openpond-sdk 0.12.26 → 0.12.27

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 (200) hide show
  1. package/PRODUCT_INTEGRATION.md +13 -6
  2. package/README.md +2 -10
  3. package/SERVERLESS_SERVING.md +52 -35
  4. package/TASK_INGESTION.md +0 -4
  5. package/dist/advanced-refiner-evaluations.js +7 -7
  6. package/dist/ai-sdk.js +20 -83
  7. package/dist/ai-sdk.js.map +1 -1
  8. package/dist/candidate-evaluations.js +6 -6
  9. package/dist/{chunk-UZAVG2WK.js → chunk-3HOGOWKU.js} +8 -6
  10. package/dist/chunk-3HOGOWKU.js.map +1 -0
  11. package/dist/{chunk-YV6PPDXZ.js → chunk-3Z5IK2UK.js} +27 -2
  12. package/dist/chunk-3Z5IK2UK.js.map +1 -0
  13. package/dist/{chunk-4REK5TTE.js → chunk-4D4QSQC3.js} +2 -2
  14. package/dist/{chunk-AMZ2SWHR.js → chunk-534W2O2S.js} +2 -2
  15. package/dist/{chunk-6LVLWRBJ.js → chunk-54BJJL7G.js} +2 -2
  16. package/dist/{chunk-XWUTYFOU.js → chunk-5KJY44FB.js} +3 -3
  17. package/dist/{chunk-QKLZZOZD.js → chunk-7MLFWVDI.js} +29 -10
  18. package/dist/chunk-7MLFWVDI.js.map +1 -0
  19. package/dist/{chunk-4A4FNG4R.js → chunk-7ZV626YS.js} +2 -2
  20. package/dist/{chunk-57ROW3XF.js → chunk-ABOZV47D.js} +2 -2
  21. package/dist/{chunk-3BILL33A.js → chunk-BIQ5TG3T.js} +4 -4
  22. package/dist/{chunk-JINWKUKR.js → chunk-BT6CC6QB.js} +3 -3
  23. package/dist/{chunk-4PEK3YOW.js → chunk-CNFN6GHY.js} +2 -2
  24. package/dist/{chunk-X5AT253K.js → chunk-CVYGJUWM.js} +2 -2
  25. package/dist/{chunk-D26RJQU7.js → chunk-CX7ETU2F.js} +2 -2
  26. package/dist/{chunk-4CDLPMJK.js → chunk-D7LV3USE.js} +2 -2
  27. package/dist/{chunk-WZKKQNYY.js → chunk-GA7OSSEA.js} +2 -2
  28. package/dist/{chunk-EL6KZO43.js → chunk-GFWXDWEE.js} +3 -3
  29. package/dist/{chunk-XYK3MAIC.js → chunk-JYWIKBGD.js} +122 -112
  30. package/dist/chunk-JYWIKBGD.js.map +1 -0
  31. package/dist/{chunk-AQV7RQZH.js → chunk-K4WU4JOP.js} +2 -2
  32. package/dist/{chunk-CAOMKM55.js → chunk-L7MVGGBH.js} +2 -2
  33. package/dist/{chunk-P56JWQYV.js → chunk-LRFVISLA.js} +3 -3
  34. package/dist/{chunk-2U6BQ5C4.js → chunk-MJHQMQEW.js} +3 -3
  35. package/dist/chunk-ON3KSXWR.js +639 -0
  36. package/dist/chunk-ON3KSXWR.js.map +1 -0
  37. package/dist/{chunk-PMVSO23F.js → chunk-RMXZISHK.js} +10 -1
  38. package/dist/chunk-RMXZISHK.js.map +1 -0
  39. package/dist/{chunk-JWEIAEQ6.js → chunk-TMNEDWA7.js} +10 -3
  40. package/dist/chunk-TMNEDWA7.js.map +1 -0
  41. package/dist/{chunk-C7I4ZG6Q.js → chunk-TX552FQG.js} +3 -3
  42. package/dist/{chunk-4CRLQMJ2.js → chunk-VIDRZCPQ.js} +6 -6
  43. package/dist/{chunk-MN4FE52A.js → chunk-VRUAIMM3.js} +7 -7
  44. package/dist/chunk-VRUAIMM3.js.map +1 -0
  45. package/dist/{chunk-H5KHLBC5.js → chunk-WJZHSIGD.js} +6 -8
  46. package/dist/{chunk-H5KHLBC5.js.map → chunk-WJZHSIGD.js.map} +1 -1
  47. package/dist/{chunk-HCUZYGUK.js → chunk-WPTFJZ4L.js} +3 -3
  48. package/dist/connected-evidence.js +7 -7
  49. package/dist/dataset-workspaces.js +10 -10
  50. package/dist/errors.js +1 -1
  51. package/dist/experiment-evaluation-schedules.js +16 -16
  52. package/dist/experiment-improvements.js +7 -7
  53. package/dist/experiments.js +15 -15
  54. package/dist/index.js +23 -23
  55. package/dist/learning.js +15 -4
  56. package/dist/learning.js.map +1 -1
  57. package/dist/model-projects.js +1 -1
  58. package/dist/model-starter-attempts.js +2 -2
  59. package/dist/model-starter-catalog.js +4 -4
  60. package/dist/model-starters.js +4 -4
  61. package/dist/model-taskset-runs.js +3 -3
  62. package/dist/post-training.js +6 -6
  63. package/dist/taskset-drafts.js +6 -6
  64. package/dist/taskset-packages.js +12 -10
  65. package/dist/training-bundle.js +46 -20
  66. package/dist/training-bundle.js.map +1 -1
  67. package/dist/training-hardware.js +2 -2
  68. package/dist/training.js +5620 -4342
  69. package/dist/training.js.map +1 -1
  70. package/dist/types/packages/sdk/src/ai-sdk.d.ts.map +1 -1
  71. package/dist/types/packages/sdk/src/alphaedit-model-serving.d.ts +8 -4
  72. package/dist/types/packages/sdk/src/alphaedit-model-serving.d.ts.map +1 -1
  73. package/dist/types/packages/sdk/src/candidate-evaluation-client.d.ts +162 -162
  74. package/dist/types/packages/sdk/src/decision-model-serving.d.ts +4 -2
  75. package/dist/types/packages/sdk/src/decision-model-serving.d.ts.map +1 -1
  76. package/dist/types/packages/sdk/src/engram-model-serving.d.ts +3 -1
  77. package/dist/types/packages/sdk/src/engram-model-serving.d.ts.map +1 -1
  78. package/dist/types/packages/sdk/src/experiment-client.d.ts +699 -691
  79. package/dist/types/packages/sdk/src/experiment-client.d.ts.map +1 -1
  80. package/dist/types/packages/sdk/src/experiment-draft-storage.d.ts +34 -28
  81. package/dist/types/packages/sdk/src/experiment-draft-storage.d.ts.map +1 -1
  82. package/dist/types/packages/sdk/src/experiment-improvement-client.d.ts +1889 -1889
  83. package/dist/types/packages/sdk/src/experiment-run-details.d.ts +7 -7
  84. package/dist/types/packages/sdk/src/experiment-scoring-contracts.d.ts +21 -3
  85. package/dist/types/packages/sdk/src/experiment-scoring-contracts.d.ts.map +1 -1
  86. package/dist/types/packages/sdk/src/learning.d.ts +1 -0
  87. package/dist/types/packages/sdk/src/learning.d.ts.map +1 -1
  88. package/dist/types/packages/sdk/src/managed-alphaedit-executions.d.ts +272 -16
  89. package/dist/types/packages/sdk/src/managed-alphaedit-executions.d.ts.map +1 -1
  90. package/dist/types/packages/sdk/src/managed-compute-price.d.ts +53 -1
  91. package/dist/types/packages/sdk/src/managed-compute-price.d.ts.map +1 -1
  92. package/dist/types/packages/sdk/src/managed-decision-executions.d.ts +612 -356
  93. package/dist/types/packages/sdk/src/managed-decision-executions.d.ts.map +1 -1
  94. package/dist/types/packages/sdk/src/managed-engram-executions.d.ts +272 -16
  95. package/dist/types/packages/sdk/src/managed-engram-executions.d.ts.map +1 -1
  96. package/dist/types/packages/sdk/src/managed-modification-executions.d.ts +208 -16
  97. package/dist/types/packages/sdk/src/managed-modification-executions.d.ts.map +1 -1
  98. package/dist/types/packages/sdk/src/managed-neuron-training-executions.d.ts +272 -16
  99. package/dist/types/packages/sdk/src/managed-neuron-training-executions.d.ts.map +1 -1
  100. package/dist/types/packages/sdk/src/managed-serving-usage.d.ts +3 -0
  101. package/dist/types/packages/sdk/src/managed-serving-usage.d.ts.map +1 -1
  102. package/dist/types/packages/sdk/src/managed-training-preparation.d.ts +25 -25
  103. package/dist/types/packages/sdk/src/managed-training-preparation.d.ts.map +1 -1
  104. package/dist/types/packages/sdk/src/model-learning-capabilities.d.ts +4 -0
  105. package/dist/types/packages/sdk/src/model-learning-capabilities.d.ts.map +1 -0
  106. package/dist/types/packages/sdk/src/model-learning-policy.d.ts.map +1 -1
  107. package/dist/types/packages/sdk/src/model-supervised-training-defaults.d.ts.map +1 -1
  108. package/dist/types/packages/sdk/src/model-training-defaults.d.ts +5 -2
  109. package/dist/types/packages/sdk/src/model-training-defaults.d.ts.map +1 -1
  110. package/dist/types/packages/sdk/src/neuron-training-model-serving.d.ts +8 -4
  111. package/dist/types/packages/sdk/src/neuron-training-model-serving.d.ts.map +1 -1
  112. package/dist/types/packages/sdk/src/post-training.d.ts +7 -7
  113. package/dist/types/packages/sdk/src/product-model-response.d.ts +6 -2
  114. package/dist/types/packages/sdk/src/product-model-response.d.ts.map +1 -1
  115. package/dist/types/packages/sdk/src/product-model-transport.d.ts +99 -0
  116. package/dist/types/packages/sdk/src/product-model-transport.d.ts.map +1 -0
  117. package/dist/types/packages/sdk/src/product-model-wire.d.ts +104 -0
  118. package/dist/types/packages/sdk/src/product-model-wire.d.ts.map +1 -0
  119. package/dist/types/packages/sdk/src/product-models.d.ts +18 -15
  120. package/dist/types/packages/sdk/src/product-models.d.ts.map +1 -1
  121. package/dist/types/packages/sdk/src/serverless-serving.d.ts +416 -9
  122. package/dist/types/packages/sdk/src/serverless-serving.d.ts.map +1 -1
  123. package/dist/types/packages/sdk/src/steering-model-serving.d.ts +2144 -703
  124. package/dist/types/packages/sdk/src/steering-model-serving.d.ts.map +1 -1
  125. package/dist/types/packages/sdk/src/steering-profile-publication.d.ts +631 -0
  126. package/dist/types/packages/sdk/src/steering-profile-publication.d.ts.map +1 -1
  127. package/dist/types/packages/sdk/src/taskset-draft-client.d.ts +120 -120
  128. package/dist/types/packages/sdk/src/taskset-package-projection.d.ts +20 -20
  129. package/dist/types/packages/sdk/src/taskset-packages.d.ts +1 -0
  130. package/dist/types/packages/sdk/src/taskset-packages.d.ts.map +1 -1
  131. package/dist/types/packages/sdk/src/taskset-text-policy.d.ts +11 -0
  132. package/dist/types/packages/sdk/src/taskset-text-policy.d.ts.map +1 -0
  133. package/dist/types/packages/sdk/src/training-compute-selection.d.ts +6 -19
  134. package/dist/types/packages/sdk/src/training-compute-selection.d.ts.map +1 -1
  135. package/dist/types/packages/sdk/src/training-estimates.d.ts +1 -1
  136. package/dist/types/packages/sdk/src/training-experiment-setup.d.ts +4 -1
  137. package/dist/types/packages/sdk/src/training-experiment-setup.d.ts.map +1 -1
  138. package/dist/types/packages/sdk/src/training-harness-sources.d.ts +2 -2
  139. package/dist/types/packages/sdk/src/training-learning-batch.d.ts +40 -40
  140. package/dist/types/packages/sdk/src/training-model-serving-control.d.ts +406 -8
  141. package/dist/types/packages/sdk/src/training-model-serving-control.d.ts.map +1 -1
  142. package/dist/types/packages/sdk/src/training-model-serving.d.ts +1136 -30
  143. package/dist/types/packages/sdk/src/training-model-serving.d.ts.map +1 -1
  144. package/dist/types/packages/sdk/src/training-operations.d.ts +11197 -0
  145. package/dist/types/packages/sdk/src/training-operations.d.ts.map +1 -0
  146. package/dist/types/packages/sdk/src/training-qualification-preparation.d.ts +25 -25
  147. package/dist/types/packages/sdk/src/training-supervised.d.ts +13 -0
  148. package/dist/types/packages/sdk/src/training-supervised.d.ts.map +1 -1
  149. package/dist/types/packages/sdk/src/training-transport.d.ts +4 -1
  150. package/dist/types/packages/sdk/src/training-transport.d.ts.map +1 -1
  151. package/dist/types/packages/sdk/src/training.d.ts +15885 -4662
  152. package/dist/types/packages/sdk/src/training.d.ts.map +1 -1
  153. package/dist/types/packages/sdk/src/weight-inference.d.ts +8 -5
  154. package/dist/types/packages/sdk/src/weight-inference.d.ts.map +1 -1
  155. package/dist/types/packages/sdk/src/weight-model-versions.d.ts +2089 -398
  156. package/dist/types/packages/sdk/src/weight-model-versions.d.ts.map +1 -1
  157. package/dist/types/packages/sdk/src/weight-space-artifact.d.ts +433 -0
  158. package/dist/types/packages/sdk/src/weight-space-artifact.d.ts.map +1 -0
  159. package/dist/types/packages/sdk/src/weight-space-evaluation.d.ts +1437 -0
  160. package/dist/types/packages/sdk/src/weight-space-evaluation.d.ts.map +1 -0
  161. package/dist/types/packages/sdk/src/weight-space-model-serving.d.ts +2910 -0
  162. package/dist/types/packages/sdk/src/weight-space-model-serving.d.ts.map +1 -0
  163. package/dist/types/packages/sdk/src/weight-space-programming.d.ts +845 -0
  164. package/dist/types/packages/sdk/src/weight-space-programming.d.ts.map +1 -0
  165. package/dist/types/packages/sdk/src/weight-space-sessions.d.ts +4600 -0
  166. package/dist/types/packages/sdk/src/weight-space-sessions.d.ts.map +1 -0
  167. package/fixtures/python-foundation/operation-matrix.json +369 -0
  168. package/fixtures/python-foundation/schemas.json +34304 -27905
  169. package/fixtures/python-foundation/training-estimate-http-workflow.json +1 -1
  170. package/package.json +2 -2
  171. package/dist/chunk-JWEIAEQ6.js.map +0 -1
  172. package/dist/chunk-MN4FE52A.js.map +0 -1
  173. package/dist/chunk-PMVSO23F.js.map +0 -1
  174. package/dist/chunk-QKLZZOZD.js.map +0 -1
  175. package/dist/chunk-UZAVG2WK.js.map +0 -1
  176. package/dist/chunk-XYK3MAIC.js.map +0 -1
  177. package/dist/chunk-YV6PPDXZ.js.map +0 -1
  178. package/dist/chunk-ZG4HVCJB.js +0 -248
  179. package/dist/chunk-ZG4HVCJB.js.map +0 -1
  180. /package/dist/{chunk-4REK5TTE.js.map → chunk-4D4QSQC3.js.map} +0 -0
  181. /package/dist/{chunk-AMZ2SWHR.js.map → chunk-534W2O2S.js.map} +0 -0
  182. /package/dist/{chunk-6LVLWRBJ.js.map → chunk-54BJJL7G.js.map} +0 -0
  183. /package/dist/{chunk-XWUTYFOU.js.map → chunk-5KJY44FB.js.map} +0 -0
  184. /package/dist/{chunk-4A4FNG4R.js.map → chunk-7ZV626YS.js.map} +0 -0
  185. /package/dist/{chunk-57ROW3XF.js.map → chunk-ABOZV47D.js.map} +0 -0
  186. /package/dist/{chunk-3BILL33A.js.map → chunk-BIQ5TG3T.js.map} +0 -0
  187. /package/dist/{chunk-JINWKUKR.js.map → chunk-BT6CC6QB.js.map} +0 -0
  188. /package/dist/{chunk-4PEK3YOW.js.map → chunk-CNFN6GHY.js.map} +0 -0
  189. /package/dist/{chunk-X5AT253K.js.map → chunk-CVYGJUWM.js.map} +0 -0
  190. /package/dist/{chunk-D26RJQU7.js.map → chunk-CX7ETU2F.js.map} +0 -0
  191. /package/dist/{chunk-4CDLPMJK.js.map → chunk-D7LV3USE.js.map} +0 -0
  192. /package/dist/{chunk-WZKKQNYY.js.map → chunk-GA7OSSEA.js.map} +0 -0
  193. /package/dist/{chunk-EL6KZO43.js.map → chunk-GFWXDWEE.js.map} +0 -0
  194. /package/dist/{chunk-AQV7RQZH.js.map → chunk-K4WU4JOP.js.map} +0 -0
  195. /package/dist/{chunk-CAOMKM55.js.map → chunk-L7MVGGBH.js.map} +0 -0
  196. /package/dist/{chunk-P56JWQYV.js.map → chunk-LRFVISLA.js.map} +0 -0
  197. /package/dist/{chunk-2U6BQ5C4.js.map → chunk-MJHQMQEW.js.map} +0 -0
  198. /package/dist/{chunk-C7I4ZG6Q.js.map → chunk-TX552FQG.js.map} +0 -0
  199. /package/dist/{chunk-4CRLQMJ2.js.map → chunk-VIDRZCPQ.js.map} +0 -0
  200. /package/dist/{chunk-HCUZYGUK.js.map → chunk-WPTFJZ4L.js.map} +0 -0
@@ -0,0 +1,1437 @@
1
+ import { type WeightSpaceRequest } from "./weight-space-programming.js";
2
+ /** Resolve a policy-visible panel from an already authorized immutable package.
3
+ * Hosts must authorize the exact release again before dispatch and scoring.
4
+ * The returned envelope contains no answers, graders or private context. */
5
+ export declare function resolveWeightSpaceTasksetPanel(raw: unknown, input: {
6
+ source: NonNullable<WeightSpaceRequest["frozenEvaluation"]>;
7
+ role: "development" | "preservation" | "frozen";
8
+ taskIds?: string[];
9
+ }): {
10
+ package: {
11
+ schemaVersion: "openpond.tasksetPackage.v1";
12
+ taskset: {
13
+ schemaVersion: "openpond.tasksetRelease.v2";
14
+ id: string;
15
+ revision: number;
16
+ policy: {
17
+ policyVisibleFields: string[];
18
+ privilegedFields: string[];
19
+ hiddenGraderRefs: string[];
20
+ connectedAppScopes: string[];
21
+ };
22
+ environment: {
23
+ protocolVersion: "openpond.environment.v1";
24
+ kind: "agent" | "custom_program" | "text" | "work";
25
+ entrypoint: string;
26
+ stateful: boolean;
27
+ deterministicSeeds: boolean;
28
+ lifecycle: ("collect" | "create" | "destroy" | "reset" | "step")[];
29
+ networkPolicy: "declared_read_only" | "declared_scoped" | "none";
30
+ defaultTimeoutMs: number;
31
+ };
32
+ environmentRelease?: {
33
+ id: string;
34
+ contentHash: string;
35
+ } | undefined;
36
+ tools: {
37
+ name: string;
38
+ description: string;
39
+ inputSchema: Record<string, unknown>;
40
+ inputSchemaHash: string;
41
+ sideEffect: "read" | "write";
42
+ timeoutMs: number;
43
+ }[];
44
+ capabilities: {
45
+ id: string;
46
+ required: boolean;
47
+ scopes: string[];
48
+ portability: "host_adapter" | "hosted_only" | "local_only" | "portable";
49
+ }[];
50
+ tasks: {
51
+ id: string;
52
+ clusterKey: string;
53
+ split: "frozen_eval" | "test" | "train" | "validation";
54
+ input: Record<string, unknown>;
55
+ expectedOutput: Record<string, unknown> | null;
56
+ policyVisibleContext: Record<string, unknown>;
57
+ privilegedContextRef: string | null;
58
+ artifactRefs: {
59
+ id: string;
60
+ path: string;
61
+ contentHash: string;
62
+ sizeBytes: number;
63
+ mediaType: string;
64
+ visibility: "host_private" | "policy" | "verifier";
65
+ }[];
66
+ requiredOutputs?: {
67
+ path: string;
68
+ mediaType: string;
69
+ schemaRef: {
70
+ id: string;
71
+ path: string;
72
+ contentHash: string;
73
+ sizeBytes: number;
74
+ mediaType: string;
75
+ visibility: "host_private" | "policy" | "verifier";
76
+ } | null;
77
+ submissionMode?: "structured_json" | undefined;
78
+ maxBytes: number | null;
79
+ metadata: Record<string, unknown>;
80
+ }[] | undefined;
81
+ tags: string[];
82
+ }[];
83
+ graders: ({
84
+ id: string;
85
+ taskIds?: string[] | undefined;
86
+ version: string;
87
+ weight: number;
88
+ hardGate: boolean;
89
+ rewardEligible: boolean;
90
+ privileged: boolean;
91
+ kind: "artifact" | "content" | "runtime_event" | "schema" | "state";
92
+ config: Record<string, unknown>;
93
+ } | {
94
+ id: string;
95
+ taskIds?: string[] | undefined;
96
+ version: string;
97
+ weight: number;
98
+ hardGate: boolean;
99
+ rewardEligible: boolean;
100
+ privileged: boolean;
101
+ kind: "model_judge";
102
+ rubricRef: {
103
+ id: string;
104
+ path: string;
105
+ contentHash: string;
106
+ sizeBytes: number;
107
+ mediaType: string;
108
+ visibility: "host_private" | "policy" | "verifier";
109
+ };
110
+ calibrationStatus: "failed" | "passed" | "pending";
111
+ model?: {
112
+ providerId: string;
113
+ modelId: string;
114
+ revision: string | null;
115
+ } | undefined;
116
+ temperature?: number | undefined;
117
+ } | {
118
+ id: string;
119
+ taskIds?: string[] | undefined;
120
+ version: string;
121
+ weight: number;
122
+ hardGate: boolean;
123
+ rewardEligible: boolean;
124
+ privileged: boolean;
125
+ kind: "custom_verifier";
126
+ runtime?: "isolated_javascript" | "sandbox_process" | undefined;
127
+ verifierRef: {
128
+ id: string;
129
+ path: string;
130
+ contentHash: string;
131
+ sizeBytes: number;
132
+ mediaType: string;
133
+ visibility: "host_private" | "policy" | "verifier";
134
+ };
135
+ exportName?: string | undefined;
136
+ timeoutMs: number;
137
+ networkPolicy: "none";
138
+ } | {
139
+ id: string;
140
+ taskIds?: string[] | undefined;
141
+ version: string;
142
+ weight: number;
143
+ hardGate: boolean;
144
+ rewardEligible: boolean;
145
+ privileged: boolean;
146
+ kind: "human";
147
+ rubricRef: {
148
+ id: string;
149
+ path: string;
150
+ contentHash: string;
151
+ sizeBytes: number;
152
+ mediaType: string;
153
+ visibility: "host_private" | "policy" | "verifier";
154
+ };
155
+ reviewerRole: string;
156
+ form?: {
157
+ schemaVersion: "openpond.humanForm.v1";
158
+ mode: "individual" | "pairwise";
159
+ instructions: string;
160
+ criteria: ({
161
+ id: string;
162
+ label: string;
163
+ instructions: string;
164
+ required: boolean;
165
+ allowAbstain: boolean;
166
+ requireRationale?: boolean | undefined;
167
+ anchors?: {
168
+ value: string | number | boolean;
169
+ description: string;
170
+ }[] | undefined;
171
+ kind: "score";
172
+ minimum: number;
173
+ maximum: number;
174
+ step: number;
175
+ } | {
176
+ id: string;
177
+ label: string;
178
+ instructions: string;
179
+ required: boolean;
180
+ allowAbstain: boolean;
181
+ requireRationale?: boolean | undefined;
182
+ anchors?: {
183
+ value: string | number | boolean;
184
+ description: string;
185
+ }[] | undefined;
186
+ kind: "category";
187
+ options: string[];
188
+ } | {
189
+ id: string;
190
+ label: string;
191
+ instructions: string;
192
+ required: boolean;
193
+ allowAbstain: boolean;
194
+ requireRationale?: boolean | undefined;
195
+ anchors?: {
196
+ value: string | number | boolean;
197
+ description: string;
198
+ }[] | undefined;
199
+ kind: "boolean";
200
+ } | {
201
+ id: string;
202
+ label: string;
203
+ instructions: string;
204
+ required: boolean;
205
+ allowAbstain: boolean;
206
+ requireRationale?: boolean | undefined;
207
+ anchors?: {
208
+ value: string | number | boolean;
209
+ description: string;
210
+ }[] | undefined;
211
+ kind: "text";
212
+ minimumLength?: number | undefined;
213
+ maximumLength: number;
214
+ } | {
215
+ id: string;
216
+ label: string;
217
+ instructions: string;
218
+ required: boolean;
219
+ allowAbstain: boolean;
220
+ requireRationale?: boolean | undefined;
221
+ anchors?: {
222
+ value: string | number | boolean;
223
+ description: string;
224
+ }[] | undefined;
225
+ kind: "preference";
226
+ requireStrength?: boolean | undefined;
227
+ allowTie: boolean;
228
+ })[];
229
+ } | undefined;
230
+ })[];
231
+ verifierSetRelease?: {
232
+ id: string;
233
+ contentHash: string;
234
+ } | undefined;
235
+ metrics?: {
236
+ schemaVersion: "openpond.tasksetMetricPolicy.v1";
237
+ primaryMetric: string;
238
+ aggregation: "custom" | "mean_score" | "pass_rate" | "weighted_mean";
239
+ missingReward: "exclude" | "zero";
240
+ taskWeights?: Record<string, number> | undefined;
241
+ customAggregator: {
242
+ module: string;
243
+ exportName: string;
244
+ contentHash: string;
245
+ timeoutMs: number;
246
+ networkPolicy: "none";
247
+ } | null;
248
+ } | undefined;
249
+ metadata: Record<string, unknown>;
250
+ contentHash: string;
251
+ };
252
+ environment: {
253
+ schemaVersion: "openpond.environmentRelease.v1";
254
+ id: string;
255
+ revision: number;
256
+ contract: {
257
+ protocolVersion: "openpond.environment.v1";
258
+ kind: "agent" | "custom_program" | "text" | "work";
259
+ entrypoint: string;
260
+ stateful: boolean;
261
+ deterministicSeeds: boolean;
262
+ lifecycle: ("collect" | "create" | "destroy" | "reset" | "step")[];
263
+ networkPolicy: "declared_read_only" | "declared_scoped" | "none";
264
+ defaultTimeoutMs: number;
265
+ };
266
+ actionSchemaRef: {
267
+ id: string;
268
+ path: string;
269
+ contentHash: string;
270
+ sizeBytes: number;
271
+ mediaType: string;
272
+ visibility: "host_private" | "policy" | "verifier";
273
+ } | null;
274
+ observationSchemaRef: {
275
+ id: string;
276
+ path: string;
277
+ contentHash: string;
278
+ sizeBytes: number;
279
+ mediaType: string;
280
+ visibility: "host_private" | "policy" | "verifier";
281
+ } | null;
282
+ stateSchemaRef: {
283
+ id: string;
284
+ path: string;
285
+ contentHash: string;
286
+ sizeBytes: number;
287
+ mediaType: string;
288
+ visibility: "host_private" | "policy" | "verifier";
289
+ } | null;
290
+ artifactCollection: {
291
+ maxArtifacts: number;
292
+ maxTotalBytes: number;
293
+ };
294
+ adapterConformanceHashes: Record<string, string>;
295
+ metadata: Record<string, unknown>;
296
+ contentHash: string;
297
+ };
298
+ verifierSet: {
299
+ schemaVersion: "openpond.verifierSetRelease.v1";
300
+ id: string;
301
+ revision: number;
302
+ graders: ({
303
+ id: string;
304
+ taskIds?: string[] | undefined;
305
+ version: string;
306
+ weight: number;
307
+ hardGate: boolean;
308
+ rewardEligible: boolean;
309
+ privileged: boolean;
310
+ kind: "artifact" | "content" | "runtime_event" | "schema" | "state";
311
+ config: Record<string, unknown>;
312
+ } | {
313
+ id: string;
314
+ taskIds?: string[] | undefined;
315
+ version: string;
316
+ weight: number;
317
+ hardGate: boolean;
318
+ rewardEligible: boolean;
319
+ privileged: boolean;
320
+ kind: "model_judge";
321
+ rubricRef: {
322
+ id: string;
323
+ path: string;
324
+ contentHash: string;
325
+ sizeBytes: number;
326
+ mediaType: string;
327
+ visibility: "host_private" | "policy" | "verifier";
328
+ };
329
+ calibrationStatus: "failed" | "passed" | "pending";
330
+ model?: {
331
+ providerId: string;
332
+ modelId: string;
333
+ revision: string | null;
334
+ } | undefined;
335
+ temperature?: number | undefined;
336
+ } | {
337
+ id: string;
338
+ taskIds?: string[] | undefined;
339
+ version: string;
340
+ weight: number;
341
+ hardGate: boolean;
342
+ rewardEligible: boolean;
343
+ privileged: boolean;
344
+ kind: "custom_verifier";
345
+ runtime?: "isolated_javascript" | "sandbox_process" | undefined;
346
+ verifierRef: {
347
+ id: string;
348
+ path: string;
349
+ contentHash: string;
350
+ sizeBytes: number;
351
+ mediaType: string;
352
+ visibility: "host_private" | "policy" | "verifier";
353
+ };
354
+ exportName?: string | undefined;
355
+ timeoutMs: number;
356
+ networkPolicy: "none";
357
+ } | {
358
+ id: string;
359
+ taskIds?: string[] | undefined;
360
+ version: string;
361
+ weight: number;
362
+ hardGate: boolean;
363
+ rewardEligible: boolean;
364
+ privileged: boolean;
365
+ kind: "human";
366
+ rubricRef: {
367
+ id: string;
368
+ path: string;
369
+ contentHash: string;
370
+ sizeBytes: number;
371
+ mediaType: string;
372
+ visibility: "host_private" | "policy" | "verifier";
373
+ };
374
+ reviewerRole: string;
375
+ form?: {
376
+ schemaVersion: "openpond.humanForm.v1";
377
+ mode: "individual" | "pairwise";
378
+ instructions: string;
379
+ criteria: ({
380
+ id: string;
381
+ label: string;
382
+ instructions: string;
383
+ required: boolean;
384
+ allowAbstain: boolean;
385
+ requireRationale?: boolean | undefined;
386
+ anchors?: {
387
+ value: string | number | boolean;
388
+ description: string;
389
+ }[] | undefined;
390
+ kind: "score";
391
+ minimum: number;
392
+ maximum: number;
393
+ step: number;
394
+ } | {
395
+ id: string;
396
+ label: string;
397
+ instructions: string;
398
+ required: boolean;
399
+ allowAbstain: boolean;
400
+ requireRationale?: boolean | undefined;
401
+ anchors?: {
402
+ value: string | number | boolean;
403
+ description: string;
404
+ }[] | undefined;
405
+ kind: "category";
406
+ options: string[];
407
+ } | {
408
+ id: string;
409
+ label: string;
410
+ instructions: string;
411
+ required: boolean;
412
+ allowAbstain: boolean;
413
+ requireRationale?: boolean | undefined;
414
+ anchors?: {
415
+ value: string | number | boolean;
416
+ description: string;
417
+ }[] | undefined;
418
+ kind: "boolean";
419
+ } | {
420
+ id: string;
421
+ label: string;
422
+ instructions: string;
423
+ required: boolean;
424
+ allowAbstain: boolean;
425
+ requireRationale?: boolean | undefined;
426
+ anchors?: {
427
+ value: string | number | boolean;
428
+ description: string;
429
+ }[] | undefined;
430
+ kind: "text";
431
+ minimumLength?: number | undefined;
432
+ maximumLength: number;
433
+ } | {
434
+ id: string;
435
+ label: string;
436
+ instructions: string;
437
+ required: boolean;
438
+ allowAbstain: boolean;
439
+ requireRationale?: boolean | undefined;
440
+ anchors?: {
441
+ value: string | number | boolean;
442
+ description: string;
443
+ }[] | undefined;
444
+ kind: "preference";
445
+ requireStrength?: boolean | undefined;
446
+ allowTie: boolean;
447
+ })[];
448
+ } | undefined;
449
+ })[];
450
+ isolation: {
451
+ processBoundary: "container" | "isolated_process" | "same_process";
452
+ networkPolicy: "none";
453
+ defaultTimeoutMs: number;
454
+ };
455
+ calibrationReceiptRefs: {
456
+ id: string;
457
+ contentHash: string;
458
+ }[];
459
+ metadata: Record<string, unknown>;
460
+ contentHash: string;
461
+ };
462
+ files: {
463
+ asset: {
464
+ id: string;
465
+ path: string;
466
+ contentHash: string;
467
+ sizeBytes: number;
468
+ mediaType: string;
469
+ visibility: "host_private" | "policy" | "verifier";
470
+ };
471
+ base64: string;
472
+ }[];
473
+ modelResources?: {
474
+ taskDefinition: {
475
+ schemaVersion: "openpond.taskDefinition.v1";
476
+ id: string;
477
+ revision: number;
478
+ name: string;
479
+ description: string;
480
+ instructions: string;
481
+ category: "coding" | "custom" | "question_answering" | "structured" | "tool_workflow";
482
+ familyNamespace: string;
483
+ inputSchema: Record<string, unknown>;
484
+ outputSchema: Record<string, unknown>;
485
+ requiredOutputs?: {
486
+ path: string;
487
+ mediaType: string;
488
+ schemaRef: {
489
+ id: string;
490
+ path: string;
491
+ contentHash: string;
492
+ sizeBytes: number;
493
+ mediaType: string;
494
+ visibility: "host_private" | "policy" | "verifier";
495
+ } | null;
496
+ submissionMode?: "structured_json" | undefined;
497
+ maxBytes: number | null;
498
+ metadata: Record<string, unknown>;
499
+ }[] | undefined;
500
+ contextRequirements?: string[] | undefined;
501
+ rewardBinding: {
502
+ id: string;
503
+ contentHash: string;
504
+ revision: number;
505
+ };
506
+ harness: {
507
+ id: string;
508
+ contentHash: string;
509
+ } | null;
510
+ execution: {
511
+ policy: {
512
+ policyVisibleFields: string[];
513
+ privilegedFields: string[];
514
+ hiddenGraderRefs: string[];
515
+ connectedAppScopes: string[];
516
+ };
517
+ environment: {
518
+ protocolVersion: "openpond.environment.v1";
519
+ kind: "agent" | "custom_program" | "text" | "work";
520
+ entrypoint: string;
521
+ stateful: boolean;
522
+ deterministicSeeds: boolean;
523
+ lifecycle: ("collect" | "create" | "destroy" | "reset" | "step")[];
524
+ networkPolicy: "declared_read_only" | "declared_scoped" | "none";
525
+ defaultTimeoutMs: number;
526
+ };
527
+ environmentRelease?: {
528
+ id: string;
529
+ contentHash: string;
530
+ } | undefined;
531
+ tools: {
532
+ name: string;
533
+ description: string;
534
+ inputSchema: Record<string, unknown>;
535
+ inputSchemaHash: string;
536
+ sideEffect: "read" | "write";
537
+ timeoutMs: number;
538
+ }[];
539
+ capabilities: {
540
+ id: string;
541
+ required: boolean;
542
+ scopes: string[];
543
+ portability: "host_adapter" | "hosted_only" | "local_only" | "portable";
544
+ }[];
545
+ verifierSetRelease?: {
546
+ id: string;
547
+ contentHash: string;
548
+ } | undefined;
549
+ };
550
+ contentHash: string;
551
+ };
552
+ rewardBinding: {
553
+ schemaVersion: "openpond.rewardBinding.v1";
554
+ id: string;
555
+ revision: number;
556
+ name?: string | undefined;
557
+ description?: string | undefined;
558
+ recipeRef?: {
559
+ id: string;
560
+ contentHash: string;
561
+ revision: number;
562
+ } | undefined;
563
+ sources: {
564
+ graderId: string;
565
+ taskIds?: string[] | undefined;
566
+ reward: {
567
+ id: string;
568
+ contentHash: string;
569
+ revision: number;
570
+ };
571
+ role: "evaluation" | "training";
572
+ normalization: {
573
+ kind: "identity";
574
+ } | {
575
+ kind: "linear";
576
+ minimum: number;
577
+ maximum: number;
578
+ direction: "higher" | "lower";
579
+ };
580
+ weight: number;
581
+ required: boolean;
582
+ hardGate: boolean;
583
+ privileged: boolean;
584
+ fixtureRefs: {
585
+ id: string;
586
+ contentHash: string;
587
+ }[];
588
+ }[];
589
+ aggregation: "weighted_mean";
590
+ unscorable: "exclude_optional_require_all_required";
591
+ contentHash: string;
592
+ };
593
+ rewards: {
594
+ schemaVersion: "openpond.rewardRelease.v1";
595
+ id: string;
596
+ revision: number;
597
+ name: string;
598
+ description: string;
599
+ feedbackKey?: string | undefined;
600
+ implementation: {
601
+ kind: "artifact" | "content" | "runtime_event" | "schema" | "state";
602
+ config: Record<string, unknown>;
603
+ } | {
604
+ kind: "model_judge";
605
+ rubricRef: {
606
+ id: string;
607
+ path: string;
608
+ contentHash: string;
609
+ sizeBytes: number;
610
+ mediaType: string;
611
+ visibility: "host_private" | "policy" | "verifier";
612
+ };
613
+ calibrationStatus: "failed" | "passed" | "pending";
614
+ model?: {
615
+ providerId: string;
616
+ modelId: string;
617
+ revision: string | null;
618
+ } | undefined;
619
+ temperature?: number | undefined;
620
+ } | {
621
+ kind: "custom_verifier";
622
+ runtime?: "isolated_javascript" | "sandbox_process" | undefined;
623
+ verifierRef: {
624
+ id: string;
625
+ path: string;
626
+ contentHash: string;
627
+ sizeBytes: number;
628
+ mediaType: string;
629
+ visibility: "host_private" | "policy" | "verifier";
630
+ };
631
+ exportName?: string | undefined;
632
+ timeoutMs: number;
633
+ networkPolicy: "none";
634
+ } | {
635
+ kind: "human";
636
+ rubricRef: {
637
+ id: string;
638
+ path: string;
639
+ contentHash: string;
640
+ sizeBytes: number;
641
+ mediaType: string;
642
+ visibility: "host_private" | "policy" | "verifier";
643
+ };
644
+ reviewerRole: string;
645
+ form?: {
646
+ schemaVersion: "openpond.humanForm.v1";
647
+ mode: "individual" | "pairwise";
648
+ instructions: string;
649
+ criteria: ({
650
+ id: string;
651
+ label: string;
652
+ instructions: string;
653
+ required: boolean;
654
+ allowAbstain: boolean;
655
+ requireRationale?: boolean | undefined;
656
+ anchors?: {
657
+ value: string | number | boolean;
658
+ description: string;
659
+ }[] | undefined;
660
+ kind: "score";
661
+ minimum: number;
662
+ maximum: number;
663
+ step: number;
664
+ } | {
665
+ id: string;
666
+ label: string;
667
+ instructions: string;
668
+ required: boolean;
669
+ allowAbstain: boolean;
670
+ requireRationale?: boolean | undefined;
671
+ anchors?: {
672
+ value: string | number | boolean;
673
+ description: string;
674
+ }[] | undefined;
675
+ kind: "category";
676
+ options: string[];
677
+ } | {
678
+ id: string;
679
+ label: string;
680
+ instructions: string;
681
+ required: boolean;
682
+ allowAbstain: boolean;
683
+ requireRationale?: boolean | undefined;
684
+ anchors?: {
685
+ value: string | number | boolean;
686
+ description: string;
687
+ }[] | undefined;
688
+ kind: "boolean";
689
+ } | {
690
+ id: string;
691
+ label: string;
692
+ instructions: string;
693
+ required: boolean;
694
+ allowAbstain: boolean;
695
+ requireRationale?: boolean | undefined;
696
+ anchors?: {
697
+ value: string | number | boolean;
698
+ description: string;
699
+ }[] | undefined;
700
+ kind: "text";
701
+ minimumLength?: number | undefined;
702
+ maximumLength: number;
703
+ } | {
704
+ id: string;
705
+ label: string;
706
+ instructions: string;
707
+ required: boolean;
708
+ allowAbstain: boolean;
709
+ requireRationale?: boolean | undefined;
710
+ anchors?: {
711
+ value: string | number | boolean;
712
+ description: string;
713
+ }[] | undefined;
714
+ kind: "preference";
715
+ requireStrength?: boolean | undefined;
716
+ allowTie: boolean;
717
+ })[];
718
+ } | undefined;
719
+ } | {
720
+ kind: "learned_model";
721
+ modelVersion: {
722
+ id: string;
723
+ contentHash: string;
724
+ };
725
+ inputContract: {
726
+ id: string;
727
+ path: string;
728
+ contentHash: string;
729
+ sizeBytes: number;
730
+ mediaType: string;
731
+ visibility: "host_private" | "policy" | "verifier";
732
+ };
733
+ };
734
+ rawScore: {
735
+ minimum: number;
736
+ maximum: number;
737
+ };
738
+ assets: {
739
+ id: string;
740
+ path: string;
741
+ contentHash: string;
742
+ sizeBytes: number;
743
+ mediaType: string;
744
+ visibility: "host_private" | "policy" | "verifier";
745
+ }[];
746
+ fixtureSetRef?: {
747
+ id: string;
748
+ path: string;
749
+ contentHash: string;
750
+ sizeBytes: number;
751
+ mediaType: string;
752
+ visibility: "host_private" | "policy" | "verifier";
753
+ } | undefined;
754
+ calibrationCheckRef?: {
755
+ id: string;
756
+ contentHash: string;
757
+ revision: number;
758
+ } | undefined;
759
+ contentHash: string;
760
+ }[];
761
+ assets: {
762
+ schemaVersion: "openpond.learningTextAsset.v1";
763
+ id: string;
764
+ revision: 1;
765
+ asset: {
766
+ id: string;
767
+ path: string;
768
+ contentHash: string;
769
+ sizeBytes: number;
770
+ mediaType: string;
771
+ visibility: "host_private" | "policy" | "verifier";
772
+ };
773
+ text: string;
774
+ contentHash: string;
775
+ }[];
776
+ execution?: {
777
+ environment: {
778
+ schemaVersion: "openpond.environmentRelease.v1";
779
+ id: string;
780
+ revision: number;
781
+ contract: {
782
+ protocolVersion: "openpond.environment.v1";
783
+ kind: "agent" | "custom_program" | "text" | "work";
784
+ entrypoint: string;
785
+ stateful: boolean;
786
+ deterministicSeeds: boolean;
787
+ lifecycle: ("collect" | "create" | "destroy" | "reset" | "step")[];
788
+ networkPolicy: "declared_read_only" | "declared_scoped" | "none";
789
+ defaultTimeoutMs: number;
790
+ };
791
+ actionSchemaRef: {
792
+ id: string;
793
+ path: string;
794
+ contentHash: string;
795
+ sizeBytes: number;
796
+ mediaType: string;
797
+ visibility: "host_private" | "policy" | "verifier";
798
+ } | null;
799
+ observationSchemaRef: {
800
+ id: string;
801
+ path: string;
802
+ contentHash: string;
803
+ sizeBytes: number;
804
+ mediaType: string;
805
+ visibility: "host_private" | "policy" | "verifier";
806
+ } | null;
807
+ stateSchemaRef: {
808
+ id: string;
809
+ path: string;
810
+ contentHash: string;
811
+ sizeBytes: number;
812
+ mediaType: string;
813
+ visibility: "host_private" | "policy" | "verifier";
814
+ } | null;
815
+ artifactCollection: {
816
+ maxArtifacts: number;
817
+ maxTotalBytes: number;
818
+ };
819
+ adapterConformanceHashes: Record<string, string>;
820
+ metadata: Record<string, unknown>;
821
+ contentHash: string;
822
+ };
823
+ verifierSet: {
824
+ schemaVersion: "openpond.verifierSetRelease.v1";
825
+ id: string;
826
+ revision: number;
827
+ graders: ({
828
+ id: string;
829
+ taskIds?: string[] | undefined;
830
+ version: string;
831
+ weight: number;
832
+ hardGate: boolean;
833
+ rewardEligible: boolean;
834
+ privileged: boolean;
835
+ kind: "artifact" | "content" | "runtime_event" | "schema" | "state";
836
+ config: Record<string, unknown>;
837
+ } | {
838
+ id: string;
839
+ taskIds?: string[] | undefined;
840
+ version: string;
841
+ weight: number;
842
+ hardGate: boolean;
843
+ rewardEligible: boolean;
844
+ privileged: boolean;
845
+ kind: "model_judge";
846
+ rubricRef: {
847
+ id: string;
848
+ path: string;
849
+ contentHash: string;
850
+ sizeBytes: number;
851
+ mediaType: string;
852
+ visibility: "host_private" | "policy" | "verifier";
853
+ };
854
+ calibrationStatus: "failed" | "passed" | "pending";
855
+ model?: {
856
+ providerId: string;
857
+ modelId: string;
858
+ revision: string | null;
859
+ } | undefined;
860
+ temperature?: number | undefined;
861
+ } | {
862
+ id: string;
863
+ taskIds?: string[] | undefined;
864
+ version: string;
865
+ weight: number;
866
+ hardGate: boolean;
867
+ rewardEligible: boolean;
868
+ privileged: boolean;
869
+ kind: "custom_verifier";
870
+ runtime?: "isolated_javascript" | "sandbox_process" | undefined;
871
+ verifierRef: {
872
+ id: string;
873
+ path: string;
874
+ contentHash: string;
875
+ sizeBytes: number;
876
+ mediaType: string;
877
+ visibility: "host_private" | "policy" | "verifier";
878
+ };
879
+ exportName?: string | undefined;
880
+ timeoutMs: number;
881
+ networkPolicy: "none";
882
+ } | {
883
+ id: string;
884
+ taskIds?: string[] | undefined;
885
+ version: string;
886
+ weight: number;
887
+ hardGate: boolean;
888
+ rewardEligible: boolean;
889
+ privileged: boolean;
890
+ kind: "human";
891
+ rubricRef: {
892
+ id: string;
893
+ path: string;
894
+ contentHash: string;
895
+ sizeBytes: number;
896
+ mediaType: string;
897
+ visibility: "host_private" | "policy" | "verifier";
898
+ };
899
+ reviewerRole: string;
900
+ form?: {
901
+ schemaVersion: "openpond.humanForm.v1";
902
+ mode: "individual" | "pairwise";
903
+ instructions: string;
904
+ criteria: ({
905
+ id: string;
906
+ label: string;
907
+ instructions: string;
908
+ required: boolean;
909
+ allowAbstain: boolean;
910
+ requireRationale?: boolean | undefined;
911
+ anchors?: {
912
+ value: string | number | boolean;
913
+ description: string;
914
+ }[] | undefined;
915
+ kind: "score";
916
+ minimum: number;
917
+ maximum: number;
918
+ step: number;
919
+ } | {
920
+ id: string;
921
+ label: string;
922
+ instructions: string;
923
+ required: boolean;
924
+ allowAbstain: boolean;
925
+ requireRationale?: boolean | undefined;
926
+ anchors?: {
927
+ value: string | number | boolean;
928
+ description: string;
929
+ }[] | undefined;
930
+ kind: "category";
931
+ options: string[];
932
+ } | {
933
+ id: string;
934
+ label: string;
935
+ instructions: string;
936
+ required: boolean;
937
+ allowAbstain: boolean;
938
+ requireRationale?: boolean | undefined;
939
+ anchors?: {
940
+ value: string | number | boolean;
941
+ description: string;
942
+ }[] | undefined;
943
+ kind: "boolean";
944
+ } | {
945
+ id: string;
946
+ label: string;
947
+ instructions: string;
948
+ required: boolean;
949
+ allowAbstain: boolean;
950
+ requireRationale?: boolean | undefined;
951
+ anchors?: {
952
+ value: string | number | boolean;
953
+ description: string;
954
+ }[] | undefined;
955
+ kind: "text";
956
+ minimumLength?: number | undefined;
957
+ maximumLength: number;
958
+ } | {
959
+ id: string;
960
+ label: string;
961
+ instructions: string;
962
+ required: boolean;
963
+ allowAbstain: boolean;
964
+ requireRationale?: boolean | undefined;
965
+ anchors?: {
966
+ value: string | number | boolean;
967
+ description: string;
968
+ }[] | undefined;
969
+ kind: "preference";
970
+ requireStrength?: boolean | undefined;
971
+ allowTie: boolean;
972
+ })[];
973
+ } | undefined;
974
+ })[];
975
+ isolation: {
976
+ processBoundary: "container" | "isolated_process" | "same_process";
977
+ networkPolicy: "none";
978
+ defaultTimeoutMs: number;
979
+ };
980
+ calibrationReceiptRefs: {
981
+ id: string;
982
+ contentHash: string;
983
+ }[];
984
+ metadata: Record<string, unknown>;
985
+ contentHash: string;
986
+ };
987
+ javascript: {
988
+ schemaVersion: "openpond.javascriptEnvironment.v1";
989
+ id: string;
990
+ revision: number;
991
+ module: {
992
+ id: string;
993
+ path: string;
994
+ contentHash: string;
995
+ sizeBytes: number;
996
+ mediaType: string;
997
+ visibility: "host_private" | "policy" | "verifier";
998
+ };
999
+ tools: {
1000
+ name: string;
1001
+ description: string;
1002
+ inputSchema: Record<string, unknown>;
1003
+ inputSchemaHash: string;
1004
+ sideEffect: "read" | "write";
1005
+ timeoutMs: number;
1006
+ }[];
1007
+ maxSteps: number;
1008
+ maxStateBytes: number;
1009
+ maxObservationBytes: number;
1010
+ operationTimeoutMs: number;
1011
+ executionServices?: ({
1012
+ id: string;
1013
+ operation: "collect" | "step";
1014
+ toolName?: string | undefined;
1015
+ timeoutMs: number;
1016
+ kind: "sqlite.v1";
1017
+ sql: {
1018
+ scope: "arguments" | "initialState" | "input" | "state";
1019
+ path: string[];
1020
+ };
1021
+ snapshot: {
1022
+ scope: "arguments" | "initialState" | "input" | "state";
1023
+ path: string[];
1024
+ };
1025
+ maxRows: number;
1026
+ maxResultBytes: number;
1027
+ } | {
1028
+ id: string;
1029
+ operation: "collect" | "step";
1030
+ toolName?: string | undefined;
1031
+ timeoutMs: number;
1032
+ kind: "javascript.v1";
1033
+ source: {
1034
+ scope: "arguments" | "initialState" | "input" | "state";
1035
+ path: string[];
1036
+ };
1037
+ cases: {
1038
+ scope: "arguments" | "initialState" | "input" | "state";
1039
+ path: string[];
1040
+ };
1041
+ exportName: string;
1042
+ maxCases: number;
1043
+ maxResultBytes: number;
1044
+ })[] | undefined;
1045
+ contentHash: string;
1046
+ };
1047
+ } | undefined;
1048
+ instructionMode?: "per_task" | undefined;
1049
+ } | undefined;
1050
+ learningResources?: {
1051
+ batch: {
1052
+ schemaVersion: "openpond.taskBatch.v1";
1053
+ id: string;
1054
+ revision: 1;
1055
+ taskDefinition: {
1056
+ id: string;
1057
+ contentHash: string;
1058
+ revision: number;
1059
+ };
1060
+ rewardBinding: {
1061
+ id: string;
1062
+ contentHash: string;
1063
+ revision: number;
1064
+ };
1065
+ purpose: "evaluation" | "reward_training" | "supervised_training";
1066
+ examples: {
1067
+ evidence: {
1068
+ id: string;
1069
+ contentHash: string;
1070
+ revision: number;
1071
+ };
1072
+ decision: {
1073
+ id: string;
1074
+ contentHash: string;
1075
+ revision: number;
1076
+ };
1077
+ familyKey: string;
1078
+ inputHash: string;
1079
+ split: "frozen_eval" | "test" | "train" | "validation";
1080
+ }[];
1081
+ sealedAt: string;
1082
+ sealedBy: string;
1083
+ contentHash: string;
1084
+ };
1085
+ evidence: {
1086
+ schemaVersion: "openpond.taskEvidence.v1";
1087
+ id: string;
1088
+ revision: number;
1089
+ source: {
1090
+ id: string;
1091
+ contentHash: string;
1092
+ revision: number;
1093
+ };
1094
+ submission: {
1095
+ schemaVersion: "openpond.taskExample.v1";
1096
+ sourceId: string;
1097
+ idempotencyKey: string;
1098
+ taskDefinition: {
1099
+ id: string;
1100
+ contentHash: string;
1101
+ revision: number;
1102
+ };
1103
+ exampleId: string;
1104
+ attemptId: string;
1105
+ occurredAt: string;
1106
+ familyKey: string | null;
1107
+ split: "frozen_eval" | "test" | "train" | "validation";
1108
+ input: Record<string, import("zod").JSONType>;
1109
+ policyVisibleContext?: Record<string, import("zod").JSONType> | undefined;
1110
+ observedOutput: Record<string, import("zod").JSONType> | null;
1111
+ expected: Record<string, import("zod").JSONType> | null;
1112
+ evaluatorContext: Record<string, import("zod").JSONType> | null;
1113
+ assets: {
1114
+ id: string;
1115
+ path: string;
1116
+ contentHash: string;
1117
+ sizeBytes: number;
1118
+ mediaType: string;
1119
+ visibility: "host_private" | "policy" | "verifier";
1120
+ }[];
1121
+ requiredOutputs?: {
1122
+ path: string;
1123
+ mediaType: string;
1124
+ schemaRef: {
1125
+ id: string;
1126
+ path: string;
1127
+ contentHash: string;
1128
+ sizeBytes: number;
1129
+ mediaType: string;
1130
+ visibility: "host_private" | "policy" | "verifier";
1131
+ } | null;
1132
+ submissionMode?: "structured_json" | undefined;
1133
+ maxBytes: number | null;
1134
+ metadata: Record<string, unknown>;
1135
+ }[] | undefined;
1136
+ provenance: {
1137
+ sourceRecordRef: string | null;
1138
+ mappingHash: string | null;
1139
+ };
1140
+ };
1141
+ supersedes: {
1142
+ id: string;
1143
+ contentHash: string;
1144
+ revision: number;
1145
+ } | null;
1146
+ correctionFeedbackId: string | null;
1147
+ receivedAt: string;
1148
+ contentHash: string;
1149
+ }[];
1150
+ decisions: {
1151
+ schemaVersion: "openpond.taskAdmissionDecision.v1";
1152
+ id: string;
1153
+ revision: number;
1154
+ evidence: {
1155
+ id: string;
1156
+ contentHash: string;
1157
+ revision: number;
1158
+ };
1159
+ supersedes: {
1160
+ id: string;
1161
+ contentHash: string;
1162
+ revision: number;
1163
+ } | null;
1164
+ actor: {
1165
+ kind: "human" | "qualified_policy";
1166
+ id: string;
1167
+ policy: {
1168
+ id: string;
1169
+ contentHash: string;
1170
+ revision: number;
1171
+ } | null;
1172
+ };
1173
+ evidenceValidity: "invalid" | "pending" | "valid";
1174
+ taskAdmissibility: "approved" | "pending" | "rejected";
1175
+ observedQuality: "failed" | "passed" | "unavailable" | "unscored";
1176
+ targetApproval: "approved" | "not_required" | "pending" | "rejected";
1177
+ approvedTarget: Record<string, import("zod").JSONType> | null;
1178
+ grade: {
1179
+ schemaVersion: "openpond.rewardComposition.v1";
1180
+ binding: {
1181
+ id: string;
1182
+ contentHash: string;
1183
+ revision: number;
1184
+ };
1185
+ taskHash: string;
1186
+ taskId?: string | undefined;
1187
+ outputHash: string;
1188
+ results: {
1189
+ graderId: string;
1190
+ reward: {
1191
+ id: string;
1192
+ contentHash: string;
1193
+ revision: number;
1194
+ };
1195
+ role: "evaluation" | "training";
1196
+ status: "failed" | "not_applicable" | "pending" | "scored" | "unavailable";
1197
+ rawScore: number | null;
1198
+ normalizedScore: number | null;
1199
+ passed: boolean | null;
1200
+ evidenceHashes: string[];
1201
+ graderEvidence?: {
1202
+ schemaVersion: "openpond.graderEvidence.v1";
1203
+ graderId: string;
1204
+ graderVersion: string;
1205
+ status?: "failed" | "pending" | "scored" | "unavailable" | undefined;
1206
+ score: number | null;
1207
+ passed: boolean;
1208
+ rewardEligible: boolean;
1209
+ failureClass: "cancelled" | "environment_failure" | "grader_failure" | "infrastructure_failure" | "policy_failure" | "timeout" | null;
1210
+ feedback: string[];
1211
+ visibleEvidenceRefs: string[];
1212
+ privilegedEvidenceRefs: string[];
1213
+ modelJudgeReceipt?: {
1214
+ schemaVersion: "openpond.modelJudgeReceipt.v1";
1215
+ providerId: string;
1216
+ modelId: string;
1217
+ modelRevision: string | null;
1218
+ responseId: string | null;
1219
+ requestHash: string;
1220
+ responseHash: string;
1221
+ inputTokens: number | null;
1222
+ outputTokens: number | null;
1223
+ costUsd: number | null;
1224
+ transportPolicy?: {
1225
+ schemaVersion: "openpond.modelJudgeTransportPolicy.v1";
1226
+ responseFormat: "judgment_json_schema";
1227
+ reasoningEffort: "high" | "low" | "max" | "medium" | "off" | "xhigh";
1228
+ maximumOutputTokens: number;
1229
+ } | undefined;
1230
+ effectiveRequestHash?: string | undefined;
1231
+ finishReason?: string | null | undefined;
1232
+ reasoningTokens?: number | null | undefined;
1233
+ } | undefined;
1234
+ contentHash: string;
1235
+ } | undefined;
1236
+ message: string | null;
1237
+ }[];
1238
+ training: {
1239
+ status: "not_configured" | "pending" | "scored" | "unscorable";
1240
+ score: number | null;
1241
+ passed: boolean | null;
1242
+ };
1243
+ evaluation: {
1244
+ status: "not_configured" | "pending" | "scored" | "unscorable";
1245
+ score: number | null;
1246
+ passed: boolean | null;
1247
+ };
1248
+ contentHash: string;
1249
+ } | null;
1250
+ targetGrade: {
1251
+ schemaVersion: "openpond.rewardComposition.v1";
1252
+ binding: {
1253
+ id: string;
1254
+ contentHash: string;
1255
+ revision: number;
1256
+ };
1257
+ taskHash: string;
1258
+ taskId?: string | undefined;
1259
+ outputHash: string;
1260
+ results: {
1261
+ graderId: string;
1262
+ reward: {
1263
+ id: string;
1264
+ contentHash: string;
1265
+ revision: number;
1266
+ };
1267
+ role: "evaluation" | "training";
1268
+ status: "failed" | "not_applicable" | "pending" | "scored" | "unavailable";
1269
+ rawScore: number | null;
1270
+ normalizedScore: number | null;
1271
+ passed: boolean | null;
1272
+ evidenceHashes: string[];
1273
+ graderEvidence?: {
1274
+ schemaVersion: "openpond.graderEvidence.v1";
1275
+ graderId: string;
1276
+ graderVersion: string;
1277
+ status?: "failed" | "pending" | "scored" | "unavailable" | undefined;
1278
+ score: number | null;
1279
+ passed: boolean;
1280
+ rewardEligible: boolean;
1281
+ failureClass: "cancelled" | "environment_failure" | "grader_failure" | "infrastructure_failure" | "policy_failure" | "timeout" | null;
1282
+ feedback: string[];
1283
+ visibleEvidenceRefs: string[];
1284
+ privilegedEvidenceRefs: string[];
1285
+ modelJudgeReceipt?: {
1286
+ schemaVersion: "openpond.modelJudgeReceipt.v1";
1287
+ providerId: string;
1288
+ modelId: string;
1289
+ modelRevision: string | null;
1290
+ responseId: string | null;
1291
+ requestHash: string;
1292
+ responseHash: string;
1293
+ inputTokens: number | null;
1294
+ outputTokens: number | null;
1295
+ costUsd: number | null;
1296
+ transportPolicy?: {
1297
+ schemaVersion: "openpond.modelJudgeTransportPolicy.v1";
1298
+ responseFormat: "judgment_json_schema";
1299
+ reasoningEffort: "high" | "low" | "max" | "medium" | "off" | "xhigh";
1300
+ maximumOutputTokens: number;
1301
+ } | undefined;
1302
+ effectiveRequestHash?: string | undefined;
1303
+ finishReason?: string | null | undefined;
1304
+ reasoningTokens?: number | null | undefined;
1305
+ } | undefined;
1306
+ contentHash: string;
1307
+ } | undefined;
1308
+ message: string | null;
1309
+ }[];
1310
+ training: {
1311
+ status: "not_configured" | "pending" | "scored" | "unscorable";
1312
+ score: number | null;
1313
+ passed: boolean | null;
1314
+ };
1315
+ evaluation: {
1316
+ status: "not_configured" | "pending" | "scored" | "unscorable";
1317
+ score: number | null;
1318
+ passed: boolean | null;
1319
+ };
1320
+ contentHash: string;
1321
+ } | null;
1322
+ note: string;
1323
+ decidedAt: string;
1324
+ contentHash: string;
1325
+ }[];
1326
+ sources: {
1327
+ schemaVersion: "openpond.learningSource.v1";
1328
+ id: string;
1329
+ revision: number;
1330
+ name: string;
1331
+ kind: "benchmark" | "direct" | "opentelemetry" | "provider" | "work";
1332
+ taskDefinition: {
1333
+ id: string;
1334
+ contentHash: string;
1335
+ revision: number;
1336
+ };
1337
+ enabled: boolean;
1338
+ allowedSplits: ("frozen_eval" | "test" | "train" | "validation")[];
1339
+ mapping: {
1340
+ schemaVersion: "openpond.taskSourceMapping.v1";
1341
+ exampleId: string;
1342
+ attemptId: string;
1343
+ occurredAt: string;
1344
+ input: string;
1345
+ observedOutput: string | null;
1346
+ expected: string | null;
1347
+ evaluatorContext: string | null;
1348
+ familyKey: string;
1349
+ split: "frozen_eval" | "test" | "train" | "validation";
1350
+ } | null;
1351
+ adapterVersion: string | null;
1352
+ reviewOrigin?: {
1353
+ modelId: string;
1354
+ packageHash: string;
1355
+ taskset: {
1356
+ id: string;
1357
+ contentHash: string;
1358
+ revision: number;
1359
+ };
1360
+ batch: {
1361
+ id: string;
1362
+ contentHash: string;
1363
+ revision: number;
1364
+ };
1365
+ rewardBinding: {
1366
+ id: string;
1367
+ contentHash: string;
1368
+ revision: number;
1369
+ };
1370
+ } | undefined;
1371
+ datasetOrigin?: {
1372
+ receiptId: string;
1373
+ packageHash: string;
1374
+ taskset: {
1375
+ id: string;
1376
+ contentHash: string;
1377
+ revision: number;
1378
+ };
1379
+ } | undefined;
1380
+ capturedOrigin?: {
1381
+ receiptId: string;
1382
+ ownerUserId: string;
1383
+ modelId: string;
1384
+ packageHash: string;
1385
+ taskset: {
1386
+ id: string;
1387
+ contentHash: string;
1388
+ revision: number;
1389
+ };
1390
+ } | undefined;
1391
+ contentHash: string;
1392
+ }[];
1393
+ assets: {
1394
+ schemaVersion: "openpond.learningTextAsset.v1";
1395
+ id: string;
1396
+ revision: 1;
1397
+ asset: {
1398
+ id: string;
1399
+ path: string;
1400
+ contentHash: string;
1401
+ sizeBytes: number;
1402
+ mediaType: string;
1403
+ visibility: "host_private" | "policy" | "verifier";
1404
+ };
1405
+ text: string;
1406
+ contentHash: string;
1407
+ }[];
1408
+ } | undefined;
1409
+ contentHash: string;
1410
+ };
1411
+ panel: {
1412
+ source: {
1413
+ id: string;
1414
+ contentHash: string;
1415
+ revision: number;
1416
+ };
1417
+ prompts: {
1418
+ id: string;
1419
+ text: string;
1420
+ system?: string | undefined;
1421
+ }[];
1422
+ } | {
1423
+ id: string;
1424
+ source: {
1425
+ id: string;
1426
+ contentHash: string;
1427
+ revision: number;
1428
+ } | null;
1429
+ role: "development" | "preservation";
1430
+ prompts: {
1431
+ id: string;
1432
+ text: string;
1433
+ system?: string | undefined;
1434
+ }[];
1435
+ };
1436
+ };
1437
+ //# sourceMappingURL=weight-space-evaluation.d.ts.map