256",
134
+ evaluatorTaskContainer: {
135
+ image: "ghcr.io/example/task",
136
+ indexDigest: "sha256:aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
137
+ manifestDigest: "sha256:bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
138
+ platform: {
139
+ architecture: "amd64",
140
+ os: "linux"
141
+ },
142
+ source: "evaluator-task-container"
143
+ },
144
+ grader: {
145
+ artifact: {
146
+ byteLength: 25,
147
+ locator: {
148
+ bucket: "candidate-test-artifacts",
149
+ key: "grader/9a354495185587b25e8c98104060068871a44c719e4f57d345a905fc2d72a6ad",
150
+ kind: "s3"
151
+ },
152
+ sha256: "sha256:9a354495185587b25e8c98104060068871a44c719e4f57d345a905fc2d72a6ad"
153
+ },
154
+ format: "tangle-grader",
155
+ name: "fixture-executable-grader",
156
+ version: "1.0.0"
157
+ },
158
+ instruction: "Fix the failing behavior without changing the public API.",
159
+ kind: "agent-candidate-benchmark-task",
160
+ limits: {
161
+ maxCostUsd: 5,
162
+ maxInputTokens: 1e5,
163
+ maxModelCalls: 50,
164
+ maxOutputTokens: 5e4,
165
+ maxSteps: 100,
166
+ timeoutMs: 6e4
167
+ },
168
+ model: {
169
+ model: "model-snapshot",
170
+ provider: "provider",
171
+ reasoningEffort: "high",
172
+ requested: "provider/model",
173
+ snapshot: "model-snapshot-2026-07-01"
174
+ },
175
+ outcome: {
176
+ kind: "output",
177
+ maxBytes: 1024,
178
+ mediaType: "text/plain"
179
+ },
180
+ scenario: {
181
+ id: "owner-repo-1",
182
+ kind: "coding",
183
+ scenarioDigest: "sha256:8888888888888888888888888888888888888888888888888888888888888888"
184
+ },
185
+ workspace: {
186
+ archive: {
187
+ byteLength: 17,
188
+ content: "ZW1wdHk6b3V0cHV0LXRhc2s=",
189
+ encoding: "base64",
190
+ sha256: "sha256:332dd9c74beab8a177396a4510dab46a50ca2074b203b10694e4eaa3071e2ef0"
191
+ },
192
+ digest: "sha256:3e9fdad46f14ce5b7c78a93e4fee0960483ae9d1cda21f0c390cdbad40bdac02",
193
+ kind: "agent-candidate-workspace-snapshot",
194
+ manifest: {
195
+ byteLength: 56,
196
+ content: "eyJmaWxlcyI6W10sImtpbmQiOiJhZ2VudC1jYW5kaWRhdGUtd29ya3NwYWNlLW1hbmlmZXN0In0=",
197
+ encoding: "base64",
198
+ sha256: "sha256:3e9fdad46f14ce5b7c78a93e4fee0960483ae9d1cda21f0c390cdbad40bdac02"
199
+ },
200
+ material: {
201
+ files: [],
202
+ kind: "agent-candidate-workspace-manifest"
203
+ }
204
+ }
205
+ }
206
+ ]
207
+ },
208
+ candidate: {
209
+ code: {
210
+ kind: "disabled"
211
+ },
212
+ digest: "sha256:60fcbb1c728194bd51d7d19cb732d1c3f1881dce7e0a6266b41c8b98cfd65693",
213
+ digestAlgorithm: "rfc8785-sha256",
214
+ execution: {
215
+ cwd: {
216
+ path: ".",
217
+ workspace: "task"
218
+ },
219
+ env: {
220
+ PATH: {
221
+ kind: "public",
222
+ value: "/usr/local/bin:/usr/bin:/bin"
223
+ }
224
+ },
225
+ environment: {
226
+ kind: "evaluator-task-container"
227
+ },
228
+ harness: "codex",
229
+ harnessVersion: "1.2.3",
230
+ instructionDelivery: {
231
+ kind: "stdin-utf8"
232
+ },
233
+ isolation: {
234
+ candidateSecrets: "disabled",
235
+ network: "disabled",
236
+ remoteIntegrations: "disabled"
237
+ },
238
+ launch: {
239
+ executable: "codex",
240
+ kind: "container-command"
241
+ }
242
+ },
243
+ kind: "agent-candidate-bundle",
244
+ memory: {
245
+ mode: "disabled"
246
+ },
247
+ profile: {
248
+ harness: "codex",
249
+ model: {
250
+ default: "provider/model",
251
+ reasoningEffort: "high"
252
+ },
253
+ name: "candidate",
254
+ prompt: {
255
+ instructions: ["Inspect the repository, implement the fix, and run tests."],
256
+ systemPrompt: "Return the exact measured answer."
257
+ },
258
+ resources: {
259
+ failOnError: true
260
+ }
261
+ }
262
+ },
263
+ candidateLineage: {
264
+ source: "human"
265
+ },
266
+ digest: "sha256:c47a0dad80e6cc4856cdb8a02a279da621ab7d9409e3127650b67f1fc1c14b2d",
267
+ digestAlgorithm: "rfc8785-sha256",
268
+ kind: "agent-candidate-experiment",
269
+ policy: {
270
+ bootstrapSeed: 1337,
271
+ budgetUsd: 1,
272
+ confidenceLevel: 0.95,
273
+ criticalDimensions: ["quality"],
274
+ deltaThreshold: 0,
275
+ minProductiveRuns: 3,
276
+ regressionTolerance: 0.05,
277
+ resamples: 500
278
+ }
279
+ },
280
+ kind: "agent-improvement-measured-comparison",
281
+ measurements: [
282
+ {
283
+ baseline: {
284
+ digest: "sha256:6bf8492d4fb6cb795715f861ad1150ce8194aaf6075ce898e411fdbaf4e6f534",
285
+ kind: "agent-candidate-execution-evidence",
286
+ materializationReceipt: {
287
+ benchmark: {
288
+ suite: {
289
+ digest: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5",
290
+ material: {
291
+ byteLength: 198,
292
+ content: "eyJkaWdlc3RBbGdvcml0aG0iOiJyZmM4Nzg1LXNoYTI1NiIsImtpbmQiOiJhZ2VudC1jYW5kaWRhdGUtYmVuY2htYXJrLXN1aXRlIiwicmVwcyI6Mywic2VlZHMiOlsxMDEsMTAyLDEwM10sInRhc2tEaWdlc3RzIjpbInNoYTI1Njo2YWE3MTgzMzFiNmRhMzEwZGNhYjZkZjJlMzM5N2NiNzcwODQwZTU1NWE4MmUwNjg5NTc3MTU3Yjg5YjRlMjc3Il19",
293
+ encoding: "base64",
294
+ sha256: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5"
295
+ }
296
+ },
297
+ task: {
298
+ digest: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277",
299
+ material: {
300
+ byteLength: 2096,
301
+ content: "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",
302
+ encoding: "base64",
303
+ sha256: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277"
304
+ }
305
+ }
306
+ },
307
+ bundleDigest: "sha256:5c21ee53e513fc604cb09754e21c392b24a424da0ef37dbf8f1ee4a8a0b08f09",
308
+ codeKind: "disabled",
309
+ container: {
310
+ image: "ghcr.io/example/task",
311
+ indexDigest: "sha256:aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
312
+ manifestDigest: "sha256:bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
313
+ platform: {
314
+ architecture: "amd64",
315
+ os: "linux"
316
+ },
317
+ source: "evaluator-task-container"
318
+ },
319
+ digest: "sha256:50573a8655c2f83044a105990573b1a5ed84d3a8d9241e01cda565eead22b7cc",
320
+ digestAlgorithm: "rfc8785-sha256",
321
+ executionPlan: {
322
+ artifact: {
323
+ byteLength: 2030,
324
+ content: "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",
325
+ encoding: "base64",
326
+ sha256: "sha256:205204302122b2c90373cf000406d8841c6bf8cd99f621637fe7c6824e243acf"
327
+ },
328
+ digest: "sha256:205204302122b2c90373cf000406d8841c6bf8cd99f621637fe7c6824e243acf",
329
+ kind: "agent-candidate-execution-plan",
330
+ material: {
331
+ codeKind: "disabled",
332
+ container: {
333
+ image: "ghcr.io/example/task",
334
+ indexDigest: "sha256:aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
335
+ manifestDigest: "sha256:bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
336
+ platform: {
337
+ architecture: "amd64",
338
+ os: "linux"
339
+ },
340
+ source: "evaluator-task-container"
341
+ },
342
+ executionId: "experiment-baseline-0-0",
343
+ harness: "codex",
344
+ harnessVersion: "1.2.3",
345
+ instructionDelivery: {
346
+ kind: "stdin-utf8"
347
+ },
348
+ kind: "agent-candidate-execution-plan-material",
349
+ launch: {
350
+ args: [],
351
+ cwd: {
352
+ path: ".",
353
+ workspace: "task"
354
+ },
355
+ env: {
356
+ PATH: {
357
+ kind: "public",
358
+ value: "/usr/local/bin:/usr/bin:/bin"
359
+ }
360
+ },
361
+ executable: "codex"
362
+ },
363
+ limits: {
364
+ maxCostUsd: 5,
365
+ maxInputTokens: 1e5,
366
+ maxModelCalls: 50,
367
+ maxOutputTokens: 5e4,
368
+ maxSteps: 100,
369
+ timeoutMs: 6e4
370
+ },
371
+ memory: {
372
+ mode: "disabled"
373
+ },
374
+ model: {
375
+ access: {
376
+ grantDigest: "sha256:cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc",
377
+ kind: "evaluator-mediated",
378
+ network: {
379
+ domains: ["router.tangle.tools"],
380
+ mode: "gateway-only"
381
+ }
382
+ },
383
+ policy: "single",
384
+ resolved: {
385
+ model: "model-snapshot",
386
+ provider: "provider",
387
+ reasoningEffort: "high",
388
+ requested: "provider/model",
389
+ snapshot: "model-snapshot-2026-07-01"
390
+ },
391
+ routes: [
392
+ {
393
+ kind: "primary",
394
+ requested: "provider/model"
395
+ }
396
+ ]
397
+ },
398
+ network: {
399
+ mode: "disabled"
400
+ },
401
+ profile: {
402
+ mountPaths: ["AGENTS.md"],
403
+ planDigest: "sha256:af65bd67cf53561cd7bfac4e8c913ef18e36c0c2fe20a0bca1df55e59b21fab5",
404
+ targetWorkspace: "task"
405
+ },
406
+ runCell: {
407
+ arm: "baseline",
408
+ attempt: 1,
409
+ bundleDigest: "sha256:5c21ee53e513fc604cb09754e21c392b24a424da0ef37dbf8f1ee4a8a0b08f09",
410
+ digest: "sha256:5246bcb23929d3ef84c78c5185390aab5e2309f1267d38e0d1c0ddc18232b6d4",
411
+ experimentDigest: "sha256:c47a0dad80e6cc4856cdb8a02a279da621ab7d9409e3127650b67f1fc1c14b2d",
412
+ kind: "agent-candidate-run-cell",
413
+ repetition: 0,
414
+ seed: 101,
415
+ suiteDigest: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5",
416
+ taskDigest: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277",
417
+ taskIndex: 0
418
+ },
419
+ workspaces: {
420
+ taskRoot: "/workspace/task"
421
+ }
422
+ }
423
+ },
424
+ harness: "codex",
425
+ harnessVersion: "1.2.3",
426
+ kind: "agent-candidate-materialization",
427
+ profileActivation: {
428
+ digest: "sha256:273b700e4080e7e6cc4bd58a755466d18ce323eaaa195bb4202fea456422bb66",
429
+ files: [
430
+ {
431
+ content: "Inspect the repository, implement the fix, and run tests.\n",
432
+ mode: 420,
433
+ path: "AGENTS.md"
434
+ }
435
+ ],
436
+ kind: "agent-candidate-profile-activation",
437
+ profilePlan: {
438
+ artifact: {
439
+ byteLength: 287,
440
+ content: "eyJlbnYiOnt9LCJmaWxlcyI6W3siY29udGVudFNoYTI1NiI6InNoYTI1NjozOWFkYmI5YjdmNTcxZjBmZDU4NmU3Nzk1MDIzZDI1OTA5YWRiZTk1ODZiYjIxZWUxMzRkM2Q3MjQ4YmYxNTk5IiwibW9kZSI6NDIwLCJyZWxQYXRoIjoiQUdFTlRTLm1kIn1dLCJmbGFncyI6W10sImhhcm5lc3MiOiJjb2RleCIsInNvdXJjZVByb2ZpbGVEaWdlc3QiOiJzaGEyNTY6ZTUzZTFjOGY5NjMyYmNiNWRiOTAzM2UzMmYzM2E5MGI5YmQ5OWFiZWNlZjBiYjUyZmY5MDY5ZjA3NTA5ZTJmOSIsInVuc3VwcG9ydGVkIjpbXX0=",
441
+ encoding: "base64",
442
+ sha256: "sha256:af65bd67cf53561cd7bfac4e8c913ef18e36c0c2fe20a0bca1df55e59b21fab5"
443
+ },
444
+ digest: "sha256:af65bd67cf53561cd7bfac4e8c913ef18e36c0c2fe20a0bca1df55e59b21fab5",
445
+ kind: "agent-profile-workspace-plan",
446
+ material: {
447
+ env: {},
448
+ files: [
449
+ {
450
+ contentSha256: "sha256:39adbb9b7f571f0fd586e7795023d25909adbe9586bb21ee134d3d7248bf1599",
451
+ mode: 420,
452
+ relPath: "AGENTS.md"
453
+ }
454
+ ],
455
+ flags: [],
456
+ harness: "codex",
457
+ sourceProfileDigest: "sha256:e53e1c8f9632bcb5db9033e32f33a90b9bd99abecef0bb52ff9069f07509e2f9",
458
+ unsupported: []
459
+ }
460
+ }
461
+ },
462
+ resolvedModel: {
463
+ model: "model-snapshot",
464
+ provider: "provider",
465
+ reasoningEffort: "high",
466
+ requested: "provider/model",
467
+ snapshot: "model-snapshot-2026-07-01"
468
+ }
469
+ },
470
+ receipt: {
471
+ benchmarkResult: {
472
+ artifact: {
473
+ byteLength: 1122,
474
+ locator: {
475
+ bucket: "candidate-test-artifacts",
476
+ key: "benchmark-result/e9a709a18c6a51d8f81080c79f856162eb740e7664eba2906bc5888752eb27a6",
477
+ kind: "s3"
478
+ },
479
+ sha256: "sha256:e9a709a18c6a51d8f81080c79f856162eb740e7664eba2906bc5888752eb27a6"
480
+ },
481
+ digest: "sha256:e9a709a18c6a51d8f81080c79f856162eb740e7664eba2906bc5888752eb27a6",
482
+ kind: "agent-candidate-benchmark-result",
483
+ material: {
484
+ dimensions: [
485
+ {
486
+ name: "quality",
487
+ score: 0
488
+ }
489
+ ],
490
+ evidence: {
491
+ byteLength: 11,
492
+ locator: {
493
+ bucket: "candidate-test-artifacts",
494
+ key: "grader-evidence/628cb4ead38e34e34a34f78fa4481f2a0ccf8e9139acfae69d79dc997e90e36b",
495
+ kind: "s3"
496
+ },
497
+ sha256: "sha256:628cb4ead38e34e34a34f78fa4481f2a0ccf8e9139acfae69d79dc997e90e36b"
498
+ },
499
+ executionPlanDigest: "sha256:205204302122b2c90373cf000406d8841c6bf8cd99f621637fe7c6824e243acf",
500
+ grader: {
501
+ artifact: {
502
+ byteLength: 25,
503
+ locator: {
504
+ bucket: "candidate-test-artifacts",
505
+ key: "grader/9a354495185587b25e8c98104060068871a44c719e4f57d345a905fc2d72a6ad",
506
+ kind: "s3"
507
+ },
508
+ sha256: "sha256:9a354495185587b25e8c98104060068871a44c719e4f57d345a905fc2d72a6ad"
509
+ },
510
+ format: "tangle-grader",
511
+ name: "fixture-executable-grader",
512
+ version: "1.0.0"
513
+ },
514
+ grading: {
515
+ timing: {
516
+ durationMs: 0,
517
+ endedAtMs: 17836416e5,
518
+ startedAtMs: 17836416e5
519
+ },
520
+ usage: {
521
+ cachedInputTokens: 0,
522
+ costUsdNanos: 0,
523
+ inputTokens: 0,
524
+ modelCalls: 0,
525
+ outputTokens: 0,
526
+ reasoningTokens: 0
527
+ }
528
+ },
529
+ kind: "agent-candidate-benchmark-result-material",
530
+ passed: false,
531
+ score: 0,
532
+ taskOutcomeDigest: "sha256:96d90a54ff0cdc7ef2dc49e80fad20175c10a761eb6abd3bf127da35340994ed"
533
+ }
534
+ },
535
+ bundleDigest: "sha256:5c21ee53e513fc604cb09754e21c392b24a424da0ef37dbf8f1ee4a8a0b08f09",
536
+ digest: "sha256:792a1fc3d525e029e74f9fe88940277b6d7c162959b4ecc943b25cbaffe3d86f",
537
+ digestAlgorithm: "rfc8785-sha256",
538
+ executionPlanDigest: "sha256:205204302122b2c90373cf000406d8841c6bf8cd99f621637fe7c6824e243acf",
539
+ executorCapture: {
540
+ byteLength: 477,
541
+ locator: {
542
+ bucket: "candidate-test-artifacts",
543
+ key: "executor-capture/64bb8b780db79c2a5bf89c46cefb755060f18271084748143c194d5716c5bedd",
544
+ kind: "s3"
545
+ },
546
+ sha256: "sha256:64bb8b780db79c2a5bf89c46cefb755060f18271084748143c194d5716c5bedd"
547
+ },
548
+ kind: "agent-candidate-run",
549
+ materializationReceiptDigest: "sha256:50573a8655c2f83044a105990573b1a5ed84d3a8d9241e01cda565eead22b7cc",
550
+ memory: {
551
+ mode: "disabled"
552
+ },
553
+ modelSettlement: {
554
+ artifact: {
555
+ byteLength: 615,
556
+ locator: {
557
+ bucket: "candidate-test-artifacts",
558
+ key: "model-settlement/625a1690d64a8ccee6b2d616f7ee7a9648edb1ddf9ea58711bef5a4a306cdba0",
559
+ kind: "s3"
560
+ },
561
+ sha256: "sha256:625a1690d64a8ccee6b2d616f7ee7a9648edb1ddf9ea58711bef5a4a306cdba0"
562
+ },
563
+ digest: "sha256:625a1690d64a8ccee6b2d616f7ee7a9648edb1ddf9ea58711bef5a4a306cdba0",
564
+ kind: "agent-candidate-model-settlement",
565
+ material: {
566
+ calls: [],
567
+ closed: true,
568
+ executionPlanDigest: "sha256:205204302122b2c90373cf000406d8841c6bf8cd99f621637fe7c6824e243acf",
569
+ grantDigest: "sha256:cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc",
570
+ kind: "agent-candidate-model-settlement-material",
571
+ preparationId: "candidate-preparation.nBqsDbKYlsiY4Tto5XaxkR3RCTop9PA9_kX-kOSzR8U",
572
+ resolved: {
573
+ model: "model-snapshot",
574
+ provider: "provider",
575
+ reasoningEffort: "high",
576
+ requested: "provider/model",
577
+ snapshot: "model-snapshot-2026-07-01"
578
+ },
579
+ usage: {
580
+ cachedInputTokens: 0,
581
+ costUsdNanos: 0,
582
+ inputTokens: 0,
583
+ modelCalls: 0,
584
+ outputTokens: 0,
585
+ reasoningTokens: 0
586
+ }
587
+ }
588
+ },
589
+ runCellDigest: "sha256:5246bcb23929d3ef84c78c5185390aab5e2309f1267d38e0d1c0ddc18232b6d4",
590
+ taskOutcome: {
591
+ artifact: {
592
+ byteLength: 478,
593
+ locator: {
594
+ bucket: "candidate-test-artifacts",
595
+ key: "task-outcome/96d90a54ff0cdc7ef2dc49e80fad20175c10a761eb6abd3bf127da35340994ed",
596
+ kind: "s3"
597
+ },
598
+ sha256: "sha256:96d90a54ff0cdc7ef2dc49e80fad20175c10a761eb6abd3bf127da35340994ed"
599
+ },
600
+ digest: "sha256:96d90a54ff0cdc7ef2dc49e80fad20175c10a761eb6abd3bf127da35340994ed",
601
+ kind: "agent-candidate-task-outcome",
602
+ material: {
603
+ executionPlanDigest: "sha256:205204302122b2c90373cf000406d8841c6bf8cd99f621637fe7c6824e243acf",
604
+ kind: "agent-candidate-task-outcome-material",
605
+ outcome: {
606
+ artifact: {
607
+ byteLength: 4,
608
+ locator: {
609
+ bucket: "candidate-test-artifacts",
610
+ key: "task-output/481ba4019c9d2710c3386537382592c093ef02bf5b056c30237b3d92b0e19a1d",
611
+ kind: "s3"
612
+ },
613
+ sha256: "sha256:481ba4019c9d2710c3386537382592c093ef02bf5b056c30237b3d92b0e19a1d"
614
+ },
615
+ kind: "output",
616
+ spec: {
617
+ maxBytes: 1024,
618
+ mediaType: "text/plain"
619
+ }
620
+ }
621
+ }
622
+ },
623
+ termination: {
624
+ exitCode: 0,
625
+ kind: "exit"
626
+ },
627
+ timing: {
628
+ durationMs: 50,
629
+ endedAtMs: 1050,
630
+ startedAtMs: 1e3
631
+ },
632
+ trace: {
633
+ artifact: {
634
+ byteLength: 748,
635
+ locator: {
636
+ bucket: "candidate-test-artifacts",
637
+ key: "trace/2fce8b3ba11deb5251d3b7b9de66d1b4788c18a2b56aafe77ddd57423db41593",
638
+ kind: "s3"
639
+ },
640
+ sha256: "sha256:2fce8b3ba11deb5251d3b7b9de66d1b4788c18a2b56aafe77ddd57423db41593"
641
+ },
642
+ eventCount: 1,
643
+ modelCallCount: 0
644
+ }
645
+ }
646
+ },
647
+ candidate: {
648
+ digest: "sha256:d99b9e877621de7c5fe20850fcdcf6b88119e46b0505dc4a7bf867201b0afafe",
649
+ kind: "agent-candidate-execution-evidence",
650
+ materializationReceipt: {
651
+ benchmark: {
652
+ suite: {
653
+ digest: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5",
654
+ material: {
655
+ byteLength: 198,
656
+ content: "eyJkaWdlc3RBbGdvcml0aG0iOiJyZmM4Nzg1LXNoYTI1NiIsImtpbmQiOiJhZ2VudC1jYW5kaWRhdGUtYmVuY2htYXJrLXN1aXRlIiwicmVwcyI6Mywic2VlZHMiOlsxMDEsMTAyLDEwM10sInRhc2tEaWdlc3RzIjpbInNoYTI1Njo2YWE3MTgzMzFiNmRhMzEwZGNhYjZkZjJlMzM5N2NiNzcwODQwZTU1NWE4MmUwNjg5NTc3MTU3Yjg5YjRlMjc3Il19",
657
+ encoding: "base64",
658
+ sha256: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5"
659
+ }
660
+ },
661
+ task: {
662
+ digest: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277",
663
+ material: {
664
+ byteLength: 2096,
665
+ content: "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",
666
+ encoding: "base64",
667
+ sha256: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277"
668
+ }
669
+ }
670
+ },
671
+ bundleDigest: "sha256:60fcbb1c728194bd51d7d19cb732d1c3f1881dce7e0a6266b41c8b98cfd65693",
672
+ codeKind: "disabled",
673
+ container: {
674
+ image: "ghcr.io/example/task",
675
+ indexDigest: "sha256:aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
676
+ manifestDigest: "sha256:bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
677
+ platform: {
678
+ architecture: "amd64",
679
+ os: "linux"
680
+ },
681
+ source: "evaluator-task-container"
682
+ },
683
+ digest: "sha256:dc45d1b3f6fb88d4b0b709f6a2ff63164df737dddd582a8749ff7d91368662d9",
684
+ digestAlgorithm: "rfc8785-sha256",
685
+ executionPlan: {
686
+ artifact: {
687
+ byteLength: 2032,
688
+ content: "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",
689
+ encoding: "base64",
690
+ sha256: "sha256:9ee298c2aa3a2b814f4ce41bf8698e20aab372d28922dca97a5c829516ea44f6"
691
+ },
692
+ digest: "sha256:9ee298c2aa3a2b814f4ce41bf8698e20aab372d28922dca97a5c829516ea44f6",
693
+ kind: "agent-candidate-execution-plan",
694
+ material: {
695
+ codeKind: "disabled",
696
+ container: {
697
+ image: "ghcr.io/example/task",
698
+ indexDigest: "sha256:aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
699
+ manifestDigest: "sha256:bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
700
+ platform: {
701
+ architecture: "amd64",
702
+ os: "linux"
703
+ },
704
+ source: "evaluator-task-container"
705
+ },
706
+ executionId: "experiment-candidate-0-0",
707
+ harness: "codex",
708
+ harnessVersion: "1.2.3",
709
+ instructionDelivery: {
710
+ kind: "stdin-utf8"
711
+ },
712
+ kind: "agent-candidate-execution-plan-material",
713
+ launch: {
714
+ args: [],
715
+ cwd: {
716
+ path: ".",
717
+ workspace: "task"
718
+ },
719
+ env: {
720
+ PATH: {
721
+ kind: "public",
722
+ value: "/usr/local/bin:/usr/bin:/bin"
723
+ }
724
+ },
725
+ executable: "codex"
726
+ },
727
+ limits: {
728
+ maxCostUsd: 5,
729
+ maxInputTokens: 1e5,
730
+ maxModelCalls: 50,
731
+ maxOutputTokens: 5e4,
732
+ maxSteps: 100,
733
+ timeoutMs: 6e4
734
+ },
735
+ memory: {
736
+ mode: "disabled"
737
+ },
738
+ model: {
739
+ access: {
740
+ grantDigest: "sha256:cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc",
741
+ kind: "evaluator-mediated",
742
+ network: {
743
+ domains: ["router.tangle.tools"],
744
+ mode: "gateway-only"
745
+ }
746
+ },
747
+ policy: "single",
748
+ resolved: {
749
+ model: "model-snapshot",
750
+ provider: "provider",
751
+ reasoningEffort: "high",
752
+ requested: "provider/model",
753
+ snapshot: "model-snapshot-2026-07-01"
754
+ },
755
+ routes: [
756
+ {
757
+ kind: "primary",
758
+ requested: "provider/model"
759
+ }
760
+ ]
761
+ },
762
+ network: {
763
+ mode: "disabled"
764
+ },
765
+ profile: {
766
+ mountPaths: ["AGENTS.md"],
767
+ planDigest: "sha256:4b4214d1bae911df6dda3a5d8585c65df2f0b2380126eb40c0cbd46a4a2fba5b",
768
+ targetWorkspace: "task"
769
+ },
770
+ runCell: {
771
+ arm: "candidate",
772
+ attempt: 1,
773
+ bundleDigest: "sha256:60fcbb1c728194bd51d7d19cb732d1c3f1881dce7e0a6266b41c8b98cfd65693",
774
+ digest: "sha256:e8faaad9b6ae2779d56d4f56d98f3197ed3d375298084ac08410c54632fb8b61",
775
+ experimentDigest: "sha256:c47a0dad80e6cc4856cdb8a02a279da621ab7d9409e3127650b67f1fc1c14b2d",
776
+ kind: "agent-candidate-run-cell",
777
+ repetition: 0,
778
+ seed: 101,
779
+ suiteDigest: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5",
780
+ taskDigest: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277",
781
+ taskIndex: 0
782
+ },
783
+ workspaces: {
784
+ taskRoot: "/workspace/task"
785
+ }
786
+ }
787
+ },
788
+ harness: "codex",
789
+ harnessVersion: "1.2.3",
790
+ kind: "agent-candidate-materialization",
791
+ profileActivation: {
792
+ digest: "sha256:3103f8282be70a843eada63410c41d9eb5ac7b824c758eed6c2f8610de169f45",
793
+ files: [
794
+ {
795
+ content: "Inspect the repository, implement the fix, and run tests.\n",
796
+ mode: 420,
797
+ path: "AGENTS.md"
798
+ }
799
+ ],
800
+ kind: "agent-candidate-profile-activation",
801
+ profilePlan: {
802
+ artifact: {
803
+ byteLength: 364,
804
+ content: "eyJlbnYiOnt9LCJmaWxlcyI6W3siY29udGVudFNoYTI1NiI6InNoYTI1NjozOWFkYmI5YjdmNTcxZjBmZDU4NmU3Nzk1MDIzZDI1OTA5YWRiZTk1ODZiYjIxZWUxMzRkM2Q3MjQ4YmYxNTk5IiwibW9kZSI6NDIwLCJyZWxQYXRoIjoiQUdFTlRTLm1kIn1dLCJmbGFncyI6W10sImhhcm5lc3MiOiJjb2RleCIsInNvdXJjZVByb2ZpbGVEaWdlc3QiOiJzaGEyNTY6OGIxZTJmMGY3YjgzMjk1ODM4MWMyOWJhZWJlNGE5NjFiMTUxYWY5N2EwY2U5ZDYxMWExNWFkMDc0MWFlYWE3NSIsInN5c3RlbVByb21wdCI6eyJraW5kIjoicHVibGljIiwidmFsdWUiOiJSZXR1cm4gdGhlIGV4YWN0IG1lYXN1cmVkIGFuc3dlci4ifSwidW5zdXBwb3J0ZWQiOltdfQ==",
805
+ encoding: "base64",
806
+ sha256: "sha256:4b4214d1bae911df6dda3a5d8585c65df2f0b2380126eb40c0cbd46a4a2fba5b"
807
+ },
808
+ digest: "sha256:4b4214d1bae911df6dda3a5d8585c65df2f0b2380126eb40c0cbd46a4a2fba5b",
809
+ kind: "agent-profile-workspace-plan",
810
+ material: {
811
+ env: {},
812
+ files: [
813
+ {
814
+ contentSha256: "sha256:39adbb9b7f571f0fd586e7795023d25909adbe9586bb21ee134d3d7248bf1599",
815
+ mode: 420,
816
+ relPath: "AGENTS.md"
817
+ }
818
+ ],
819
+ flags: [],
820
+ harness: "codex",
821
+ sourceProfileDigest: "sha256:8b1e2f0f7b832958381c29baebe4a961b151af97a0ce9d611a15ad0741aeaa75",
822
+ systemPrompt: {
823
+ kind: "public",
824
+ value: "Return the exact measured answer."
825
+ },
826
+ unsupported: []
827
+ }
828
+ }
829
+ },
830
+ resolvedModel: {
831
+ model: "model-snapshot",
832
+ provider: "provider",
833
+ reasoningEffort: "high",
834
+ requested: "provider/model",
835
+ snapshot: "model-snapshot-2026-07-01"
836
+ }
837
+ },
838
+ receipt: {
839
+ benchmarkResult: {
840
+ artifact: {
841
+ byteLength: 1121,
842
+ locator: {
843
+ bucket: "candidate-test-artifacts",
844
+ key: "benchmark-result/78dab25681b91e59579ac46cc8b3d2bfb58d9f57430eb2625e9f2cd534d2b15a",
845
+ kind: "s3"
846
+ },
847
+ sha256: "sha256:78dab25681b91e59579ac46cc8b3d2bfb58d9f57430eb2625e9f2cd534d2b15a"
848
+ },
849
+ digest: "sha256:78dab25681b91e59579ac46cc8b3d2bfb58d9f57430eb2625e9f2cd534d2b15a",
850
+ kind: "agent-candidate-benchmark-result",
851
+ material: {
852
+ dimensions: [
853
+ {
854
+ name: "quality",
855
+ score: 1
856
+ }
857
+ ],
858
+ evidence: {
859
+ byteLength: 11,
860
+ locator: {
861
+ bucket: "candidate-test-artifacts",
862
+ key: "grader-evidence/9b9b3a1471309177261cfe65ea9c298e0dd372e4b5d087f8d35d7b732485373d",
863
+ kind: "s3"
864
+ },
865
+ sha256: "sha256:9b9b3a1471309177261cfe65ea9c298e0dd372e4b5d087f8d35d7b732485373d"
866
+ },
867
+ executionPlanDigest: "sha256:9ee298c2aa3a2b814f4ce41bf8698e20aab372d28922dca97a5c829516ea44f6",
868
+ grader: {
869
+ artifact: {
870
+ byteLength: 25,
871
+ locator: {
872
+ bucket: "candidate-test-artifacts",
873
+ key: "grader/9a354495185587b25e8c98104060068871a44c719e4f57d345a905fc2d72a6ad",
874
+ kind: "s3"
875
+ },
876
+ sha256: "sha256:9a354495185587b25e8c98104060068871a44c719e4f57d345a905fc2d72a6ad"
877
+ },
878
+ format: "tangle-grader",
879
+ name: "fixture-executable-grader",
880
+ version: "1.0.0"
881
+ },
882
+ grading: {
883
+ timing: {
884
+ durationMs: 0,
885
+ endedAtMs: 17836416e5,
886
+ startedAtMs: 17836416e5
887
+ },
888
+ usage: {
889
+ cachedInputTokens: 0,
890
+ costUsdNanos: 0,
891
+ inputTokens: 0,
892
+ modelCalls: 0,
893
+ outputTokens: 0,
894
+ reasoningTokens: 0
895
+ }
896
+ },
897
+ kind: "agent-candidate-benchmark-result-material",
898
+ passed: true,
899
+ score: 1,
900
+ taskOutcomeDigest: "sha256:1674e3e6fc00ca4e74f44dc8247f3f6fd334c62f25b24eda25e37ca91d5fb016"
901
+ }
902
+ },
903
+ bundleDigest: "sha256:60fcbb1c728194bd51d7d19cb732d1c3f1881dce7e0a6266b41c8b98cfd65693",
904
+ digest: "sha256:3e003fb6cf3bde16a59f28901d29c752595794b000b79268b4b7be853adb4551",
905
+ digestAlgorithm: "rfc8785-sha256",
906
+ executionPlanDigest: "sha256:9ee298c2aa3a2b814f4ce41bf8698e20aab372d28922dca97a5c829516ea44f6",
907
+ executorCapture: {
908
+ byteLength: 477,
909
+ locator: {
910
+ bucket: "candidate-test-artifacts",
911
+ key: "executor-capture/7b8a015c52a055cc2befaf553c1fe14e2f4daeafd520f085be2f5e29682baff3",
912
+ kind: "s3"
913
+ },
914
+ sha256: "sha256:7b8a015c52a055cc2befaf553c1fe14e2f4daeafd520f085be2f5e29682baff3"
915
+ },
916
+ kind: "agent-candidate-run",
917
+ materializationReceiptDigest: "sha256:dc45d1b3f6fb88d4b0b709f6a2ff63164df737dddd582a8749ff7d91368662d9",
918
+ memory: {
919
+ mode: "disabled"
920
+ },
921
+ modelSettlement: {
922
+ artifact: {
923
+ byteLength: 615,
924
+ locator: {
925
+ bucket: "candidate-test-artifacts",
926
+ key: "model-settlement/6436cd10c36b64083ac00b0925efc03ef8d5e873f64cc2be5bdce4533366c8c4",
927
+ kind: "s3"
928
+ },
929
+ sha256: "sha256:6436cd10c36b64083ac00b0925efc03ef8d5e873f64cc2be5bdce4533366c8c4"
930
+ },
931
+ digest: "sha256:6436cd10c36b64083ac00b0925efc03ef8d5e873f64cc2be5bdce4533366c8c4",
932
+ kind: "agent-candidate-model-settlement",
933
+ material: {
934
+ calls: [],
935
+ closed: true,
936
+ executionPlanDigest: "sha256:9ee298c2aa3a2b814f4ce41bf8698e20aab372d28922dca97a5c829516ea44f6",
937
+ grantDigest: "sha256:cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc",
938
+ kind: "agent-candidate-model-settlement-material",
939
+ preparationId: "candidate-preparation.nBqsDbKYlsiY4Tto5XaxkR3RCTop9PA9_kX-kOSzR8U",
940
+ resolved: {
941
+ model: "model-snapshot",
942
+ provider: "provider",
943
+ reasoningEffort: "high",
944
+ requested: "provider/model",
945
+ snapshot: "model-snapshot-2026-07-01"
946
+ },
947
+ usage: {
948
+ cachedInputTokens: 0,
949
+ costUsdNanos: 0,
950
+ inputTokens: 0,
951
+ modelCalls: 0,
952
+ outputTokens: 0,
953
+ reasoningTokens: 0
954
+ }
955
+ }
956
+ },
957
+ runCellDigest: "sha256:e8faaad9b6ae2779d56d4f56d98f3197ed3d375298084ac08410c54632fb8b61",
958
+ taskOutcome: {
959
+ artifact: {
960
+ byteLength: 478,
961
+ locator: {
962
+ bucket: "candidate-test-artifacts",
963
+ key: "task-outcome/1674e3e6fc00ca4e74f44dc8247f3f6fd334c62f25b24eda25e37ca91d5fb016",
964
+ kind: "s3"
965
+ },
966
+ sha256: "sha256:1674e3e6fc00ca4e74f44dc8247f3f6fd334c62f25b24eda25e37ca91d5fb016"
967
+ },
968
+ digest: "sha256:1674e3e6fc00ca4e74f44dc8247f3f6fd334c62f25b24eda25e37ca91d5fb016",
969
+ kind: "agent-candidate-task-outcome",
970
+ material: {
971
+ executionPlanDigest: "sha256:9ee298c2aa3a2b814f4ce41bf8698e20aab372d28922dca97a5c829516ea44f6",
972
+ kind: "agent-candidate-task-outcome-material",
973
+ outcome: {
974
+ artifact: {
975
+ byteLength: 6,
976
+ locator: {
977
+ bucket: "candidate-test-artifacts",
978
+ key: "task-output/eda15ce4834ac303532bc6973df5ff8a5deb75d7dfceb9fc2cdf1bb450a01cf5",
979
+ kind: "s3"
980
+ },
981
+ sha256: "sha256:eda15ce4834ac303532bc6973df5ff8a5deb75d7dfceb9fc2cdf1bb450a01cf5"
982
+ },
983
+ kind: "output",
984
+ spec: {
985
+ maxBytes: 1024,
986
+ mediaType: "text/plain"
987
+ }
988
+ }
989
+ }
990
+ },
991
+ termination: {
992
+ exitCode: 0,
993
+ kind: "exit"
994
+ },
995
+ timing: {
996
+ durationMs: 50,
997
+ endedAtMs: 1050,
998
+ startedAtMs: 1e3
999
+ },
1000
+ trace: {
1001
+ artifact: {
1002
+ byteLength: 750,
1003
+ locator: {
1004
+ bucket: "candidate-test-artifacts",
1005
+ key: "trace/f40b48a52282eaf104755ceb8dd6d713f6cc93fda296767f2656ef66ac4a55ff",
1006
+ kind: "s3"
1007
+ },
1008
+ sha256: "sha256:f40b48a52282eaf104755ceb8dd6d713f6cc93fda296767f2656ef66ac4a55ff"
1009
+ },
1010
+ eventCount: 1,
1011
+ modelCallCount: 0
1012
+ }
1013
+ }
1014
+ }
1015
+ },
1016
+ {
1017
+ baseline: {
1018
+ digest: "sha256:9795afbb773b45c2d794522ad8c1fde600feda44cfd978572d62b93e9afb5724",
1019
+ kind: "agent-candidate-execution-evidence",
1020
+ materializationReceipt: {
1021
+ benchmark: {
1022
+ suite: {
1023
+ digest: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5",
1024
+ material: {
1025
+ byteLength: 198,
1026
+ content: "eyJkaWdlc3RBbGdvcml0aG0iOiJyZmM4Nzg1LXNoYTI1NiIsImtpbmQiOiJhZ2VudC1jYW5kaWRhdGUtYmVuY2htYXJrLXN1aXRlIiwicmVwcyI6Mywic2VlZHMiOlsxMDEsMTAyLDEwM10sInRhc2tEaWdlc3RzIjpbInNoYTI1Njo2YWE3MTgzMzFiNmRhMzEwZGNhYjZkZjJlMzM5N2NiNzcwODQwZTU1NWE4MmUwNjg5NTc3MTU3Yjg5YjRlMjc3Il19",
1027
+ encoding: "base64",
1028
+ sha256: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5"
1029
+ }
1030
+ },
1031
+ task: {
1032
+ digest: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277",
1033
+ material: {
1034
+ byteLength: 2096,
1035
+ content: "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",
1036
+ encoding: "base64",
1037
+ sha256: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277"
1038
+ }
1039
+ }
1040
+ },
1041
+ bundleDigest: "sha256:5c21ee53e513fc604cb09754e21c392b24a424da0ef37dbf8f1ee4a8a0b08f09",
1042
+ codeKind: "disabled",
1043
+ container: {
1044
+ image: "ghcr.io/example/task",
1045
+ indexDigest: "sha256:aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
1046
+ manifestDigest: "sha256:bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
1047
+ platform: {
1048
+ architecture: "amd64",
1049
+ os: "linux"
1050
+ },
1051
+ source: "evaluator-task-container"
1052
+ },
1053
+ digest: "sha256:0406c21aababc8747dd9de4bc66e83dbc1bb48ad55048168178dcaa4016dc48f",
1054
+ digestAlgorithm: "rfc8785-sha256",
1055
+ executionPlan: {
1056
+ artifact: {
1057
+ byteLength: 2030,
1058
+ content: "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",
1059
+ encoding: "base64",
1060
+ sha256: "sha256:e719f546b4f6585879b1210ff9c2263859efc625b5d78cfbab46f5a180e29287"
1061
+ },
1062
+ digest: "sha256:e719f546b4f6585879b1210ff9c2263859efc625b5d78cfbab46f5a180e29287",
1063
+ kind: "agent-candidate-execution-plan",
1064
+ material: {
1065
+ codeKind: "disabled",
1066
+ container: {
1067
+ image: "ghcr.io/example/task",
1068
+ indexDigest: "sha256:aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
1069
+ manifestDigest: "sha256:bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
1070
+ platform: {
1071
+ architecture: "amd64",
1072
+ os: "linux"
1073
+ },
1074
+ source: "evaluator-task-container"
1075
+ },
1076
+ executionId: "experiment-baseline-0-1",
1077
+ harness: "codex",
1078
+ harnessVersion: "1.2.3",
1079
+ instructionDelivery: {
1080
+ kind: "stdin-utf8"
1081
+ },
1082
+ kind: "agent-candidate-execution-plan-material",
1083
+ launch: {
1084
+ args: [],
1085
+ cwd: {
1086
+ path: ".",
1087
+ workspace: "task"
1088
+ },
1089
+ env: {
1090
+ PATH: {
1091
+ kind: "public",
1092
+ value: "/usr/local/bin:/usr/bin:/bin"
1093
+ }
1094
+ },
1095
+ executable: "codex"
1096
+ },
1097
+ limits: {
1098
+ maxCostUsd: 5,
1099
+ maxInputTokens: 1e5,
1100
+ maxModelCalls: 50,
1101
+ maxOutputTokens: 5e4,
1102
+ maxSteps: 100,
1103
+ timeoutMs: 6e4
1104
+ },
1105
+ memory: {
1106
+ mode: "disabled"
1107
+ },
1108
+ model: {
1109
+ access: {
1110
+ grantDigest: "sha256:cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc",
1111
+ kind: "evaluator-mediated",
1112
+ network: {
1113
+ domains: ["router.tangle.tools"],
1114
+ mode: "gateway-only"
1115
+ }
1116
+ },
1117
+ policy: "single",
1118
+ resolved: {
1119
+ model: "model-snapshot",
1120
+ provider: "provider",
1121
+ reasoningEffort: "high",
1122
+ requested: "provider/model",
1123
+ snapshot: "model-snapshot-2026-07-01"
1124
+ },
1125
+ routes: [
1126
+ {
1127
+ kind: "primary",
1128
+ requested: "provider/model"
1129
+ }
1130
+ ]
1131
+ },
1132
+ network: {
1133
+ mode: "disabled"
1134
+ },
1135
+ profile: {
1136
+ mountPaths: ["AGENTS.md"],
1137
+ planDigest: "sha256:af65bd67cf53561cd7bfac4e8c913ef18e36c0c2fe20a0bca1df55e59b21fab5",
1138
+ targetWorkspace: "task"
1139
+ },
1140
+ runCell: {
1141
+ arm: "baseline",
1142
+ attempt: 1,
1143
+ bundleDigest: "sha256:5c21ee53e513fc604cb09754e21c392b24a424da0ef37dbf8f1ee4a8a0b08f09",
1144
+ digest: "sha256:1c64f3d18b9d8c5a34231c5c7c2d3496e225051c9f2930ca97d01176492e8e76",
1145
+ experimentDigest: "sha256:c47a0dad80e6cc4856cdb8a02a279da621ab7d9409e3127650b67f1fc1c14b2d",
1146
+ kind: "agent-candidate-run-cell",
1147
+ repetition: 1,
1148
+ seed: 102,
1149
+ suiteDigest: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5",
1150
+ taskDigest: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277",
1151
+ taskIndex: 0
1152
+ },
1153
+ workspaces: {
1154
+ taskRoot: "/workspace/task"
1155
+ }
1156
+ }
1157
+ },
1158
+ harness: "codex",
1159
+ harnessVersion: "1.2.3",
1160
+ kind: "agent-candidate-materialization",
1161
+ profileActivation: {
1162
+ digest: "sha256:273b700e4080e7e6cc4bd58a755466d18ce323eaaa195bb4202fea456422bb66",
1163
+ files: [
1164
+ {
1165
+ content: "Inspect the repository, implement the fix, and run tests.\n",
1166
+ mode: 420,
1167
+ path: "AGENTS.md"
1168
+ }
1169
+ ],
1170
+ kind: "agent-candidate-profile-activation",
1171
+ profilePlan: {
1172
+ artifact: {
1173
+ byteLength: 287,
1174
+ content: "eyJlbnYiOnt9LCJmaWxlcyI6W3siY29udGVudFNoYTI1NiI6InNoYTI1NjozOWFkYmI5YjdmNTcxZjBmZDU4NmU3Nzk1MDIzZDI1OTA5YWRiZTk1ODZiYjIxZWUxMzRkM2Q3MjQ4YmYxNTk5IiwibW9kZSI6NDIwLCJyZWxQYXRoIjoiQUdFTlRTLm1kIn1dLCJmbGFncyI6W10sImhhcm5lc3MiOiJjb2RleCIsInNvdXJjZVByb2ZpbGVEaWdlc3QiOiJzaGEyNTY6ZTUzZTFjOGY5NjMyYmNiNWRiOTAzM2UzMmYzM2E5MGI5YmQ5OWFiZWNlZjBiYjUyZmY5MDY5ZjA3NTA5ZTJmOSIsInVuc3VwcG9ydGVkIjpbXX0=",
1175
+ encoding: "base64",
1176
+ sha256: "sha256:af65bd67cf53561cd7bfac4e8c913ef18e36c0c2fe20a0bca1df55e59b21fab5"
1177
+ },
1178
+ digest: "sha256:af65bd67cf53561cd7bfac4e8c913ef18e36c0c2fe20a0bca1df55e59b21fab5",
1179
+ kind: "agent-profile-workspace-plan",
1180
+ material: {
1181
+ env: {},
1182
+ files: [
1183
+ {
1184
+ contentSha256: "sha256:39adbb9b7f571f0fd586e7795023d25909adbe9586bb21ee134d3d7248bf1599",
1185
+ mode: 420,
1186
+ relPath: "AGENTS.md"
1187
+ }
1188
+ ],
1189
+ flags: [],
1190
+ harness: "codex",
1191
+ sourceProfileDigest: "sha256:e53e1c8f9632bcb5db9033e32f33a90b9bd99abecef0bb52ff9069f07509e2f9",
1192
+ unsupported: []
1193
+ }
1194
+ }
1195
+ },
1196
+ resolvedModel: {
1197
+ model: "model-snapshot",
1198
+ provider: "provider",
1199
+ reasoningEffort: "high",
1200
+ requested: "provider/model",
1201
+ snapshot: "model-snapshot-2026-07-01"
1202
+ }
1203
+ },
1204
+ receipt: {
1205
+ benchmarkResult: {
1206
+ artifact: {
1207
+ byteLength: 1122,
1208
+ locator: {
1209
+ bucket: "candidate-test-artifacts",
1210
+ key: "benchmark-result/c9891b43d8cc45bb7b1d76eccce93421af4c329983e4d2cbfd4411df4cd228da",
1211
+ kind: "s3"
1212
+ },
1213
+ sha256: "sha256:c9891b43d8cc45bb7b1d76eccce93421af4c329983e4d2cbfd4411df4cd228da"
1214
+ },
1215
+ digest: "sha256:c9891b43d8cc45bb7b1d76eccce93421af4c329983e4d2cbfd4411df4cd228da",
1216
+ kind: "agent-candidate-benchmark-result",
1217
+ material: {
1218
+ dimensions: [
1219
+ {
1220
+ name: "quality",
1221
+ score: 0
1222
+ }
1223
+ ],
1224
+ evidence: {
1225
+ byteLength: 11,
1226
+ locator: {
1227
+ bucket: "candidate-test-artifacts",
1228
+ key: "grader-evidence/628cb4ead38e34e34a34f78fa4481f2a0ccf8e9139acfae69d79dc997e90e36b",
1229
+ kind: "s3"
1230
+ },
1231
+ sha256: "sha256:628cb4ead38e34e34a34f78fa4481f2a0ccf8e9139acfae69d79dc997e90e36b"
1232
+ },
1233
+ executionPlanDigest: "sha256:e719f546b4f6585879b1210ff9c2263859efc625b5d78cfbab46f5a180e29287",
1234
+ grader: {
1235
+ artifact: {
1236
+ byteLength: 25,
1237
+ locator: {
1238
+ bucket: "candidate-test-artifacts",
1239
+ key: "grader/9a354495185587b25e8c98104060068871a44c719e4f57d345a905fc2d72a6ad",
1240
+ kind: "s3"
1241
+ },
1242
+ sha256: "sha256:9a354495185587b25e8c98104060068871a44c719e4f57d345a905fc2d72a6ad"
1243
+ },
1244
+ format: "tangle-grader",
1245
+ name: "fixture-executable-grader",
1246
+ version: "1.0.0"
1247
+ },
1248
+ grading: {
1249
+ timing: {
1250
+ durationMs: 0,
1251
+ endedAtMs: 17836416e5,
1252
+ startedAtMs: 17836416e5
1253
+ },
1254
+ usage: {
1255
+ cachedInputTokens: 0,
1256
+ costUsdNanos: 0,
1257
+ inputTokens: 0,
1258
+ modelCalls: 0,
1259
+ outputTokens: 0,
1260
+ reasoningTokens: 0
1261
+ }
1262
+ },
1263
+ kind: "agent-candidate-benchmark-result-material",
1264
+ passed: false,
1265
+ score: 0,
1266
+ taskOutcomeDigest: "sha256:24cd5d7f59ab6334118cd341ad8f8fd4d20156a561ddb2b55fe86c61c2895aa2"
1267
+ }
1268
+ },
1269
+ bundleDigest: "sha256:5c21ee53e513fc604cb09754e21c392b24a424da0ef37dbf8f1ee4a8a0b08f09",
1270
+ digest: "sha256:362e0b403b592b762ec92734f1286ecaf0c08e40970525de14b3052b335b8618",
1271
+ digestAlgorithm: "rfc8785-sha256",
1272
+ executionPlanDigest: "sha256:e719f546b4f6585879b1210ff9c2263859efc625b5d78cfbab46f5a180e29287",
1273
+ executorCapture: {
1274
+ byteLength: 477,
1275
+ locator: {
1276
+ bucket: "candidate-test-artifacts",
1277
+ key: "executor-capture/da5fc6561b02133890cdfb4cdc0b3495d17059040b27c1be20b8a123d2a6524a",
1278
+ kind: "s3"
1279
+ },
1280
+ sha256: "sha256:da5fc6561b02133890cdfb4cdc0b3495d17059040b27c1be20b8a123d2a6524a"
1281
+ },
1282
+ kind: "agent-candidate-run",
1283
+ materializationReceiptDigest: "sha256:0406c21aababc8747dd9de4bc66e83dbc1bb48ad55048168178dcaa4016dc48f",
1284
+ memory: {
1285
+ mode: "disabled"
1286
+ },
1287
+ modelSettlement: {
1288
+ artifact: {
1289
+ byteLength: 615,
1290
+ locator: {
1291
+ bucket: "candidate-test-artifacts",
1292
+ key: "model-settlement/446db011d1af4c39f2d0fd21a7e64774b9aaa685efe56e928d8e67d8ea1b3fa3",
1293
+ kind: "s3"
1294
+ },
1295
+ sha256: "sha256:446db011d1af4c39f2d0fd21a7e64774b9aaa685efe56e928d8e67d8ea1b3fa3"
1296
+ },
1297
+ digest: "sha256:446db011d1af4c39f2d0fd21a7e64774b9aaa685efe56e928d8e67d8ea1b3fa3",
1298
+ kind: "agent-candidate-model-settlement",
1299
+ material: {
1300
+ calls: [],
1301
+ closed: true,
1302
+ executionPlanDigest: "sha256:e719f546b4f6585879b1210ff9c2263859efc625b5d78cfbab46f5a180e29287",
1303
+ grantDigest: "sha256:cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc",
1304
+ kind: "agent-candidate-model-settlement-material",
1305
+ preparationId: "candidate-preparation.nBqsDbKYlsiY4Tto5XaxkR3RCTop9PA9_kX-kOSzR8U",
1306
+ resolved: {
1307
+ model: "model-snapshot",
1308
+ provider: "provider",
1309
+ reasoningEffort: "high",
1310
+ requested: "provider/model",
1311
+ snapshot: "model-snapshot-2026-07-01"
1312
+ },
1313
+ usage: {
1314
+ cachedInputTokens: 0,
1315
+ costUsdNanos: 0,
1316
+ inputTokens: 0,
1317
+ modelCalls: 0,
1318
+ outputTokens: 0,
1319
+ reasoningTokens: 0
1320
+ }
1321
+ }
1322
+ },
1323
+ runCellDigest: "sha256:1c64f3d18b9d8c5a34231c5c7c2d3496e225051c9f2930ca97d01176492e8e76",
1324
+ taskOutcome: {
1325
+ artifact: {
1326
+ byteLength: 478,
1327
+ locator: {
1328
+ bucket: "candidate-test-artifacts",
1329
+ key: "task-outcome/24cd5d7f59ab6334118cd341ad8f8fd4d20156a561ddb2b55fe86c61c2895aa2",
1330
+ kind: "s3"
1331
+ },
1332
+ sha256: "sha256:24cd5d7f59ab6334118cd341ad8f8fd4d20156a561ddb2b55fe86c61c2895aa2"
1333
+ },
1334
+ digest: "sha256:24cd5d7f59ab6334118cd341ad8f8fd4d20156a561ddb2b55fe86c61c2895aa2",
1335
+ kind: "agent-candidate-task-outcome",
1336
+ material: {
1337
+ executionPlanDigest: "sha256:e719f546b4f6585879b1210ff9c2263859efc625b5d78cfbab46f5a180e29287",
1338
+ kind: "agent-candidate-task-outcome-material",
1339
+ outcome: {
1340
+ artifact: {
1341
+ byteLength: 4,
1342
+ locator: {
1343
+ bucket: "candidate-test-artifacts",
1344
+ key: "task-output/481ba4019c9d2710c3386537382592c093ef02bf5b056c30237b3d92b0e19a1d",
1345
+ kind: "s3"
1346
+ },
1347
+ sha256: "sha256:481ba4019c9d2710c3386537382592c093ef02bf5b056c30237b3d92b0e19a1d"
1348
+ },
1349
+ kind: "output",
1350
+ spec: {
1351
+ maxBytes: 1024,
1352
+ mediaType: "text/plain"
1353
+ }
1354
+ }
1355
+ }
1356
+ },
1357
+ termination: {
1358
+ exitCode: 0,
1359
+ kind: "exit"
1360
+ },
1361
+ timing: {
1362
+ durationMs: 50,
1363
+ endedAtMs: 1150,
1364
+ startedAtMs: 1100
1365
+ },
1366
+ trace: {
1367
+ artifact: {
1368
+ byteLength: 748,
1369
+ locator: {
1370
+ bucket: "candidate-test-artifacts",
1371
+ key: "trace/c03c3bd51fb421ffc65985a27619e57b16623ba3f2bc8d22b2f963cdae6e41f6",
1372
+ kind: "s3"
1373
+ },
1374
+ sha256: "sha256:c03c3bd51fb421ffc65985a27619e57b16623ba3f2bc8d22b2f963cdae6e41f6"
1375
+ },
1376
+ eventCount: 1,
1377
+ modelCallCount: 0
1378
+ }
1379
+ }
1380
+ },
1381
+ candidate: {
1382
+ digest: "sha256:90aa644ce8a4a7bd8968146189ec279b74c550ae6ff02b08adfa984008b68ce5",
1383
+ kind: "agent-candidate-execution-evidence",
1384
+ materializationReceipt: {
1385
+ benchmark: {
1386
+ suite: {
1387
+ digest: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5",
1388
+ material: {
1389
+ byteLength: 198,
1390
+ content: "eyJkaWdlc3RBbGdvcml0aG0iOiJyZmM4Nzg1LXNoYTI1NiIsImtpbmQiOiJhZ2VudC1jYW5kaWRhdGUtYmVuY2htYXJrLXN1aXRlIiwicmVwcyI6Mywic2VlZHMiOlsxMDEsMTAyLDEwM10sInRhc2tEaWdlc3RzIjpbInNoYTI1Njo2YWE3MTgzMzFiNmRhMzEwZGNhYjZkZjJlMzM5N2NiNzcwODQwZTU1NWE4MmUwNjg5NTc3MTU3Yjg5YjRlMjc3Il19",
1391
+ encoding: "base64",
1392
+ sha256: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5"
1393
+ }
1394
+ },
1395
+ task: {
1396
+ digest: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277",
1397
+ material: {
1398
+ byteLength: 2096,
1399
+ content: "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",
1400
+ encoding: "base64",
1401
+ sha256: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277"
1402
+ }
1403
+ }
1404
+ },
1405
+ bundleDigest: "sha256:60fcbb1c728194bd51d7d19cb732d1c3f1881dce7e0a6266b41c8b98cfd65693",
1406
+ codeKind: "disabled",
1407
+ container: {
1408
+ image: "ghcr.io/example/task",
1409
+ indexDigest: "sha256:aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
1410
+ manifestDigest: "sha256:bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
1411
+ platform: {
1412
+ architecture: "amd64",
1413
+ os: "linux"
1414
+ },
1415
+ source: "evaluator-task-container"
1416
+ },
1417
+ digest: "sha256:d6e837e515c14c86a79f6fb099384a0ac97d74cbfebc2e1a095a8b7660ef1fed",
1418
+ digestAlgorithm: "rfc8785-sha256",
1419
+ executionPlan: {
1420
+ artifact: {
1421
+ byteLength: 2032,
1422
+ content: "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",
1423
+ encoding: "base64",
1424
+ sha256: "sha256:8d12d5d6466c0dd8f2592be9f6d66b5b6a2139bfe8cdc5b793bde9e28ebec4c6"
1425
+ },
1426
+ digest: "sha256:8d12d5d6466c0dd8f2592be9f6d66b5b6a2139bfe8cdc5b793bde9e28ebec4c6",
1427
+ kind: "agent-candidate-execution-plan",
1428
+ material: {
1429
+ codeKind: "disabled",
1430
+ container: {
1431
+ image: "ghcr.io/example/task",
1432
+ indexDigest: "sha256:aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
1433
+ manifestDigest: "sha256:bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
1434
+ platform: {
1435
+ architecture: "amd64",
1436
+ os: "linux"
1437
+ },
1438
+ source: "evaluator-task-container"
1439
+ },
1440
+ executionId: "experiment-candidate-0-1",
1441
+ harness: "codex",
1442
+ harnessVersion: "1.2.3",
1443
+ instructionDelivery: {
1444
+ kind: "stdin-utf8"
1445
+ },
1446
+ kind: "agent-candidate-execution-plan-material",
1447
+ launch: {
1448
+ args: [],
1449
+ cwd: {
1450
+ path: ".",
1451
+ workspace: "task"
1452
+ },
1453
+ env: {
1454
+ PATH: {
1455
+ kind: "public",
1456
+ value: "/usr/local/bin:/usr/bin:/bin"
1457
+ }
1458
+ },
1459
+ executable: "codex"
1460
+ },
1461
+ limits: {
1462
+ maxCostUsd: 5,
1463
+ maxInputTokens: 1e5,
1464
+ maxModelCalls: 50,
1465
+ maxOutputTokens: 5e4,
1466
+ maxSteps: 100,
1467
+ timeoutMs: 6e4
1468
+ },
1469
+ memory: {
1470
+ mode: "disabled"
1471
+ },
1472
+ model: {
1473
+ access: {
1474
+ grantDigest: "sha256:cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc",
1475
+ kind: "evaluator-mediated",
1476
+ network: {
1477
+ domains: ["router.tangle.tools"],
1478
+ mode: "gateway-only"
1479
+ }
1480
+ },
1481
+ policy: "single",
1482
+ resolved: {
1483
+ model: "model-snapshot",
1484
+ provider: "provider",
1485
+ reasoningEffort: "high",
1486
+ requested: "provider/model",
1487
+ snapshot: "model-snapshot-2026-07-01"
1488
+ },
1489
+ routes: [
1490
+ {
1491
+ kind: "primary",
1492
+ requested: "provider/model"
1493
+ }
1494
+ ]
1495
+ },
1496
+ network: {
1497
+ mode: "disabled"
1498
+ },
1499
+ profile: {
1500
+ mountPaths: ["AGENTS.md"],
1501
+ planDigest: "sha256:4b4214d1bae911df6dda3a5d8585c65df2f0b2380126eb40c0cbd46a4a2fba5b",
1502
+ targetWorkspace: "task"
1503
+ },
1504
+ runCell: {
1505
+ arm: "candidate",
1506
+ attempt: 1,
1507
+ bundleDigest: "sha256:60fcbb1c728194bd51d7d19cb732d1c3f1881dce7e0a6266b41c8b98cfd65693",
1508
+ digest: "sha256:3056f0119ac4643dc3defb95fb324552e10f94e40621231d1f60d148f59d1d39",
1509
+ experimentDigest: "sha256:c47a0dad80e6cc4856cdb8a02a279da621ab7d9409e3127650b67f1fc1c14b2d",
1510
+ kind: "agent-candidate-run-cell",
1511
+ repetition: 1,
1512
+ seed: 102,
1513
+ suiteDigest: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5",
1514
+ taskDigest: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277",
1515
+ taskIndex: 0
1516
+ },
1517
+ workspaces: {
1518
+ taskRoot: "/workspace/task"
1519
+ }
1520
+ }
1521
+ },
1522
+ harness: "codex",
1523
+ harnessVersion: "1.2.3",
1524
+ kind: "agent-candidate-materialization",
1525
+ profileActivation: {
1526
+ digest: "sha256:3103f8282be70a843eada63410c41d9eb5ac7b824c758eed6c2f8610de169f45",
1527
+ files: [
1528
+ {
1529
+ content: "Inspect the repository, implement the fix, and run tests.\n",
1530
+ mode: 420,
1531
+ path: "AGENTS.md"
1532
+ }
1533
+ ],
1534
+ kind: "agent-candidate-profile-activation",
1535
+ profilePlan: {
1536
+ artifact: {
1537
+ byteLength: 364,
1538
+ content: "eyJlbnYiOnt9LCJmaWxlcyI6W3siY29udGVudFNoYTI1NiI6InNoYTI1NjozOWFkYmI5YjdmNTcxZjBmZDU4NmU3Nzk1MDIzZDI1OTA5YWRiZTk1ODZiYjIxZWUxMzRkM2Q3MjQ4YmYxNTk5IiwibW9kZSI6NDIwLCJyZWxQYXRoIjoiQUdFTlRTLm1kIn1dLCJmbGFncyI6W10sImhhcm5lc3MiOiJjb2RleCIsInNvdXJjZVByb2ZpbGVEaWdlc3QiOiJzaGEyNTY6OGIxZTJmMGY3YjgzMjk1ODM4MWMyOWJhZWJlNGE5NjFiMTUxYWY5N2EwY2U5ZDYxMWExNWFkMDc0MWFlYWE3NSIsInN5c3RlbVByb21wdCI6eyJraW5kIjoicHVibGljIiwidmFsdWUiOiJSZXR1cm4gdGhlIGV4YWN0IG1lYXN1cmVkIGFuc3dlci4ifSwidW5zdXBwb3J0ZWQiOltdfQ==",
1539
+ encoding: "base64",
1540
+ sha256: "sha256:4b4214d1bae911df6dda3a5d8585c65df2f0b2380126eb40c0cbd46a4a2fba5b"
1541
+ },
1542
+ digest: "sha256:4b4214d1bae911df6dda3a5d8585c65df2f0b2380126eb40c0cbd46a4a2fba5b",
1543
+ kind: "agent-profile-workspace-plan",
1544
+ material: {
1545
+ env: {},
1546
+ files: [
1547
+ {
1548
+ contentSha256: "sha256:39adbb9b7f571f0fd586e7795023d25909adbe9586bb21ee134d3d7248bf1599",
1549
+ mode: 420,
1550
+ relPath: "AGENTS.md"
1551
+ }
1552
+ ],
1553
+ flags: [],
1554
+ harness: "codex",
1555
+ sourceProfileDigest: "sha256:8b1e2f0f7b832958381c29baebe4a961b151af97a0ce9d611a15ad0741aeaa75",
1556
+ systemPrompt: {
1557
+ kind: "public",
1558
+ value: "Return the exact measured answer."
1559
+ },
1560
+ unsupported: []
1561
+ }
1562
+ }
1563
+ },
1564
+ resolvedModel: {
1565
+ model: "model-snapshot",
1566
+ provider: "provider",
1567
+ reasoningEffort: "high",
1568
+ requested: "provider/model",
1569
+ snapshot: "model-snapshot-2026-07-01"
1570
+ }
1571
+ },
1572
+ receipt: {
1573
+ benchmarkResult: {
1574
+ artifact: {
1575
+ byteLength: 1121,
1576
+ locator: {
1577
+ bucket: "candidate-test-artifacts",
1578
+ key: "benchmark-result/8b1ecbcec1ebdb0ec02a3785ff61796b8de31c83830d446c560029dc127a45ca",
1579
+ kind: "s3"
1580
+ },
1581
+ sha256: "sha256:8b1ecbcec1ebdb0ec02a3785ff61796b8de31c83830d446c560029dc127a45ca"
1582
+ },
1583
+ digest: "sha256:8b1ecbcec1ebdb0ec02a3785ff61796b8de31c83830d446c560029dc127a45ca",
1584
+ kind: "agent-candidate-benchmark-result",
1585
+ material: {
1586
+ dimensions: [
1587
+ {
1588
+ name: "quality",
1589
+ score: 1
1590
+ }
1591
+ ],
1592
+ evidence: {
1593
+ byteLength: 11,
1594
+ locator: {
1595
+ bucket: "candidate-test-artifacts",
1596
+ key: "grader-evidence/9b9b3a1471309177261cfe65ea9c298e0dd372e4b5d087f8d35d7b732485373d",
1597
+ kind: "s3"
1598
+ },
1599
+ sha256: "sha256:9b9b3a1471309177261cfe65ea9c298e0dd372e4b5d087f8d35d7b732485373d"
1600
+ },
1601
+ executionPlanDigest: "sha256:8d12d5d6466c0dd8f2592be9f6d66b5b6a2139bfe8cdc5b793bde9e28ebec4c6",
1602
+ grader: {
1603
+ artifact: {
1604
+ byteLength: 25,
1605
+ locator: {
1606
+ bucket: "candidate-test-artifacts",
1607
+ key: "grader/9a354495185587b25e8c98104060068871a44c719e4f57d345a905fc2d72a6ad",
1608
+ kind: "s3"
1609
+ },
1610
+ sha256: "sha256:9a354495185587b25e8c98104060068871a44c719e4f57d345a905fc2d72a6ad"
1611
+ },
1612
+ format: "tangle-grader",
1613
+ name: "fixture-executable-grader",
1614
+ version: "1.0.0"
1615
+ },
1616
+ grading: {
1617
+ timing: {
1618
+ durationMs: 0,
1619
+ endedAtMs: 17836416e5,
1620
+ startedAtMs: 17836416e5
1621
+ },
1622
+ usage: {
1623
+ cachedInputTokens: 0,
1624
+ costUsdNanos: 0,
1625
+ inputTokens: 0,
1626
+ modelCalls: 0,
1627
+ outputTokens: 0,
1628
+ reasoningTokens: 0
1629
+ }
1630
+ },
1631
+ kind: "agent-candidate-benchmark-result-material",
1632
+ passed: true,
1633
+ score: 1,
1634
+ taskOutcomeDigest: "sha256:f24062b48a7ba4bfe3e2566ef7e0d4e97d3c38d7a54e86bf86446adf7fa1cef4"
1635
+ }
1636
+ },
1637
+ bundleDigest: "sha256:60fcbb1c728194bd51d7d19cb732d1c3f1881dce7e0a6266b41c8b98cfd65693",
1638
+ digest: "sha256:bbfcb8dfe7d34f06ce47c8a6562ab15fcaa1dc083524eb692890055e2e2ccc27",
1639
+ digestAlgorithm: "rfc8785-sha256",
1640
+ executionPlanDigest: "sha256:8d12d5d6466c0dd8f2592be9f6d66b5b6a2139bfe8cdc5b793bde9e28ebec4c6",
1641
+ executorCapture: {
1642
+ byteLength: 477,
1643
+ locator: {
1644
+ bucket: "candidate-test-artifacts",
1645
+ key: "executor-capture/5d5db54b6c1e1d5e8d3c1248b3617dcdc8cefb80aa69cf750a1abacbadf9efdb",
1646
+ kind: "s3"
1647
+ },
1648
+ sha256: "sha256:5d5db54b6c1e1d5e8d3c1248b3617dcdc8cefb80aa69cf750a1abacbadf9efdb"
1649
+ },
1650
+ kind: "agent-candidate-run",
1651
+ materializationReceiptDigest: "sha256:d6e837e515c14c86a79f6fb099384a0ac97d74cbfebc2e1a095a8b7660ef1fed",
1652
+ memory: {
1653
+ mode: "disabled"
1654
+ },
1655
+ modelSettlement: {
1656
+ artifact: {
1657
+ byteLength: 615,
1658
+ locator: {
1659
+ bucket: "candidate-test-artifacts",
1660
+ key: "model-settlement/87ad72eeccca5a0d4593d75b73d7cd4d0d3ffba748b30becc3b496e74ad5d83b",
1661
+ kind: "s3"
1662
+ },
1663
+ sha256: "sha256:87ad72eeccca5a0d4593d75b73d7cd4d0d3ffba748b30becc3b496e74ad5d83b"
1664
+ },
1665
+ digest: "sha256:87ad72eeccca5a0d4593d75b73d7cd4d0d3ffba748b30becc3b496e74ad5d83b",
1666
+ kind: "agent-candidate-model-settlement",
1667
+ material: {
1668
+ calls: [],
1669
+ closed: true,
1670
+ executionPlanDigest: "sha256:8d12d5d6466c0dd8f2592be9f6d66b5b6a2139bfe8cdc5b793bde9e28ebec4c6",
1671
+ grantDigest: "sha256:cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc",
1672
+ kind: "agent-candidate-model-settlement-material",
1673
+ preparationId: "candidate-preparation.nBqsDbKYlsiY4Tto5XaxkR3RCTop9PA9_kX-kOSzR8U",
1674
+ resolved: {
1675
+ model: "model-snapshot",
1676
+ provider: "provider",
1677
+ reasoningEffort: "high",
1678
+ requested: "provider/model",
1679
+ snapshot: "model-snapshot-2026-07-01"
1680
+ },
1681
+ usage: {
1682
+ cachedInputTokens: 0,
1683
+ costUsdNanos: 0,
1684
+ inputTokens: 0,
1685
+ modelCalls: 0,
1686
+ outputTokens: 0,
1687
+ reasoningTokens: 0
1688
+ }
1689
+ }
1690
+ },
1691
+ runCellDigest: "sha256:3056f0119ac4643dc3defb95fb324552e10f94e40621231d1f60d148f59d1d39",
1692
+ taskOutcome: {
1693
+ artifact: {
1694
+ byteLength: 478,
1695
+ locator: {
1696
+ bucket: "candidate-test-artifacts",
1697
+ key: "task-outcome/f24062b48a7ba4bfe3e2566ef7e0d4e97d3c38d7a54e86bf86446adf7fa1cef4",
1698
+ kind: "s3"
1699
+ },
1700
+ sha256: "sha256:f24062b48a7ba4bfe3e2566ef7e0d4e97d3c38d7a54e86bf86446adf7fa1cef4"
1701
+ },
1702
+ digest: "sha256:f24062b48a7ba4bfe3e2566ef7e0d4e97d3c38d7a54e86bf86446adf7fa1cef4",
1703
+ kind: "agent-candidate-task-outcome",
1704
+ material: {
1705
+ executionPlanDigest: "sha256:8d12d5d6466c0dd8f2592be9f6d66b5b6a2139bfe8cdc5b793bde9e28ebec4c6",
1706
+ kind: "agent-candidate-task-outcome-material",
1707
+ outcome: {
1708
+ artifact: {
1709
+ byteLength: 6,
1710
+ locator: {
1711
+ bucket: "candidate-test-artifacts",
1712
+ key: "task-output/eda15ce4834ac303532bc6973df5ff8a5deb75d7dfceb9fc2cdf1bb450a01cf5",
1713
+ kind: "s3"
1714
+ },
1715
+ sha256: "sha256:eda15ce4834ac303532bc6973df5ff8a5deb75d7dfceb9fc2cdf1bb450a01cf5"
1716
+ },
1717
+ kind: "output",
1718
+ spec: {
1719
+ maxBytes: 1024,
1720
+ mediaType: "text/plain"
1721
+ }
1722
+ }
1723
+ }
1724
+ },
1725
+ termination: {
1726
+ exitCode: 0,
1727
+ kind: "exit"
1728
+ },
1729
+ timing: {
1730
+ durationMs: 50,
1731
+ endedAtMs: 1150,
1732
+ startedAtMs: 1100
1733
+ },
1734
+ trace: {
1735
+ artifact: {
1736
+ byteLength: 750,
1737
+ locator: {
1738
+ bucket: "candidate-test-artifacts",
1739
+ key: "trace/e2cdd682a809a037e0049d1b3d3280d7b4099a3e0507b3591fc05b2872641166",
1740
+ kind: "s3"
1741
+ },
1742
+ sha256: "sha256:e2cdd682a809a037e0049d1b3d3280d7b4099a3e0507b3591fc05b2872641166"
1743
+ },
1744
+ eventCount: 1,
1745
+ modelCallCount: 0
1746
+ }
1747
+ }
1748
+ }
1749
+ },
1750
+ {
1751
+ baseline: {
1752
+ digest: "sha256:7ebfc740cb6c0156ce54d450015af9e418a57473299022f926555f2a5b706454",
1753
+ kind: "agent-candidate-execution-evidence",
1754
+ materializationReceipt: {
1755
+ benchmark: {
1756
+ suite: {
1757
+ digest: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5",
1758
+ material: {
1759
+ byteLength: 198,
1760
+ content: "eyJkaWdlc3RBbGdvcml0aG0iOiJyZmM4Nzg1LXNoYTI1NiIsImtpbmQiOiJhZ2VudC1jYW5kaWRhdGUtYmVuY2htYXJrLXN1aXRlIiwicmVwcyI6Mywic2VlZHMiOlsxMDEsMTAyLDEwM10sInRhc2tEaWdlc3RzIjpbInNoYTI1Njo2YWE3MTgzMzFiNmRhMzEwZGNhYjZkZjJlMzM5N2NiNzcwODQwZTU1NWE4MmUwNjg5NTc3MTU3Yjg5YjRlMjc3Il19",
1761
+ encoding: "base64",
1762
+ sha256: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5"
1763
+ }
1764
+ },
1765
+ task: {
1766
+ digest: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277",
1767
+ material: {
1768
+ byteLength: 2096,
1769
+ content: "eyJhdHRlbXB0Ijp7Im1heEF0dGVtcHRzIjoxLCJyZXRyeVBvbGljeSI6Im5vbmUifSwiYmVuY2htYXJrIjp7Im5hbWUiOiJyZXBvc2l0b3J5LWRpc2pvaW50LXNtb2tlIiwic3BsaXREaWdlc3QiOiJzaGEyNTY6ZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZGRkZCIsInZlcnNpb24iOiIxIn0sImRpZ2VzdEFsZ29yaXRobSI6InJmYzg3ODUtc2hhMjU2IiwiZXZhbHVhdG9yVGFza0NvbnRhaW5lciI6eyJpbWFnZSI6ImdoY3IuaW8vZXhhbXBsZS90YXNrIiwiaW5kZXhEaWdlc3QiOiJzaGEyNTY6YWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYSIsIm1hbmlmZXN0RGlnZXN0Ijoic2hhMjU2OmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmIiLCJwbGF0Zm9ybSI6eyJhcmNoaXRlY3R1cmUiOiJhbWQ2NCIsIm9zIjoibGludXgifSwic291cmNlIjoiZXZhbHVhdG9yLXRhc2stY29udGFpbmVyIn0sImdyYWRlciI6eyJhcnRpZmFjdCI6eyJieXRlTGVuZ3RoIjoyNSwibG9jYXRvciI6eyJidWNrZXQiOiJjYW5kaWRhdGUtdGVzdC1hcnRpZmFjdHMiLCJrZXkiOiJncmFkZXIvOWEzNTQ0OTUxODU1ODdiMjVlOGM5ODEwNDA2MDA2ODg3MWE0NGM3MTllNGY1N2QzNDVhOTA1ZmMyZDcyYTZhZCIsImtpbmQiOiJzMyJ9LCJzaGEyNTYiOiJzaGEyNTY6OWEzNTQ0OTUxODU1ODdiMjVlOGM5ODEwNDA2MDA2ODg3MWE0NGM3MTllNGY1N2QzNDVhOTA1ZmMyZDcyYTZhZCJ9LCJmb3JtYXQiOiJ0YW5nbGUtZ3JhZGVyIiwibmFtZSI6ImZpeHR1cmUtZXhlY3V0YWJsZS1ncmFkZXIiLCJ2ZXJzaW9uIjoiMS4wLjAifSwiaW5zdHJ1Y3Rpb24iOiJGaXggdGhlIGZhaWxpbmcgYmVoYXZpb3Igd2l0aG91dCBjaGFuZ2luZyB0aGUgcHVibGljIEFQSS4iLCJraW5kIjoiYWdlbnQtY2FuZGlkYXRlLWJlbmNobWFyay10YXNrIiwibGltaXRzIjp7Im1heENvc3RVc2QiOjUsIm1heElucHV0VG9rZW5zIjoxMDAwMDAsIm1heE1vZGVsQ2FsbHMiOjUwLCJtYXhPdXRwdXRUb2tlbnMiOjUwMDAwLCJtYXhTdGVwcyI6MTAwLCJ0aW1lb3V0TXMiOjYwMDAwfSwibW9kZWwiOnsibW9kZWwiOiJtb2RlbC1zbmFwc2hvdCIsInByb3ZpZGVyIjoicHJvdmlkZXIiLCJyZWFzb25pbmdFZmZvcnQiOiJoaWdoIiwicmVxdWVzdGVkIjoicHJvdmlkZXIvbW9kZWwiLCJzbmFwc2hvdCI6Im1vZGVsLXNuYXBzaG90LTIwMjYtMDctMDEifSwib3V0Y29tZSI6eyJraW5kIjoib3V0cHV0IiwibWF4Qnl0ZXMiOjEwMjQsIm1lZGlhVHlwZSI6InRleHQvcGxhaW4ifSwic2NlbmFyaW8iOnsiaWQiOiJvd25lci1yZXBvLTEiLCJraW5kIjoiY29kaW5nIiwic2NlbmFyaW9EaWdlc3QiOiJzaGEyNTY6ODg4ODg4ODg4ODg4ODg4ODg4ODg4ODg4ODg4ODg4ODg4ODg4ODg4ODg4ODg4ODg4ODg4ODg4ODg4ODg4ODg4OCJ9LCJ3b3Jrc3BhY2UiOnsiYXJjaGl2ZSI6eyJieXRlTGVuZ3RoIjoxNywiY29udGVudCI6IlpXMXdkSGs2YjNWMGNIVjBMWFJoYzJzPSIsImVuY29kaW5nIjoiYmFzZTY0Iiwic2hhMjU2Ijoic2hhMjU2OjMzMmRkOWM3NGJlYWI4YTE3NzM5NmE0NTEwZGFiNDZhNTBjYTIwNzRiMjAzYjEwNjk0ZTRlYWEzMDcxZTJlZjAifSwiZGlnZXN0Ijoic2hhMjU2OjNlOWZkYWQ0NmYxNGNlNWI3Yzc4YTkzZTRmZWUwOTYwNDgzYWU5ZDFjZGEyMWYwYzM5MGNkYmFkNDBiZGFjMDIiLCJraW5kIjoiYWdlbnQtY2FuZGlkYXRlLXdvcmtzcGFjZS1zbmFwc2hvdCIsIm1hbmlmZXN0Ijp7ImJ5dGVMZW5ndGgiOjU2LCJjb250ZW50IjoiZXlKbWFXeGxjeUk2VzEwc0ltdHBibVFpT2lKaFoyVnVkQzFqWVc1a2FXUmhkR1V0ZDI5eWEzTndZV05sTFcxaGJtbG1aWE4wSW4wPSIsImVuY29kaW5nIjoiYmFzZTY0Iiwic2hhMjU2Ijoic2hhMjU2OjNlOWZkYWQ0NmYxNGNlNWI3Yzc4YTkzZTRmZWUwOTYwNDgzYWU5ZDFjZGEyMWYwYzM5MGNkYmFkNDBiZGFjMDIifSwibWF0ZXJpYWwiOnsiZmlsZXMiOltdLCJraW5kIjoiYWdlbnQtY2FuZGlkYXRlLXdvcmtzcGFjZS1tYW5pZmVzdCJ9fX0=",
1770
+ encoding: "base64",
1771
+ sha256: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277"
1772
+ }
1773
+ }
1774
+ },
1775
+ bundleDigest: "sha256:5c21ee53e513fc604cb09754e21c392b24a424da0ef37dbf8f1ee4a8a0b08f09",
1776
+ codeKind: "disabled",
1777
+ container: {
1778
+ image: "ghcr.io/example/task",
1779
+ indexDigest: "sha256:aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
1780
+ manifestDigest: "sha256:bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
1781
+ platform: {
1782
+ architecture: "amd64",
1783
+ os: "linux"
1784
+ },
1785
+ source: "evaluator-task-container"
1786
+ },
1787
+ digest: "sha256:0e3da90681970b46a99808f5d5a623329bae9529bf0f46ccebe7cac7a202ec8a",
1788
+ digestAlgorithm: "rfc8785-sha256",
1789
+ executionPlan: {
1790
+ artifact: {
1791
+ byteLength: 2030,
1792
+ content: "eyJjb2RlS2luZCI6ImRpc2FibGVkIiwiY29udGFpbmVyIjp7ImltYWdlIjoiZ2hjci5pby9leGFtcGxlL3Rhc2siLCJpbmRleERpZ2VzdCI6InNoYTI1NjphYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhYWFhIiwibWFuaWZlc3REaWdlc3QiOiJzaGEyNTY6YmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYmJiYiIsInBsYXRmb3JtIjp7ImFyY2hpdGVjdHVyZSI6ImFtZDY0Iiwib3MiOiJsaW51eCJ9LCJzb3VyY2UiOiJldmFsdWF0b3ItdGFzay1jb250YWluZXIifSwiZXhlY3V0aW9uSWQiOiJleHBlcmltZW50LWJhc2VsaW5lLTAtMiIsImhhcm5lc3MiOiJjb2RleCIsImhhcm5lc3NWZXJzaW9uIjoiMS4yLjMiLCJpbnN0cnVjdGlvbkRlbGl2ZXJ5Ijp7ImtpbmQiOiJzdGRpbi11dGY4In0sImtpbmQiOiJhZ2VudC1jYW5kaWRhdGUtZXhlY3V0aW9uLXBsYW4tbWF0ZXJpYWwiLCJsYXVuY2giOnsiYXJncyI6W10sImN3ZCI6eyJwYXRoIjoiLiIsIndvcmtzcGFjZSI6InRhc2sifSwiZW52Ijp7IlBBVEgiOnsia2luZCI6InB1YmxpYyIsInZhbHVlIjoiL3Vzci9sb2NhbC9iaW46L3Vzci9iaW46L2JpbiJ9fSwiZXhlY3V0YWJsZSI6ImNvZGV4In0sImxpbWl0cyI6eyJtYXhDb3N0VXNkIjo1LCJtYXhJbnB1dFRva2VucyI6MTAwMDAwLCJtYXhNb2RlbENhbGxzIjo1MCwibWF4T3V0cHV0VG9rZW5zIjo1MDAwMCwibWF4U3RlcHMiOjEwMCwidGltZW91dE1zIjo2MDAwMH0sIm1lbW9yeSI6eyJtb2RlIjoiZGlzYWJsZWQifSwibW9kZWwiOnsiYWNjZXNzIjp7ImdyYW50RGlnZXN0Ijoic2hhMjU2OmNjY2NjY2NjY2NjY2NjY2NjY2NjY2NjY2NjY2NjY2NjY2NjY2NjY2NjY2NjY2NjY2NjY2NjY2NjY2NjY2NjY2MiLCJraW5kIjoiZXZhbHVhdG9yLW1lZGlhdGVkIiwibmV0d29yayI6eyJkb21haW5zIjpbInJvdXRlci50YW5nbGUudG9vbHMiXSwibW9kZSI6ImdhdGV3YXktb25seSJ9fSwicG9saWN5Ijoic2luZ2xlIiwicmVzb2x2ZWQiOnsibW9kZWwiOiJtb2RlbC1zbmFwc2hvdCIsInByb3ZpZGVyIjoicHJvdmlkZXIiLCJyZWFzb25pbmdFZmZvcnQiOiJoaWdoIiwicmVxdWVzdGVkIjoicHJvdmlkZXIvbW9kZWwiLCJzbmFwc2hvdCI6Im1vZGVsLXNuYXBzaG90LTIwMjYtMDctMDEifSwicm91dGVzIjpbeyJraW5kIjoicHJpbWFyeSIsInJlcXVlc3RlZCI6InByb3ZpZGVyL21vZGVsIn1dfSwibmV0d29yayI6eyJtb2RlIjoiZGlzYWJsZWQifSwicHJvZmlsZSI6eyJtb3VudFBhdGhzIjpbIkFHRU5UUy5tZCJdLCJwbGFuRGlnZXN0Ijoic2hhMjU2OmFmNjViZDY3Y2Y1MzU2MWNkN2JmYWM0ZThjOTEzZWYxOGUzNmMwYzJmZTIwYTBiY2ExZGY1NWU1OWIyMWZhYjUiLCJ0YXJnZXRXb3Jrc3BhY2UiOiJ0YXNrIn0sInJ1bkNlbGwiOnsiYXJtIjoiYmFzZWxpbmUiLCJhdHRlbXB0IjoxLCJidW5kbGVEaWdlc3QiOiJzaGEyNTY6NWMyMWVlNTNlNTEzZmM2MDRjYjA5NzU0ZTIxYzM5MmIyNGE0MjRkYTBlZjM3ZGJmOGYxZWU0YThhMGIwOGYwOSIsImRpZ2VzdCI6InNoYTI1Njo5YzQyN2NjNjc1MDdjZGE4ZGNkYWMzMjdmMzkwYWRmYTdlMDQ1ZDA2OWU2NzdkNjI5MjE2NzhiNDZhOTZkNjZjIiwiZXhwZXJpbWVudERpZ2VzdCI6InNoYTI1NjpjNDdhMGRhZDgwZTZjYzQ4NTZjZGI4YTAyYTI3OWRhNjIxYWI3ZDk0MDllMzEyNzY1MGI2N2YxZmMxYzE0YjJkIiwia2luZCI6ImFnZW50LWNhbmRpZGF0ZS1ydW4tY2VsbCIsInJlcGV0aXRpb24iOjIsInNlZWQiOjEwMywic3VpdGVEaWdlc3QiOiJzaGEyNTY6OTYzZDZhZDdhYmUzNTljOTYzZjMxYzIxMDM2YWM2MTc2ZmU3ZDVmMDk3YjIyMDBhYjQ0MzBjMGQxYmYxYmNkNSIsInRhc2tEaWdlc3QiOiJzaGEyNTY6NmFhNzE4MzMxYjZkYTMxMGRjYWI2ZGYyZTMzOTdjYjc3MDg0MGU1NTVhODJlMDY4OTU3NzE1N2I4OWI0ZTI3NyIsInRhc2tJbmRleCI6MH0sIndvcmtzcGFjZXMiOnsidGFza1Jvb3QiOiIvd29ya3NwYWNlL3Rhc2sifX0=",
1793
+ encoding: "base64",
1794
+ sha256: "sha256:a8fe2f7c6d5bc4d0832a2aa3a0ec6b0c311a1f1bafe9a0fc406991802a87685e"
1795
+ },
1796
+ digest: "sha256:a8fe2f7c6d5bc4d0832a2aa3a0ec6b0c311a1f1bafe9a0fc406991802a87685e",
1797
+ kind: "agent-candidate-execution-plan",
1798
+ material: {
1799
+ codeKind: "disabled",
1800
+ container: {
1801
+ image: "ghcr.io/example/task",
1802
+ indexDigest: "sha256:aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
1803
+ manifestDigest: "sha256:bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
1804
+ platform: {
1805
+ architecture: "amd64",
1806
+ os: "linux"
1807
+ },
1808
+ source: "evaluator-task-container"
1809
+ },
1810
+ executionId: "experiment-baseline-0-2",
1811
+ harness: "codex",
1812
+ harnessVersion: "1.2.3",
1813
+ instructionDelivery: {
1814
+ kind: "stdin-utf8"
1815
+ },
1816
+ kind: "agent-candidate-execution-plan-material",
1817
+ launch: {
1818
+ args: [],
1819
+ cwd: {
1820
+ path: ".",
1821
+ workspace: "task"
1822
+ },
1823
+ env: {
1824
+ PATH: {
1825
+ kind: "public",
1826
+ value: "/usr/local/bin:/usr/bin:/bin"
1827
+ }
1828
+ },
1829
+ executable: "codex"
1830
+ },
1831
+ limits: {
1832
+ maxCostUsd: 5,
1833
+ maxInputTokens: 1e5,
1834
+ maxModelCalls: 50,
1835
+ maxOutputTokens: 5e4,
1836
+ maxSteps: 100,
1837
+ timeoutMs: 6e4
1838
+ },
1839
+ memory: {
1840
+ mode: "disabled"
1841
+ },
1842
+ model: {
1843
+ access: {
1844
+ grantDigest: "sha256:cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc",
1845
+ kind: "evaluator-mediated",
1846
+ network: {
1847
+ domains: ["router.tangle.tools"],
1848
+ mode: "gateway-only"
1849
+ }
1850
+ },
1851
+ policy: "single",
1852
+ resolved: {
1853
+ model: "model-snapshot",
1854
+ provider: "provider",
1855
+ reasoningEffort: "high",
1856
+ requested: "provider/model",
1857
+ snapshot: "model-snapshot-2026-07-01"
1858
+ },
1859
+ routes: [
1860
+ {
1861
+ kind: "primary",
1862
+ requested: "provider/model"
1863
+ }
1864
+ ]
1865
+ },
1866
+ network: {
1867
+ mode: "disabled"
1868
+ },
1869
+ profile: {
1870
+ mountPaths: ["AGENTS.md"],
1871
+ planDigest: "sha256:af65bd67cf53561cd7bfac4e8c913ef18e36c0c2fe20a0bca1df55e59b21fab5",
1872
+ targetWorkspace: "task"
1873
+ },
1874
+ runCell: {
1875
+ arm: "baseline",
1876
+ attempt: 1,
1877
+ bundleDigest: "sha256:5c21ee53e513fc604cb09754e21c392b24a424da0ef37dbf8f1ee4a8a0b08f09",
1878
+ digest: "sha256:9c427cc67507cda8dcdac327f390adfa7e045d069e677d62921678b46a96d66c",
1879
+ experimentDigest: "sha256:c47a0dad80e6cc4856cdb8a02a279da621ab7d9409e3127650b67f1fc1c14b2d",
1880
+ kind: "agent-candidate-run-cell",
1881
+ repetition: 2,
1882
+ seed: 103,
1883
+ suiteDigest: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5",
1884
+ taskDigest: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277",
1885
+ taskIndex: 0
1886
+ },
1887
+ workspaces: {
1888
+ taskRoot: "/workspace/task"
1889
+ }
1890
+ }
1891
+ },
1892
+ harness: "codex",
1893
+ harnessVersion: "1.2.3",
1894
+ kind: "agent-candidate-materialization",
1895
+ profileActivation: {
1896
+ digest: "sha256:273b700e4080e7e6cc4bd58a755466d18ce323eaaa195bb4202fea456422bb66",
1897
+ files: [
1898
+ {
1899
+ content: "Inspect the repository, implement the fix, and run tests.\n",
1900
+ mode: 420,
1901
+ path: "AGENTS.md"
1902
+ }
1903
+ ],
1904
+ kind: "agent-candidate-profile-activation",
1905
+ profilePlan: {
1906
+ artifact: {
1907
+ byteLength: 287,
1908
+ content: "eyJlbnYiOnt9LCJmaWxlcyI6W3siY29udGVudFNoYTI1NiI6InNoYTI1NjozOWFkYmI5YjdmNTcxZjBmZDU4NmU3Nzk1MDIzZDI1OTA5YWRiZTk1ODZiYjIxZWUxMzRkM2Q3MjQ4YmYxNTk5IiwibW9kZSI6NDIwLCJyZWxQYXRoIjoiQUdFTlRTLm1kIn1dLCJmbGFncyI6W10sImhhcm5lc3MiOiJjb2RleCIsInNvdXJjZVByb2ZpbGVEaWdlc3QiOiJzaGEyNTY6ZTUzZTFjOGY5NjMyYmNiNWRiOTAzM2UzMmYzM2E5MGI5YmQ5OWFiZWNlZjBiYjUyZmY5MDY5ZjA3NTA5ZTJmOSIsInVuc3VwcG9ydGVkIjpbXX0=",
1909
+ encoding: "base64",
1910
+ sha256: "sha256:af65bd67cf53561cd7bfac4e8c913ef18e36c0c2fe20a0bca1df55e59b21fab5"
1911
+ },
1912
+ digest: "sha256:af65bd67cf53561cd7bfac4e8c913ef18e36c0c2fe20a0bca1df55e59b21fab5",
1913
+ kind: "agent-profile-workspace-plan",
1914
+ material: {
1915
+ env: {},
1916
+ files: [
1917
+ {
1918
+ contentSha256: "sha256:39adbb9b7f571f0fd586e7795023d25909adbe9586bb21ee134d3d7248bf1599",
1919
+ mode: 420,
1920
+ relPath: "AGENTS.md"
1921
+ }
1922
+ ],
1923
+ flags: [],
1924
+ harness: "codex",
1925
+ sourceProfileDigest: "sha256:e53e1c8f9632bcb5db9033e32f33a90b9bd99abecef0bb52ff9069f07509e2f9",
1926
+ unsupported: []
1927
+ }
1928
+ }
1929
+ },
1930
+ resolvedModel: {
1931
+ model: "model-snapshot",
1932
+ provider: "provider",
1933
+ reasoningEffort: "high",
1934
+ requested: "provider/model",
1935
+ snapshot: "model-snapshot-2026-07-01"
1936
+ }
1937
+ },
1938
+ receipt: {
1939
+ benchmarkResult: {
1940
+ artifact: {
1941
+ byteLength: 1122,
1942
+ locator: {
1943
+ bucket: "candidate-test-artifacts",
1944
+ key: "benchmark-result/df1a5884a7ed4e54dae3d023b47c21596d27264b5ca59693b51eb221dd533284",
1945
+ kind: "s3"
1946
+ },
1947
+ sha256: "sha256:df1a5884a7ed4e54dae3d023b47c21596d27264b5ca59693b51eb221dd533284"
1948
+ },
1949
+ digest: "sha256:df1a5884a7ed4e54dae3d023b47c21596d27264b5ca59693b51eb221dd533284",
1950
+ kind: "agent-candidate-benchmark-result",
1951
+ material: {
1952
+ dimensions: [
1953
+ {
1954
+ name: "quality",
1955
+ score: 0
1956
+ }
1957
+ ],
1958
+ evidence: {
1959
+ byteLength: 11,
1960
+ locator: {
1961
+ bucket: "candidate-test-artifacts",
1962
+ key: "grader-evidence/628cb4ead38e34e34a34f78fa4481f2a0ccf8e9139acfae69d79dc997e90e36b",
1963
+ kind: "s3"
1964
+ },
1965
+ sha256: "sha256:628cb4ead38e34e34a34f78fa4481f2a0ccf8e9139acfae69d79dc997e90e36b"
1966
+ },
1967
+ executionPlanDigest: "sha256:a8fe2f7c6d5bc4d0832a2aa3a0ec6b0c311a1f1bafe9a0fc406991802a87685e",
1968
+ grader: {
1969
+ artifact: {
1970
+ byteLength: 25,
1971
+ locator: {
1972
+ bucket: "candidate-test-artifacts",
1973
+ key: "grader/9a354495185587b25e8c98104060068871a44c719e4f57d345a905fc2d72a6ad",
1974
+ kind: "s3"
1975
+ },
1976
+ sha256: "sha256:9a354495185587b25e8c98104060068871a44c719e4f57d345a905fc2d72a6ad"
1977
+ },
1978
+ format: "tangle-grader",
1979
+ name: "fixture-executable-grader",
1980
+ version: "1.0.0"
1981
+ },
1982
+ grading: {
1983
+ timing: {
1984
+ durationMs: 0,
1985
+ endedAtMs: 17836416e5,
1986
+ startedAtMs: 17836416e5
1987
+ },
1988
+ usage: {
1989
+ cachedInputTokens: 0,
1990
+ costUsdNanos: 0,
1991
+ inputTokens: 0,
1992
+ modelCalls: 0,
1993
+ outputTokens: 0,
1994
+ reasoningTokens: 0
1995
+ }
1996
+ },
1997
+ kind: "agent-candidate-benchmark-result-material",
1998
+ passed: false,
1999
+ score: 0,
2000
+ taskOutcomeDigest: "sha256:664358fd0e890ca17d0c31aa0157e434c3bc7f6ecaa77c581d5ef600b6a836ce"
2001
+ }
2002
+ },
2003
+ bundleDigest: "sha256:5c21ee53e513fc604cb09754e21c392b24a424da0ef37dbf8f1ee4a8a0b08f09",
2004
+ digest: "sha256:7dfea3ef8bf6e5c0432180fefe63cad9f62833ecb8bea1ce0de7d41ca3ef3bc5",
2005
+ digestAlgorithm: "rfc8785-sha256",
2006
+ executionPlanDigest: "sha256:a8fe2f7c6d5bc4d0832a2aa3a0ec6b0c311a1f1bafe9a0fc406991802a87685e",
2007
+ executorCapture: {
2008
+ byteLength: 477,
2009
+ locator: {
2010
+ bucket: "candidate-test-artifacts",
2011
+ key: "executor-capture/3fdb1d6105bd10c22daa6c0267015a8f6a88b13d1703e8bf2c901c0887a4b242",
2012
+ kind: "s3"
2013
+ },
2014
+ sha256: "sha256:3fdb1d6105bd10c22daa6c0267015a8f6a88b13d1703e8bf2c901c0887a4b242"
2015
+ },
2016
+ kind: "agent-candidate-run",
2017
+ materializationReceiptDigest: "sha256:0e3da90681970b46a99808f5d5a623329bae9529bf0f46ccebe7cac7a202ec8a",
2018
+ memory: {
2019
+ mode: "disabled"
2020
+ },
2021
+ modelSettlement: {
2022
+ artifact: {
2023
+ byteLength: 615,
2024
+ locator: {
2025
+ bucket: "candidate-test-artifacts",
2026
+ key: "model-settlement/39c54b821165bbff4a33cf310c51e20f6d1ff0ee10ff7630618010e91d5a1615",
2027
+ kind: "s3"
2028
+ },
2029
+ sha256: "sha256:39c54b821165bbff4a33cf310c51e20f6d1ff0ee10ff7630618010e91d5a1615"
2030
+ },
2031
+ digest: "sha256:39c54b821165bbff4a33cf310c51e20f6d1ff0ee10ff7630618010e91d5a1615",
2032
+ kind: "agent-candidate-model-settlement",
2033
+ material: {
2034
+ calls: [],
2035
+ closed: true,
2036
+ executionPlanDigest: "sha256:a8fe2f7c6d5bc4d0832a2aa3a0ec6b0c311a1f1bafe9a0fc406991802a87685e",
2037
+ grantDigest: "sha256:cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc",
2038
+ kind: "agent-candidate-model-settlement-material",
2039
+ preparationId: "candidate-preparation.nBqsDbKYlsiY4Tto5XaxkR3RCTop9PA9_kX-kOSzR8U",
2040
+ resolved: {
2041
+ model: "model-snapshot",
2042
+ provider: "provider",
2043
+ reasoningEffort: "high",
2044
+ requested: "provider/model",
2045
+ snapshot: "model-snapshot-2026-07-01"
2046
+ },
2047
+ usage: {
2048
+ cachedInputTokens: 0,
2049
+ costUsdNanos: 0,
2050
+ inputTokens: 0,
2051
+ modelCalls: 0,
2052
+ outputTokens: 0,
2053
+ reasoningTokens: 0
2054
+ }
2055
+ }
2056
+ },
2057
+ runCellDigest: "sha256:9c427cc67507cda8dcdac327f390adfa7e045d069e677d62921678b46a96d66c",
2058
+ taskOutcome: {
2059
+ artifact: {
2060
+ byteLength: 478,
2061
+ locator: {
2062
+ bucket: "candidate-test-artifacts",
2063
+ key: "task-outcome/664358fd0e890ca17d0c31aa0157e434c3bc7f6ecaa77c581d5ef600b6a836ce",
2064
+ kind: "s3"
2065
+ },
2066
+ sha256: "sha256:664358fd0e890ca17d0c31aa0157e434c3bc7f6ecaa77c581d5ef600b6a836ce"
2067
+ },
2068
+ digest: "sha256:664358fd0e890ca17d0c31aa0157e434c3bc7f6ecaa77c581d5ef600b6a836ce",
2069
+ kind: "agent-candidate-task-outcome",
2070
+ material: {
2071
+ executionPlanDigest: "sha256:a8fe2f7c6d5bc4d0832a2aa3a0ec6b0c311a1f1bafe9a0fc406991802a87685e",
2072
+ kind: "agent-candidate-task-outcome-material",
2073
+ outcome: {
2074
+ artifact: {
2075
+ byteLength: 4,
2076
+ locator: {
2077
+ bucket: "candidate-test-artifacts",
2078
+ key: "task-output/481ba4019c9d2710c3386537382592c093ef02bf5b056c30237b3d92b0e19a1d",
2079
+ kind: "s3"
2080
+ },
2081
+ sha256: "sha256:481ba4019c9d2710c3386537382592c093ef02bf5b056c30237b3d92b0e19a1d"
2082
+ },
2083
+ kind: "output",
2084
+ spec: {
2085
+ maxBytes: 1024,
2086
+ mediaType: "text/plain"
2087
+ }
2088
+ }
2089
+ }
2090
+ },
2091
+ termination: {
2092
+ exitCode: 0,
2093
+ kind: "exit"
2094
+ },
2095
+ timing: {
2096
+ durationMs: 50,
2097
+ endedAtMs: 1250,
2098
+ startedAtMs: 1200
2099
+ },
2100
+ trace: {
2101
+ artifact: {
2102
+ byteLength: 748,
2103
+ locator: {
2104
+ bucket: "candidate-test-artifacts",
2105
+ key: "trace/f1867bc538dae9c3a35e73fd7c21c06150bbcba815a56835f70b26d615a6e66d",
2106
+ kind: "s3"
2107
+ },
2108
+ sha256: "sha256:f1867bc538dae9c3a35e73fd7c21c06150bbcba815a56835f70b26d615a6e66d"
2109
+ },
2110
+ eventCount: 1,
2111
+ modelCallCount: 0
2112
+ }
2113
+ }
2114
+ },
2115
+ candidate: {
2116
+ digest: "sha256:325bbde1ccafdb3786a26b48ad407fea1912f0d7b340c4a654a1ebdc622d960c",
2117
+ kind: "agent-candidate-execution-evidence",
2118
+ materializationReceipt: {
2119
+ benchmark: {
2120
+ suite: {
2121
+ digest: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5",
2122
+ material: {
2123
+ byteLength: 198,
2124
+ content: "eyJkaWdlc3RBbGdvcml0aG0iOiJyZmM4Nzg1LXNoYTI1NiIsImtpbmQiOiJhZ2VudC1jYW5kaWRhdGUtYmVuY2htYXJrLXN1aXRlIiwicmVwcyI6Mywic2VlZHMiOlsxMDEsMTAyLDEwM10sInRhc2tEaWdlc3RzIjpbInNoYTI1Njo2YWE3MTgzMzFiNmRhMzEwZGNhYjZkZjJlMzM5N2NiNzcwODQwZTU1NWE4MmUwNjg5NTc3MTU3Yjg5YjRlMjc3Il19",
2125
+ encoding: "base64",
2126
+ sha256: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5"
2127
+ }
2128
+ },
2129
+ task: {
2130
+ digest: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277",
2131
+ material: {
2132
+ byteLength: 2096,
2133
+ content: "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",
2134
+ encoding: "base64",
2135
+ sha256: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277"
2136
+ }
2137
+ }
2138
+ },
2139
+ bundleDigest: "sha256:60fcbb1c728194bd51d7d19cb732d1c3f1881dce7e0a6266b41c8b98cfd65693",
2140
+ codeKind: "disabled",
2141
+ container: {
2142
+ image: "ghcr.io/example/task",
2143
+ indexDigest: "sha256:aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
2144
+ manifestDigest: "sha256:bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
2145
+ platform: {
2146
+ architecture: "amd64",
2147
+ os: "linux"
2148
+ },
2149
+ source: "evaluator-task-container"
2150
+ },
2151
+ digest: "sha256:b3718d6b33e7316d37af34e81e93a6521b0e9914bdec820d56e14fd7008f3e05",
2152
+ digestAlgorithm: "rfc8785-sha256",
2153
+ executionPlan: {
2154
+ artifact: {
2155
+ byteLength: 2032,
2156
+ content: "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",
2157
+ encoding: "base64",
2158
+ sha256: "sha256:c642749f2907277b811c99017d96103dd1ab6d2564fe1c8caba27bbe8d6c151f"
2159
+ },
2160
+ digest: "sha256:c642749f2907277b811c99017d96103dd1ab6d2564fe1c8caba27bbe8d6c151f",
2161
+ kind: "agent-candidate-execution-plan",
2162
+ material: {
2163
+ codeKind: "disabled",
2164
+ container: {
2165
+ image: "ghcr.io/example/task",
2166
+ indexDigest: "sha256:aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
2167
+ manifestDigest: "sha256:bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
2168
+ platform: {
2169
+ architecture: "amd64",
2170
+ os: "linux"
2171
+ },
2172
+ source: "evaluator-task-container"
2173
+ },
2174
+ executionId: "experiment-candidate-0-2",
2175
+ harness: "codex",
2176
+ harnessVersion: "1.2.3",
2177
+ instructionDelivery: {
2178
+ kind: "stdin-utf8"
2179
+ },
2180
+ kind: "agent-candidate-execution-plan-material",
2181
+ launch: {
2182
+ args: [],
2183
+ cwd: {
2184
+ path: ".",
2185
+ workspace: "task"
2186
+ },
2187
+ env: {
2188
+ PATH: {
2189
+ kind: "public",
2190
+ value: "/usr/local/bin:/usr/bin:/bin"
2191
+ }
2192
+ },
2193
+ executable: "codex"
2194
+ },
2195
+ limits: {
2196
+ maxCostUsd: 5,
2197
+ maxInputTokens: 1e5,
2198
+ maxModelCalls: 50,
2199
+ maxOutputTokens: 5e4,
2200
+ maxSteps: 100,
2201
+ timeoutMs: 6e4
2202
+ },
2203
+ memory: {
2204
+ mode: "disabled"
2205
+ },
2206
+ model: {
2207
+ access: {
2208
+ grantDigest: "sha256:cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc",
2209
+ kind: "evaluator-mediated",
2210
+ network: {
2211
+ domains: ["router.tangle.tools"],
2212
+ mode: "gateway-only"
2213
+ }
2214
+ },
2215
+ policy: "single",
2216
+ resolved: {
2217
+ model: "model-snapshot",
2218
+ provider: "provider",
2219
+ reasoningEffort: "high",
2220
+ requested: "provider/model",
2221
+ snapshot: "model-snapshot-2026-07-01"
2222
+ },
2223
+ routes: [
2224
+ {
2225
+ kind: "primary",
2226
+ requested: "provider/model"
2227
+ }
2228
+ ]
2229
+ },
2230
+ network: {
2231
+ mode: "disabled"
2232
+ },
2233
+ profile: {
2234
+ mountPaths: ["AGENTS.md"],
2235
+ planDigest: "sha256:4b4214d1bae911df6dda3a5d8585c65df2f0b2380126eb40c0cbd46a4a2fba5b",
2236
+ targetWorkspace: "task"
2237
+ },
2238
+ runCell: {
2239
+ arm: "candidate",
2240
+ attempt: 1,
2241
+ bundleDigest: "sha256:60fcbb1c728194bd51d7d19cb732d1c3f1881dce7e0a6266b41c8b98cfd65693",
2242
+ digest: "sha256:ea5cab9dff49326dfe856aab19514d2bc5bc0fab2b4b685c4fea356801dbd1c3",
2243
+ experimentDigest: "sha256:c47a0dad80e6cc4856cdb8a02a279da621ab7d9409e3127650b67f1fc1c14b2d",
2244
+ kind: "agent-candidate-run-cell",
2245
+ repetition: 2,
2246
+ seed: 103,
2247
+ suiteDigest: "sha256:963d6ad7abe359c963f31c21036ac6176fe7d5f097b2200ab4430c0d1bf1bcd5",
2248
+ taskDigest: "sha256:6aa718331b6da310dcab6df2e3397cb770840e555a82e0689577157b89b4e277",
2249
+ taskIndex: 0
2250
+ },
2251
+ workspaces: {
2252
+ taskRoot: "/workspace/task"
2253
+ }
2254
+ }
2255
+ },
2256
+ harness: "codex",
2257
+ harnessVersion: "1.2.3",
2258
+ kind: "agent-candidate-materialization",
2259
+ profileActivation: {
2260
+ digest: "sha256:3103f8282be70a843eada63410c41d9eb5ac7b824c758eed6c2f8610de169f45",
2261
+ files: [
2262
+ {
2263
+ content: "Inspect the repository, implement the fix, and run tests.\n",
2264
+ mode: 420,
2265
+ path: "AGENTS.md"
2266
+ }
2267
+ ],
2268
+ kind: "agent-candidate-profile-activation",
2269
+ profilePlan: {
2270
+ artifact: {
2271
+ byteLength: 364,
2272
+ content: "eyJlbnYiOnt9LCJmaWxlcyI6W3siY29udGVudFNoYTI1NiI6InNoYTI1NjozOWFkYmI5YjdmNTcxZjBmZDU4NmU3Nzk1MDIzZDI1OTA5YWRiZTk1ODZiYjIxZWUxMzRkM2Q3MjQ4YmYxNTk5IiwibW9kZSI6NDIwLCJyZWxQYXRoIjoiQUdFTlRTLm1kIn1dLCJmbGFncyI6W10sImhhcm5lc3MiOiJjb2RleCIsInNvdXJjZVByb2ZpbGVEaWdlc3QiOiJzaGEyNTY6OGIxZTJmMGY3YjgzMjk1ODM4MWMyOWJhZWJlNGE5NjFiMTUxYWY5N2EwY2U5ZDYxMWExNWFkMDc0MWFlYWE3NSIsInN5c3RlbVByb21wdCI6eyJraW5kIjoicHVibGljIiwidmFsdWUiOiJSZXR1cm4gdGhlIGV4YWN0IG1lYXN1cmVkIGFuc3dlci4ifSwidW5zdXBwb3J0ZWQiOltdfQ==",
2273
+ encoding: "base64",
2274
+ sha256: "sha256:4b4214d1bae911df6dda3a5d8585c65df2f0b2380126eb40c0cbd46a4a2fba5b"
2275
+ },
2276
+ digest: "sha256:4b4214d1bae911df6dda3a5d8585c65df2f0b2380126eb40c0cbd46a4a2fba5b",
2277
+ kind: "agent-profile-workspace-plan",
2278
+ material: {
2279
+ env: {},
2280
+ files: [
2281
+ {
2282
+ contentSha256: "sha256:39adbb9b7f571f0fd586e7795023d25909adbe9586bb21ee134d3d7248bf1599",
2283
+ mode: 420,
2284
+ relPath: "AGENTS.md"
2285
+ }
2286
+ ],
2287
+ flags: [],
2288
+ harness: "codex",
2289
+ sourceProfileDigest: "sha256:8b1e2f0f7b832958381c29baebe4a961b151af97a0ce9d611a15ad0741aeaa75",
2290
+ systemPrompt: {
2291
+ kind: "public",
2292
+ value: "Return the exact measured answer."
2293
+ },
2294
+ unsupported: []
2295
+ }
2296
+ }
2297
+ },
2298
+ resolvedModel: {
2299
+ model: "model-snapshot",
2300
+ provider: "provider",
2301
+ reasoningEffort: "high",
2302
+ requested: "provider/model",
2303
+ snapshot: "model-snapshot-2026-07-01"
2304
+ }
2305
+ },
2306
+ receipt: {
2307
+ benchmarkResult: {
2308
+ artifact: {
2309
+ byteLength: 1121,
2310
+ locator: {
2311
+ bucket: "candidate-test-artifacts",
2312
+ key: "benchmark-result/95541270205f83cd4d1fe61062e133c4f2afdb1978d1af56de0aef0410999aa4",
2313
+ kind: "s3"
2314
+ },
2315
+ sha256: "sha256:95541270205f83cd4d1fe61062e133c4f2afdb1978d1af56de0aef0410999aa4"
2316
+ },
2317
+ digest: "sha256:95541270205f83cd4d1fe61062e133c4f2afdb1978d1af56de0aef0410999aa4",
2318
+ kind: "agent-candidate-benchmark-result",
2319
+ material: {
2320
+ dimensions: [
2321
+ {
2322
+ name: "quality",
2323
+ score: 1
2324
+ }
2325
+ ],
2326
+ evidence: {
2327
+ byteLength: 11,
2328
+ locator: {
2329
+ bucket: "candidate-test-artifacts",
2330
+ key: "grader-evidence/9b9b3a1471309177261cfe65ea9c298e0dd372e4b5d087f8d35d7b732485373d",
2331
+ kind: "s3"
2332
+ },
2333
+ sha256: "sha256:9b9b3a1471309177261cfe65ea9c298e0dd372e4b5d087f8d35d7b732485373d"
2334
+ },
2335
+ executionPlanDigest: "sha256:c642749f2907277b811c99017d96103dd1ab6d2564fe1c8caba27bbe8d6c151f",
2336
+ grader: {
2337
+ artifact: {
2338
+ byteLength: 25,
2339
+ locator: {
2340
+ bucket: "candidate-test-artifacts",
2341
+ key: "grader/9a354495185587b25e8c98104060068871a44c719e4f57d345a905fc2d72a6ad",
2342
+ kind: "s3"
2343
+ },
2344
+ sha256: "sha256:9a354495185587b25e8c98104060068871a44c719e4f57d345a905fc2d72a6ad"
2345
+ },
2346
+ format: "tangle-grader",
2347
+ name: "fixture-executable-grader",
2348
+ version: "1.0.0"
2349
+ },
2350
+ grading: {
2351
+ timing: {
2352
+ durationMs: 0,
2353
+ endedAtMs: 17836416e5,
2354
+ startedAtMs: 17836416e5
2355
+ },
2356
+ usage: {
2357
+ cachedInputTokens: 0,
2358
+ costUsdNanos: 0,
2359
+ inputTokens: 0,
2360
+ modelCalls: 0,
2361
+ outputTokens: 0,
2362
+ reasoningTokens: 0
2363
+ }
2364
+ },
2365
+ kind: "agent-candidate-benchmark-result-material",
2366
+ passed: true,
2367
+ score: 1,
2368
+ taskOutcomeDigest: "sha256:61d53d0789c46c98b43acdc7177b1b6102f48aa2d6b3b2ff0ec5ee4b287c8d92"
2369
+ }
2370
+ },
2371
+ bundleDigest: "sha256:60fcbb1c728194bd51d7d19cb732d1c3f1881dce7e0a6266b41c8b98cfd65693",
2372
+ digest: "sha256:62fb136a6376e729f10efa210b6dba6607923b2502597e555bd055d72ee307dd",
2373
+ digestAlgorithm: "rfc8785-sha256",
2374
+ executionPlanDigest: "sha256:c642749f2907277b811c99017d96103dd1ab6d2564fe1c8caba27bbe8d6c151f",
2375
+ executorCapture: {
2376
+ byteLength: 477,
2377
+ locator: {
2378
+ bucket: "candidate-test-artifacts",
2379
+ key: "executor-capture/c8e3a1667966134d3f7b24b4df19a5f5286814522f7884f0a76f7cfd3d4832a2",
2380
+ kind: "s3"
2381
+ },
2382
+ sha256: "sha256:c8e3a1667966134d3f7b24b4df19a5f5286814522f7884f0a76f7cfd3d4832a2"
2383
+ },
2384
+ kind: "agent-candidate-run",
2385
+ materializationReceiptDigest: "sha256:b3718d6b33e7316d37af34e81e93a6521b0e9914bdec820d56e14fd7008f3e05",
2386
+ memory: {
2387
+ mode: "disabled"
2388
+ },
2389
+ modelSettlement: {
2390
+ artifact: {
2391
+ byteLength: 615,
2392
+ locator: {
2393
+ bucket: "candidate-test-artifacts",
2394
+ key: "model-settlement/397df4f1827292bad160efdfeb9f92748df604396b285b20f527299b073a8988",
2395
+ kind: "s3"
2396
+ },
2397
+ sha256: "sha256:397df4f1827292bad160efdfeb9f92748df604396b285b20f527299b073a8988"
2398
+ },
2399
+ digest: "sha256:397df4f1827292bad160efdfeb9f92748df604396b285b20f527299b073a8988",
2400
+ kind: "agent-candidate-model-settlement",
2401
+ material: {
2402
+ calls: [],
2403
+ closed: true,
2404
+ executionPlanDigest: "sha256:c642749f2907277b811c99017d96103dd1ab6d2564fe1c8caba27bbe8d6c151f",
2405
+ grantDigest: "sha256:cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc",
2406
+ kind: "agent-candidate-model-settlement-material",
2407
+ preparationId: "candidate-preparation.nBqsDbKYlsiY4Tto5XaxkR3RCTop9PA9_kX-kOSzR8U",
2408
+ resolved: {
2409
+ model: "model-snapshot",
2410
+ provider: "provider",
2411
+ reasoningEffort: "high",
2412
+ requested: "provider/model",
2413
+ snapshot: "model-snapshot-2026-07-01"
2414
+ },
2415
+ usage: {
2416
+ cachedInputTokens: 0,
2417
+ costUsdNanos: 0,
2418
+ inputTokens: 0,
2419
+ modelCalls: 0,
2420
+ outputTokens: 0,
2421
+ reasoningTokens: 0
2422
+ }
2423
+ }
2424
+ },
2425
+ runCellDigest: "sha256:ea5cab9dff49326dfe856aab19514d2bc5bc0fab2b4b685c4fea356801dbd1c3",
2426
+ taskOutcome: {
2427
+ artifact: {
2428
+ byteLength: 478,
2429
+ locator: {
2430
+ bucket: "candidate-test-artifacts",
2431
+ key: "task-outcome/61d53d0789c46c98b43acdc7177b1b6102f48aa2d6b3b2ff0ec5ee4b287c8d92",
2432
+ kind: "s3"
2433
+ },
2434
+ sha256: "sha256:61d53d0789c46c98b43acdc7177b1b6102f48aa2d6b3b2ff0ec5ee4b287c8d92"
2435
+ },
2436
+ digest: "sha256:61d53d0789c46c98b43acdc7177b1b6102f48aa2d6b3b2ff0ec5ee4b287c8d92",
2437
+ kind: "agent-candidate-task-outcome",
2438
+ material: {
2439
+ executionPlanDigest: "sha256:c642749f2907277b811c99017d96103dd1ab6d2564fe1c8caba27bbe8d6c151f",
2440
+ kind: "agent-candidate-task-outcome-material",
2441
+ outcome: {
2442
+ artifact: {
2443
+ byteLength: 6,
2444
+ locator: {
2445
+ bucket: "candidate-test-artifacts",
2446
+ key: "task-output/eda15ce4834ac303532bc6973df5ff8a5deb75d7dfceb9fc2cdf1bb450a01cf5",
2447
+ kind: "s3"
2448
+ },
2449
+ sha256: "sha256:eda15ce4834ac303532bc6973df5ff8a5deb75d7dfceb9fc2cdf1bb450a01cf5"
2450
+ },
2451
+ kind: "output",
2452
+ spec: {
2453
+ maxBytes: 1024,
2454
+ mediaType: "text/plain"
2455
+ }
2456
+ }
2457
+ }
2458
+ },
2459
+ termination: {
2460
+ exitCode: 0,
2461
+ kind: "exit"
2462
+ },
2463
+ timing: {
2464
+ durationMs: 50,
2465
+ endedAtMs: 1250,
2466
+ startedAtMs: 1200
2467
+ },
2468
+ trace: {
2469
+ artifact: {
2470
+ byteLength: 750,
2471
+ locator: {
2472
+ bucket: "candidate-test-artifacts",
2473
+ key: "trace/59f819d9564d5e61e59e52545f9b269083f0a0b57a92a5e8f93e566da560ef17",
2474
+ kind: "s3"
2475
+ },
2476
+ sha256: "sha256:59f819d9564d5e61e59e52545f9b269083f0a0b57a92a5e8f93e566da560ef17"
2477
+ },
2478
+ eventCount: 1,
2479
+ modelCallCount: 0
2480
+ }
2481
+ }
2482
+ }
2483
+ }
2484
+ ],
2485
+ metadata: {
2486
+ fixture: "agent-improvement-proposal",
2487
+ runtimeVersion: "0.102.0"
2488
+ },
2489
+ objectives: [
2490
+ {
2491
+ availability: "measured",
2492
+ baseline: 0,
2493
+ candidate: 1,
2494
+ confidenceInterval: {
2495
+ level: 0.95,
2496
+ lower: 1,
2497
+ method: "paired-bootstrap",
2498
+ resamples: 500,
2499
+ statistic: "mean",
2500
+ upper: 1
2501
+ },
2502
+ delta: 1,
2503
+ direction: "higher-is-better",
2504
+ kind: "objective",
2505
+ n: 3,
2506
+ name: "benchmark-score",
2507
+ unit: "score"
2508
+ },
2509
+ {
2510
+ availability: "measured",
2511
+ baseline: 0,
2512
+ candidate: 1,
2513
+ confidenceInterval: {
2514
+ level: 0.95,
2515
+ lower: 1,
2516
+ method: "paired-bootstrap",
2517
+ resamples: 500,
2518
+ statistic: "mean",
2519
+ upper: 1
2520
+ },
2521
+ delta: 1,
2522
+ direction: "higher-is-better",
2523
+ kind: "dimension",
2524
+ n: 3,
2525
+ name: "quality",
2526
+ objective: "benchmark-score",
2527
+ unit: "score"
2528
+ },
2529
+ {
2530
+ availability: "measured",
2531
+ baseline: 0,
2532
+ candidate: 0,
2533
+ confidenceInterval: {
2534
+ level: 0.95,
2535
+ lower: 0,
2536
+ method: "paired-bootstrap",
2537
+ resamples: 500,
2538
+ statistic: "mean",
2539
+ upper: 0
2540
+ },
2541
+ delta: 0,
2542
+ direction: "lower-is-better",
2543
+ kind: "cost",
2544
+ n: 3,
2545
+ name: "cost",
2546
+ unit: "usd"
2547
+ },
2548
+ {
2549
+ availability: "measured",
2550
+ baseline: 50,
2551
+ candidate: 50,
2552
+ confidenceInterval: {
2553
+ level: 0.95,
2554
+ lower: 0,
2555
+ method: "paired-bootstrap",
2556
+ resamples: 500,
2557
+ statistic: "mean",
2558
+ upper: 0
2559
+ },
2560
+ delta: 0,
2561
+ direction: "lower-is-better",
2562
+ kind: "latency",
2563
+ n: 3,
2564
+ name: "latency",
2565
+ unit: "milliseconds"
2566
+ }
2567
+ ],
2568
+ overall: {
2569
+ baseline: 0,
2570
+ candidate: 1,
2571
+ confidenceInterval: {
2572
+ level: 0.95,
2573
+ lower: 1,
2574
+ method: "paired-bootstrap",
2575
+ resamples: 500,
2576
+ statistic: "mean",
2577
+ upper: 1
2578
+ },
2579
+ delta: 1,
2580
+ direction: "higher-is-better",
2581
+ n: 3,
2582
+ name: "composite",
2583
+ unit: "score"
2584
+ },
2585
+ power: {
2586
+ confidenceLevel: 0.95,
2587
+ minimumDetectableDelta: 0,
2588
+ n: 3,
2589
+ reason: "Minimum detectable lift at n=3: 0.000 (baseline sd 0.000). Effects smaller than this cannot clear the gate; budget the search for effects you believe exceed it. Holdout and gate share one scoring channel: raising n/reps reduces only idiosyncratic noise \u2014 systematic judge bias remains and this MDE is a lower bound. Full debiasing needs an independent second scoring channel (different judge/benchmark family).",
2590
+ scaleAssumed: true,
2591
+ sharedScorerChannel: true,
2592
+ sufficient: true
2593
+ },
2594
+ provenance: {
2595
+ baselineContentHash: "sha256:5c21ee53e513fc604cb09754e21c392b24a424da0ef37dbf8f1ee4a8a0b08f09",
2596
+ candidateContentHash: "sha256:60fcbb1c728194bd51d7d19cb732d1c3f1881dce7e0a6266b41c8b98cfd65693",
2597
+ kind: "agent-eval-loop",
2598
+ recordDigest: "sha256:6ca133e58b214ee1b6712edbb59abba570d24c675279e327f35831f8445eafbc",
2599
+ runId: "agent-runtime-0.102.0-proposal-fixture",
2600
+ schema: "agent-candidate-experiment"
2601
+ }
2602
+ },
2603
+ findings: [
2604
+ {
2605
+ analyst_id: "runtime-testing-fixture",
2606
+ area: "prompt",
2607
+ claim: "The baseline omits the exact measured answer.",
2608
+ confidence: 0.9,
2609
+ evidence_refs: [
2610
+ {
2611
+ id: "runtime-testing-fixture-span",
2612
+ kind: "span"
2613
+ }
2614
+ ],
2615
+ finding_id: "runtime-testing-fixture-finding",
2616
+ produced_at: "2026-07-10T00:30:00.000Z",
2617
+ recommended_action: "Return the exact measured answer.",
2618
+ schema_version: "1.0.0",
2619
+ severity: "high",
2620
+ subject: "agent-profile:prompt.systemPrompt"
2621
+ }
2622
+ ],
2623
+ kind: "agent-improvement-proposal",
2624
+ proposedAt: "2026-07-10T01:00:00.000Z",
2625
+ runId: "agent-runtime-0.102.0-proposal-fixture"
2626
+ };
2627
+
2628
+ // src/testing/index.ts
2629
+ var serializedAgentImprovementProposalFixture = JSON.stringify(
2630
+ agent_improvement_proposal_default
2631
+ );
2632
+ function loadAgentImprovementProposalFixture() {
2633
+ return verifyAgentImprovementProposal(JSON.parse(serializedAgentImprovementProposalFixture));
2634
+ }
2635
+ export {
2636
+ loadAgentImprovementProposalFixture
2637
+ };
2638
+ //# sourceMappingURL=testing.js.map