@openpond/evals 0.4.0 → 0.4.2

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.
package/dist/index.js CHANGED
@@ -1,4 +1,5 @@
1
1
  export * from "./compatibility.js";
2
+ export * from "./benchmarks.js";
2
3
  export * from "./evidence/index.js";
3
4
  export * from "./graders.js";
4
5
  export * from "./harness.js";
@@ -0,0 +1,293 @@
1
+ import { z } from "zod";
2
+ import { type AttemptReceipt, type EvaluationResult } from "./runs.js";
3
+ export { harnessRefinerBenchmarkAssets, harnessRefinerBenchmarkRelease, } from "./builtin-benchmarks/harness-refiner.js";
4
+ export declare const BenchmarkMetricSchema: z.ZodEnum<{
5
+ foreground_tokens: "foreground_tokens";
6
+ success_rate: "success_rate";
7
+ latency_ms: "latency_ms";
8
+ cost_usd: "cost_usd";
9
+ }>;
10
+ export declare const BenchmarkRunPhaseSchema: z.ZodEnum<{
11
+ baseline: "baseline";
12
+ candidate: "candidate";
13
+ }>;
14
+ export declare const BenchmarkProtocolSchema: z.ZodObject<{
15
+ split: z.ZodEnum<{
16
+ train: "train";
17
+ validation: "validation";
18
+ test: "test";
19
+ frozen_eval: "frozen_eval";
20
+ }>;
21
+ taskIds: z.ZodArray<z.ZodString>;
22
+ seeds: z.ZodArray<z.ZodString>;
23
+ repetitions: z.ZodNumber;
24
+ runtimeTargetHash: z.ZodString;
25
+ environmentHash: z.ZodString;
26
+ toolContractHash: z.ZodString;
27
+ limitsHash: z.ZodString;
28
+ }, z.core.$strict>;
29
+ export declare const BenchmarkDefinitionSchema: z.ZodObject<{
30
+ schemaVersion: z.ZodLiteral<"openpond.benchmarkDefinition.v1">;
31
+ id: z.ZodString;
32
+ title: z.ZodString;
33
+ description: z.ZodString;
34
+ tasksetRelease: z.ZodObject<{
35
+ id: z.ZodString;
36
+ contentHash: z.ZodString;
37
+ }, z.core.$strict>;
38
+ adaptationSplit: z.ZodEnum<{
39
+ train: "train";
40
+ validation: "validation";
41
+ test: "test";
42
+ frozen_eval: "frozen_eval";
43
+ }>;
44
+ evaluationSplit: z.ZodEnum<{
45
+ train: "train";
46
+ validation: "validation";
47
+ test: "test";
48
+ frozen_eval: "frozen_eval";
49
+ }>;
50
+ primaryMetric: z.ZodEnum<{
51
+ foreground_tokens: "foreground_tokens";
52
+ success_rate: "success_rate";
53
+ latency_ms: "latency_ms";
54
+ cost_usd: "cost_usd";
55
+ }>;
56
+ qualityGate: z.ZodEnum<{
57
+ none: "none";
58
+ non_regression: "non_regression";
59
+ all_pass: "all_pass";
60
+ }>;
61
+ caseCounts: z.ZodObject<{
62
+ adaptation: z.ZodNumber;
63
+ evaluation: z.ZodNumber;
64
+ }, z.core.$strict>;
65
+ metadata: z.ZodDefault<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
66
+ }, z.core.$strict>;
67
+ export declare const BenchmarkRunRequestSchema: z.ZodObject<{
68
+ schemaVersion: z.ZodLiteral<"openpond.benchmarkRunRequest.v1">;
69
+ phase: z.ZodEnum<{
70
+ baseline: "baseline";
71
+ candidate: "candidate";
72
+ }>;
73
+ model: z.ZodObject<{
74
+ provider: z.ZodString;
75
+ model: z.ZodString;
76
+ revision: z.ZodDefault<z.ZodNullable<z.ZodString>>;
77
+ artifactHash: z.ZodDefault<z.ZodNullable<z.ZodString>>;
78
+ tokenizerRevision: z.ZodDefault<z.ZodNullable<z.ZodString>>;
79
+ chatTemplateHash: z.ZodDefault<z.ZodNullable<z.ZodString>>;
80
+ }, z.core.$strict>;
81
+ reasoningEffort: z.ZodNullable<z.ZodString>;
82
+ split: z.ZodEnum<{
83
+ train: "train";
84
+ validation: "validation";
85
+ test: "test";
86
+ frozen_eval: "frozen_eval";
87
+ }>;
88
+ seeds: z.ZodArray<z.ZodString>;
89
+ repetitions: z.ZodNumber;
90
+ metadata: z.ZodDefault<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
91
+ }, z.core.$strict>;
92
+ export declare const BenchmarkRunSummaryContentSchema: z.ZodObject<{
93
+ schemaVersion: z.ZodLiteral<"openpond.benchmarkRunSummary.v1">;
94
+ id: z.ZodString;
95
+ phase: z.ZodEnum<{
96
+ baseline: "baseline";
97
+ candidate: "candidate";
98
+ }>;
99
+ tasksetRelease: z.ZodObject<{
100
+ id: z.ZodString;
101
+ contentHash: z.ZodString;
102
+ }, z.core.$strict>;
103
+ harnessRelease: z.ZodObject<{
104
+ id: z.ZodString;
105
+ contentHash: z.ZodString;
106
+ }, z.core.$strict>;
107
+ evaluationResult: z.ZodObject<{
108
+ id: z.ZodString;
109
+ contentHash: z.ZodString;
110
+ }, z.core.$strict>;
111
+ model: z.ZodObject<{
112
+ provider: z.ZodString;
113
+ model: z.ZodString;
114
+ revision: z.ZodDefault<z.ZodNullable<z.ZodString>>;
115
+ artifactHash: z.ZodDefault<z.ZodNullable<z.ZodString>>;
116
+ tokenizerRevision: z.ZodDefault<z.ZodNullable<z.ZodString>>;
117
+ chatTemplateHash: z.ZodDefault<z.ZodNullable<z.ZodString>>;
118
+ }, z.core.$strict>;
119
+ reasoningEffort: z.ZodNullable<z.ZodString>;
120
+ protocol: z.ZodObject<{
121
+ split: z.ZodEnum<{
122
+ train: "train";
123
+ validation: "validation";
124
+ test: "test";
125
+ frozen_eval: "frozen_eval";
126
+ }>;
127
+ taskIds: z.ZodArray<z.ZodString>;
128
+ seeds: z.ZodArray<z.ZodString>;
129
+ repetitions: z.ZodNumber;
130
+ runtimeTargetHash: z.ZodString;
131
+ environmentHash: z.ZodString;
132
+ toolContractHash: z.ZodString;
133
+ limitsHash: z.ZodString;
134
+ }, z.core.$strict>;
135
+ attemptCount: z.ZodNumber;
136
+ passedCount: z.ZodNumber;
137
+ terminalCount: z.ZodNumber;
138
+ usage: z.ZodObject<{
139
+ inputTokens: z.ZodNumber;
140
+ outputTokens: z.ZodNumber;
141
+ totalTokens: z.ZodNumber;
142
+ }, z.core.$strict>;
143
+ costUsd: z.ZodNullable<z.ZodNumber>;
144
+ latencyMs: z.ZodNumber;
145
+ createdAt: z.ZodString;
146
+ metadata: z.ZodDefault<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
147
+ }, z.core.$strict>;
148
+ export declare const BenchmarkRunSummarySchema: z.ZodObject<{
149
+ schemaVersion: z.ZodLiteral<"openpond.benchmarkRunSummary.v1">;
150
+ id: z.ZodString;
151
+ phase: z.ZodEnum<{
152
+ baseline: "baseline";
153
+ candidate: "candidate";
154
+ }>;
155
+ tasksetRelease: z.ZodObject<{
156
+ id: z.ZodString;
157
+ contentHash: z.ZodString;
158
+ }, z.core.$strict>;
159
+ harnessRelease: z.ZodObject<{
160
+ id: z.ZodString;
161
+ contentHash: z.ZodString;
162
+ }, z.core.$strict>;
163
+ evaluationResult: z.ZodObject<{
164
+ id: z.ZodString;
165
+ contentHash: z.ZodString;
166
+ }, z.core.$strict>;
167
+ model: z.ZodObject<{
168
+ provider: z.ZodString;
169
+ model: z.ZodString;
170
+ revision: z.ZodDefault<z.ZodNullable<z.ZodString>>;
171
+ artifactHash: z.ZodDefault<z.ZodNullable<z.ZodString>>;
172
+ tokenizerRevision: z.ZodDefault<z.ZodNullable<z.ZodString>>;
173
+ chatTemplateHash: z.ZodDefault<z.ZodNullable<z.ZodString>>;
174
+ }, z.core.$strict>;
175
+ reasoningEffort: z.ZodNullable<z.ZodString>;
176
+ protocol: z.ZodObject<{
177
+ split: z.ZodEnum<{
178
+ train: "train";
179
+ validation: "validation";
180
+ test: "test";
181
+ frozen_eval: "frozen_eval";
182
+ }>;
183
+ taskIds: z.ZodArray<z.ZodString>;
184
+ seeds: z.ZodArray<z.ZodString>;
185
+ repetitions: z.ZodNumber;
186
+ runtimeTargetHash: z.ZodString;
187
+ environmentHash: z.ZodString;
188
+ toolContractHash: z.ZodString;
189
+ limitsHash: z.ZodString;
190
+ }, z.core.$strict>;
191
+ attemptCount: z.ZodNumber;
192
+ passedCount: z.ZodNumber;
193
+ terminalCount: z.ZodNumber;
194
+ usage: z.ZodObject<{
195
+ inputTokens: z.ZodNumber;
196
+ outputTokens: z.ZodNumber;
197
+ totalTokens: z.ZodNumber;
198
+ }, z.core.$strict>;
199
+ costUsd: z.ZodNullable<z.ZodNumber>;
200
+ latencyMs: z.ZodNumber;
201
+ createdAt: z.ZodString;
202
+ metadata: z.ZodDefault<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
203
+ contentHash: z.ZodString;
204
+ }, z.core.$strict>;
205
+ export declare const BenchmarkComparisonContentSchema: z.ZodObject<{
206
+ schemaVersion: z.ZodLiteral<"openpond.benchmarkComparison.v1">;
207
+ id: z.ZodString;
208
+ baseline: z.ZodObject<{
209
+ id: z.ZodString;
210
+ contentHash: z.ZodString;
211
+ }, z.core.$strict>;
212
+ candidate: z.ZodObject<{
213
+ id: z.ZodString;
214
+ contentHash: z.ZodString;
215
+ }, z.core.$strict>;
216
+ tasksetRelease: z.ZodObject<{
217
+ id: z.ZodString;
218
+ contentHash: z.ZodString;
219
+ }, z.core.$strict>;
220
+ primaryMetric: z.ZodEnum<{
221
+ foreground_tokens: "foreground_tokens";
222
+ success_rate: "success_rate";
223
+ latency_ms: "latency_ms";
224
+ cost_usd: "cost_usd";
225
+ }>;
226
+ qualityPassed: z.ZodBoolean;
227
+ baselinePassRate: z.ZodNumber;
228
+ candidatePassRate: z.ZodNumber;
229
+ foregroundTokenDelta: z.ZodNumber;
230
+ foregroundTokenDeltaPercent: z.ZodNullable<z.ZodNumber>;
231
+ improved: z.ZodBoolean;
232
+ createdAt: z.ZodString;
233
+ metadata: z.ZodDefault<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
234
+ }, z.core.$strict>;
235
+ export declare const BenchmarkComparisonSchema: z.ZodObject<{
236
+ schemaVersion: z.ZodLiteral<"openpond.benchmarkComparison.v1">;
237
+ id: z.ZodString;
238
+ baseline: z.ZodObject<{
239
+ id: z.ZodString;
240
+ contentHash: z.ZodString;
241
+ }, z.core.$strict>;
242
+ candidate: z.ZodObject<{
243
+ id: z.ZodString;
244
+ contentHash: z.ZodString;
245
+ }, z.core.$strict>;
246
+ tasksetRelease: z.ZodObject<{
247
+ id: z.ZodString;
248
+ contentHash: z.ZodString;
249
+ }, z.core.$strict>;
250
+ primaryMetric: z.ZodEnum<{
251
+ foreground_tokens: "foreground_tokens";
252
+ success_rate: "success_rate";
253
+ latency_ms: "latency_ms";
254
+ cost_usd: "cost_usd";
255
+ }>;
256
+ qualityPassed: z.ZodBoolean;
257
+ baselinePassRate: z.ZodNumber;
258
+ candidatePassRate: z.ZodNumber;
259
+ foregroundTokenDelta: z.ZodNumber;
260
+ foregroundTokenDeltaPercent: z.ZodNullable<z.ZodNumber>;
261
+ improved: z.ZodBoolean;
262
+ createdAt: z.ZodString;
263
+ metadata: z.ZodDefault<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
264
+ contentHash: z.ZodString;
265
+ }, z.core.$strict>;
266
+ export declare function createBenchmarkDefinition(input: z.input<typeof BenchmarkDefinitionSchema>): BenchmarkDefinition;
267
+ export declare function createBenchmarkRunSummary(input: {
268
+ id: string;
269
+ phase: BenchmarkRunPhase;
270
+ evaluation: EvaluationResult;
271
+ receipts: AttemptReceipt[];
272
+ reasoningEffort: string | null;
273
+ protocol: BenchmarkProtocol;
274
+ createdAt: string;
275
+ metadata?: Record<string, unknown>;
276
+ }): BenchmarkRunSummary;
277
+ export declare function compareBenchmarkRuns(input: {
278
+ id: string;
279
+ baseline: BenchmarkRunSummary;
280
+ candidate: BenchmarkRunSummary;
281
+ primaryMetric: BenchmarkMetric;
282
+ qualityGate: "none" | "non_regression" | "all_pass";
283
+ createdAt: string;
284
+ metadata?: Record<string, unknown>;
285
+ }): BenchmarkComparison;
286
+ export type BenchmarkMetric = z.infer<typeof BenchmarkMetricSchema>;
287
+ export type BenchmarkRunPhase = z.infer<typeof BenchmarkRunPhaseSchema>;
288
+ export type BenchmarkProtocol = z.infer<typeof BenchmarkProtocolSchema>;
289
+ export type BenchmarkDefinition = z.infer<typeof BenchmarkDefinitionSchema>;
290
+ export type BenchmarkRunRequest = z.infer<typeof BenchmarkRunRequestSchema>;
291
+ export type BenchmarkRunSummary = z.infer<typeof BenchmarkRunSummarySchema>;
292
+ export type BenchmarkComparison = z.infer<typeof BenchmarkComparisonSchema>;
293
+ //# sourceMappingURL=benchmarks.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"benchmarks.d.ts","sourceRoot":"","sources":["../../../../src/benchmarks.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAYxB,OAAO,EAAE,KAAK,cAAc,EAAE,KAAK,gBAAgB,EAAE,MAAM,WAAW,CAAC;AAGvE,OAAO,EACL,6BAA6B,EAC7B,8BAA8B,GAC/B,MAAM,yCAAyC,CAAC;AAEjD,eAAO,MAAM,qBAAqB;;;;;EAKhC,CAAC;AACH,eAAO,MAAM,uBAAuB;;;EAAoC,CAAC;AAEzE,eAAO,MAAM,uBAAuB;;;;;;;;;;;;;;kBASzB,CAAC;AAEZ,eAAO,MAAM,yBAAyB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;kBAe3B,CAAC;AAEZ,eAAO,MAAM,yBAAyB;;;;;;;;;;;;;;;;;;;;;;;;kBAS3B,CAAC;AAQZ,eAAO,MAAM,gCAAgC;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;kBAkBlC,CAAC;AACZ,eAAO,MAAM,yBAAyB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;kBAE3B,CAAC;AAEZ,eAAO,MAAM,gCAAgC;;;;;;;;;;;;;;;;;;;;;;;;;;;;;kBAelC,CAAC;AACZ,eAAO,MAAM,yBAAyB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;kBAE3B,CAAC;AAEZ,wBAAgB,yBAAyB,CACvC,KAAK,EAAE,CAAC,CAAC,KAAK,CAAC,OAAO,yBAAyB,CAAC,GAC/C,mBAAmB,CAErB;AAED,wBAAgB,yBAAyB,CAAC,KAAK,EAAE;IAC/C,EAAE,EAAE,MAAM,CAAC;IACX,KAAK,EAAE,iBAAiB,CAAC;IACzB,UAAU,EAAE,gBAAgB,CAAC;IAC7B,QAAQ,EAAE,cAAc,EAAE,CAAC;IAC3B,eAAe,EAAE,MAAM,GAAG,IAAI,CAAC;IAC/B,QAAQ,EAAE,iBAAiB,CAAC;IAC5B,SAAS,EAAE,MAAM,CAAC;IAClB,QAAQ,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;CACpC,GAAG,mBAAmB,CAuCtB;AAED,wBAAgB,oBAAoB,CAAC,KAAK,EAAE;IAC1C,EAAE,EAAE,MAAM,CAAC;IACX,QAAQ,EAAE,mBAAmB,CAAC;IAC9B,SAAS,EAAE,mBAAmB,CAAC;IAC/B,aAAa,EAAE,eAAe,CAAC;IAC/B,WAAW,EAAE,MAAM,GAAG,gBAAgB,GAAG,UAAU,CAAC;IACpD,SAAS,EAAE,MAAM,CAAC;IAClB,QAAQ,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;CACpC,GAAG,mBAAmB,CAuDtB;AAkDD,MAAM,MAAM,eAAe,GAAG,CAAC,CAAC,KAAK,CAAC,OAAO,qBAAqB,CAAC,CAAC;AACpE,MAAM,MAAM,iBAAiB,GAAG,CAAC,CAAC,KAAK,CAAC,OAAO,uBAAuB,CAAC,CAAC;AACxE,MAAM,MAAM,iBAAiB,GAAG,CAAC,CAAC,KAAK,CAAC,OAAO,uBAAuB,CAAC,CAAC;AACxE,MAAM,MAAM,mBAAmB,GAAG,CAAC,CAAC,KAAK,CAAC,OAAO,yBAAyB,CAAC,CAAC;AAC5E,MAAM,MAAM,mBAAmB,GAAG,CAAC,CAAC,KAAK,CAAC,OAAO,yBAAyB,CAAC,CAAC;AAC5E,MAAM,MAAM,mBAAmB,GAAG,CAAC,CAAC,KAAK,CAAC,OAAO,yBAAyB,CAAC,CAAC;AAC5E,MAAM,MAAM,mBAAmB,GAAG,CAAC,CAAC,KAAK,CAAC,OAAO,yBAAyB,CAAC,CAAC"}
@@ -0,0 +1,119 @@
1
+ export declare const harnessRefinerBenchmarkRelease: {
2
+ schemaVersion: "openpond.tasksetRelease.v2";
3
+ id: string;
4
+ revision: number;
5
+ policy: {
6
+ policyVisibleFields: string[];
7
+ privilegedFields: string[];
8
+ hiddenGraderRefs: string[];
9
+ connectedAppScopes: string[];
10
+ };
11
+ environment: {
12
+ protocolVersion: "openpond.environment.v1";
13
+ kind: "text" | "agent" | "work" | "custom_program";
14
+ entrypoint: string;
15
+ stateful: boolean;
16
+ deterministicSeeds: boolean;
17
+ lifecycle: ("create" | "reset" | "step" | "collect" | "destroy")[];
18
+ networkPolicy: "none" | "declared_read_only" | "declared_scoped";
19
+ defaultTimeoutMs: number;
20
+ };
21
+ tools: {
22
+ name: string;
23
+ description: string;
24
+ inputSchema: Record<string, unknown>;
25
+ inputSchemaHash: string;
26
+ sideEffect: "read" | "write";
27
+ timeoutMs: number;
28
+ }[];
29
+ capabilities: {
30
+ id: string;
31
+ required: boolean;
32
+ scopes: string[];
33
+ portability: "portable" | "host_adapter" | "local_only" | "hosted_only";
34
+ }[];
35
+ tasks: {
36
+ id: string;
37
+ clusterKey: string;
38
+ split: "train" | "validation" | "test" | "frozen_eval";
39
+ input: Record<string, unknown>;
40
+ expectedOutput: Record<string, unknown> | null;
41
+ policyVisibleContext: Record<string, unknown>;
42
+ privilegedContextRef: string | null;
43
+ artifactRefs: {
44
+ id: string;
45
+ path: string;
46
+ contentHash: string;
47
+ sizeBytes: number;
48
+ mediaType: string;
49
+ visibility: "policy" | "verifier" | "host_private";
50
+ }[];
51
+ tags: string[];
52
+ }[];
53
+ graders: ({
54
+ id: string;
55
+ version: string;
56
+ weight: number;
57
+ hardGate: boolean;
58
+ rewardEligible: boolean;
59
+ privileged: boolean;
60
+ kind: "content" | "schema" | "artifact" | "runtime_event" | "state";
61
+ config: Record<string, unknown>;
62
+ } | {
63
+ id: string;
64
+ version: string;
65
+ weight: number;
66
+ hardGate: boolean;
67
+ rewardEligible: boolean;
68
+ privileged: boolean;
69
+ kind: "model_judge";
70
+ rubricRef: {
71
+ id: string;
72
+ path: string;
73
+ contentHash: string;
74
+ sizeBytes: number;
75
+ mediaType: string;
76
+ visibility: "policy" | "verifier" | "host_private";
77
+ };
78
+ calibrationStatus: "pending" | "passed" | "failed";
79
+ } | {
80
+ id: string;
81
+ version: string;
82
+ weight: number;
83
+ hardGate: boolean;
84
+ rewardEligible: boolean;
85
+ privileged: boolean;
86
+ kind: "custom_verifier";
87
+ verifierRef: {
88
+ id: string;
89
+ path: string;
90
+ contentHash: string;
91
+ sizeBytes: number;
92
+ mediaType: string;
93
+ visibility: "policy" | "verifier" | "host_private";
94
+ };
95
+ timeoutMs: number;
96
+ networkPolicy: "none";
97
+ } | {
98
+ id: string;
99
+ version: string;
100
+ weight: number;
101
+ hardGate: boolean;
102
+ rewardEligible: boolean;
103
+ privileged: boolean;
104
+ kind: "human";
105
+ rubricRef: {
106
+ id: string;
107
+ path: string;
108
+ contentHash: string;
109
+ sizeBytes: number;
110
+ mediaType: string;
111
+ visibility: "policy" | "verifier" | "host_private";
112
+ };
113
+ reviewerRole: string;
114
+ })[];
115
+ metadata: Record<string, unknown>;
116
+ contentHash: string;
117
+ };
118
+ export declare const harnessRefinerBenchmarkAssets: Readonly<Record<string, string>>;
119
+ //# sourceMappingURL=harness-refiner.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"harness-refiner.d.ts","sourceRoot":"","sources":["../../../../../src/builtin-benchmarks/harness-refiner.ts"],"names":[],"mappings":"AAGA,eAAO,MAAM,8BAA8B;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CAwuC1C,CAAC;AAEF,eAAO,MAAM,6BAA6B,EAAE,QAAQ,CAAC,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CASzE,CAAC"}
@@ -23,8 +23,8 @@ export declare const completeWorkProcessTraceFixture: {
23
23
  id: string;
24
24
  }[];
25
25
  attributes: {
26
- toolCategory: "agent" | "filesystem" | "source_control" | "command" | "browser" | "connected_app" | "sandbox" | "other" | null;
27
- validationKind: "test" | "other" | "structural" | "visual" | "user_review" | null;
26
+ toolCategory: "agent" | "command" | "filesystem" | "source_control" | "browser" | "connected_app" | "sandbox" | "other" | null;
27
+ validationKind: "test" | "structural" | "visual" | "user_review" | "other" | null;
28
28
  transitionState: "timeout" | "cancelled" | "failed" | "completed" | "running" | null;
29
29
  interventionOutcome: "requested" | "approved" | "denied" | "answered" | "dismissed" | null;
30
30
  artifactCount: number;
@@ -200,8 +200,8 @@ export declare const incompleteWorkProcessTraceFixture: {
200
200
  id: string;
201
201
  }[];
202
202
  attributes: {
203
- toolCategory: "agent" | "filesystem" | "source_control" | "command" | "browser" | "connected_app" | "sandbox" | "other" | null;
204
- validationKind: "test" | "other" | "structural" | "visual" | "user_review" | null;
203
+ toolCategory: "agent" | "command" | "filesystem" | "source_control" | "browser" | "connected_app" | "sandbox" | "other" | null;
204
+ validationKind: "test" | "structural" | "visual" | "user_review" | "other" | null;
205
205
  transitionState: "timeout" | "cancelled" | "failed" | "completed" | "running" | null;
206
206
  interventionOutcome: "requested" | "approved" | "denied" | "answered" | "dismissed" | null;
207
207
  artifactCount: number;
@@ -406,8 +406,8 @@ export declare const workEvidenceConformance: {
406
406
  id: string;
407
407
  }[];
408
408
  attributes: {
409
- toolCategory: "agent" | "filesystem" | "source_control" | "command" | "browser" | "connected_app" | "sandbox" | "other" | null;
410
- validationKind: "test" | "other" | "structural" | "visual" | "user_review" | null;
409
+ toolCategory: "agent" | "command" | "filesystem" | "source_control" | "browser" | "connected_app" | "sandbox" | "other" | null;
410
+ validationKind: "test" | "structural" | "visual" | "user_review" | "other" | null;
411
411
  transitionState: "timeout" | "cancelled" | "failed" | "completed" | "running" | null;
412
412
  interventionOutcome: "requested" | "approved" | "denied" | "answered" | "dismissed" | null;
413
413
  artifactCount: number;
@@ -639,8 +639,8 @@ export declare const workEvidenceConformance: {
639
639
  id: string;
640
640
  }[];
641
641
  attributes: {
642
- toolCategory: "agent" | "filesystem" | "source_control" | "command" | "browser" | "connected_app" | "sandbox" | "other" | null;
643
- validationKind: "test" | "other" | "structural" | "visual" | "user_review" | null;
642
+ toolCategory: "agent" | "command" | "filesystem" | "source_control" | "browser" | "connected_app" | "sandbox" | "other" | null;
643
+ validationKind: "test" | "structural" | "visual" | "user_review" | "other" | null;
644
644
  transitionState: "timeout" | "cancelled" | "failed" | "completed" | "running" | null;
645
645
  interventionOutcome: "requested" | "approved" | "denied" | "answered" | "dismissed" | null;
646
646
  artifactCount: number;
@@ -57,9 +57,9 @@ export declare const WorkProcessStepStatusSchema: z.ZodEnum<{
57
57
  }>;
58
58
  export declare const WorkToolCategorySchema: z.ZodEnum<{
59
59
  agent: "agent";
60
+ command: "command";
60
61
  filesystem: "filesystem";
61
62
  source_control: "source_control";
62
- command: "command";
63
63
  browser: "browser";
64
64
  connected_app: "connected_app";
65
65
  sandbox: "sandbox";
@@ -67,10 +67,10 @@ export declare const WorkToolCategorySchema: z.ZodEnum<{
67
67
  }>;
68
68
  export declare const WorkValidationKindSchema: z.ZodEnum<{
69
69
  test: "test";
70
- other: "other";
71
70
  structural: "structural";
72
71
  visual: "visual";
73
72
  user_review: "user_review";
73
+ other: "other";
74
74
  }>;
75
75
  export declare const WorkTransitionStateSchema: z.ZodEnum<{
76
76
  timeout: "timeout";
@@ -162,9 +162,9 @@ export declare const WorkProcessStepSchema: z.ZodObject<{
162
162
  attributes: z.ZodObject<{
163
163
  toolCategory: z.ZodNullable<z.ZodEnum<{
164
164
  agent: "agent";
165
+ command: "command";
165
166
  filesystem: "filesystem";
166
167
  source_control: "source_control";
167
- command: "command";
168
168
  browser: "browser";
169
169
  connected_app: "connected_app";
170
170
  sandbox: "sandbox";
@@ -172,10 +172,10 @@ export declare const WorkProcessStepSchema: z.ZodObject<{
172
172
  }>>;
173
173
  validationKind: z.ZodNullable<z.ZodEnum<{
174
174
  test: "test";
175
- other: "other";
176
175
  structural: "structural";
177
176
  visual: "visual";
178
177
  user_review: "user_review";
178
+ other: "other";
179
179
  }>>;
180
180
  transitionState: z.ZodNullable<z.ZodEnum<{
181
181
  timeout: "timeout";
@@ -280,9 +280,9 @@ export declare const WorkProcessTraceContentSchema: z.ZodObject<{
280
280
  attributes: z.ZodObject<{
281
281
  toolCategory: z.ZodNullable<z.ZodEnum<{
282
282
  agent: "agent";
283
+ command: "command";
283
284
  filesystem: "filesystem";
284
285
  source_control: "source_control";
285
- command: "command";
286
286
  browser: "browser";
287
287
  connected_app: "connected_app";
288
288
  sandbox: "sandbox";
@@ -290,10 +290,10 @@ export declare const WorkProcessTraceContentSchema: z.ZodObject<{
290
290
  }>>;
291
291
  validationKind: z.ZodNullable<z.ZodEnum<{
292
292
  test: "test";
293
- other: "other";
294
293
  structural: "structural";
295
294
  visual: "visual";
296
295
  user_review: "user_review";
296
+ other: "other";
297
297
  }>>;
298
298
  transitionState: z.ZodNullable<z.ZodEnum<{
299
299
  timeout: "timeout";
@@ -399,9 +399,9 @@ export declare const WorkProcessTraceSchema: z.ZodObject<{
399
399
  attributes: z.ZodObject<{
400
400
  toolCategory: z.ZodNullable<z.ZodEnum<{
401
401
  agent: "agent";
402
+ command: "command";
402
403
  filesystem: "filesystem";
403
404
  source_control: "source_control";
404
- command: "command";
405
405
  browser: "browser";
406
406
  connected_app: "connected_app";
407
407
  sandbox: "sandbox";
@@ -409,10 +409,10 @@ export declare const WorkProcessTraceSchema: z.ZodObject<{
409
409
  }>>;
410
410
  validationKind: z.ZodNullable<z.ZodEnum<{
411
411
  test: "test";
412
- other: "other";
413
412
  structural: "structural";
414
413
  visual: "visual";
415
414
  user_review: "user_review";
415
+ other: "other";
416
416
  }>>;
417
417
  transitionState: z.ZodNullable<z.ZodEnum<{
418
418
  timeout: "timeout";
@@ -1,4 +1,5 @@
1
1
  export * from "./compatibility.js";
2
+ export * from "./benchmarks.js";
2
3
  export * from "./evidence/index.js";
3
4
  export * from "./graders.js";
4
5
  export * from "./harness.js";
@@ -1 +1 @@
1
- {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../../src/index.ts"],"names":[],"mappings":"AAAA,cAAc,oBAAoB,CAAC;AACnC,cAAc,qBAAqB,CAAC;AACpC,cAAc,cAAc,CAAC;AAC7B,cAAc,cAAc,CAAC;AAC7B,cAAc,WAAW,CAAC;AAC1B,cAAc,sCAAsC,CAAC;AACrD,cAAc,yBAAyB,CAAC;AACxC,cAAc,eAAe,CAAC"}
1
+ {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../../src/index.ts"],"names":[],"mappings":"AAAA,cAAc,oBAAoB,CAAC;AACnC,cAAc,iBAAiB,CAAC;AAChC,cAAc,qBAAqB,CAAC;AACpC,cAAc,cAAc,CAAC;AAC7B,cAAc,cAAc,CAAC;AAC7B,cAAc,WAAW,CAAC;AAC1B,cAAc,sCAAsC,CAAC;AACrD,cAAc,yBAAyB,CAAC;AACxC,cAAc,eAAe,CAAC"}
@@ -24,7 +24,7 @@ export declare const harnessEvaluationReviewConformance: {
24
24
  id: string;
25
25
  contentHash: string;
26
26
  };
27
- kind: "validation" | "observation" | "taskset" | "trigger" | "route_decision" | "refiner_outcome" | "proposal" | "apply_receipt" | "harness_advance" | "rollback" | "work_outcome" | "evaluation" | "training_qualification" | "model_candidate";
27
+ kind: "validation" | "evaluation" | "observation" | "taskset" | "trigger" | "route_decision" | "refiner_outcome" | "proposal" | "apply_receipt" | "harness_advance" | "rollback" | "work_outcome" | "training_qualification" | "model_candidate";
28
28
  sourceRef: string;
29
29
  sourcePolicy: {
30
30
  policy: {
@@ -61,7 +61,7 @@ export declare const harnessEvaluationReviewConformance: {
61
61
  } | null;
62
62
  classification: "runtime" | "taskset" | "no_action" | "harness_maintenance" | "product" | "model_improvement";
63
63
  triage: {
64
- layer: "model" | "tools" | "runtime" | "evaluation" | "product" | "harness" | "retrieval";
64
+ layer: "model" | "tools" | "evaluation" | "runtime" | "product" | "harness" | "retrieval";
65
65
  status: "resolved" | "not_applicable" | "unresolved" | "blocked";
66
66
  reason: string;
67
67
  evidenceRefs: {
@@ -114,7 +114,7 @@ export declare const harnessEvaluationReviewConformance: {
114
114
  id: string;
115
115
  contentHash: string;
116
116
  };
117
- kind: "validation" | "observation" | "taskset" | "trigger" | "route_decision" | "refiner_outcome" | "proposal" | "apply_receipt" | "harness_advance" | "rollback" | "work_outcome" | "evaluation" | "training_qualification" | "model_candidate";
117
+ kind: "validation" | "evaluation" | "observation" | "taskset" | "trigger" | "route_decision" | "refiner_outcome" | "proposal" | "apply_receipt" | "harness_advance" | "rollback" | "work_outcome" | "training_qualification" | "model_candidate";
118
118
  sourceRef: string;
119
119
  sourcePolicy: {
120
120
  policy: {
@@ -151,7 +151,7 @@ export declare const harnessEvaluationReviewConformance: {
151
151
  } | null;
152
152
  classification: "runtime" | "taskset" | "no_action" | "harness_maintenance" | "product" | "model_improvement";
153
153
  triage: {
154
- layer: "model" | "tools" | "runtime" | "evaluation" | "product" | "harness" | "retrieval";
154
+ layer: "model" | "tools" | "evaluation" | "runtime" | "product" | "harness" | "retrieval";
155
155
  status: "resolved" | "not_applicable" | "unresolved" | "blocked";
156
156
  reason: string;
157
157
  evidenceRefs: {
@@ -204,7 +204,7 @@ export declare const harnessEvaluationReviewConformance: {
204
204
  id: string;
205
205
  contentHash: string;
206
206
  };
207
- kind: "validation" | "observation" | "taskset" | "trigger" | "route_decision" | "refiner_outcome" | "proposal" | "apply_receipt" | "harness_advance" | "rollback" | "work_outcome" | "evaluation" | "training_qualification" | "model_candidate";
207
+ kind: "validation" | "evaluation" | "observation" | "taskset" | "trigger" | "route_decision" | "refiner_outcome" | "proposal" | "apply_receipt" | "harness_advance" | "rollback" | "work_outcome" | "training_qualification" | "model_candidate";
208
208
  sourceRef: string;
209
209
  sourcePolicy: {
210
210
  policy: {
@@ -241,7 +241,7 @@ export declare const harnessEvaluationReviewConformance: {
241
241
  } | null;
242
242
  classification: "runtime" | "taskset" | "no_action" | "harness_maintenance" | "product" | "model_improvement";
243
243
  triage: {
244
- layer: "model" | "tools" | "runtime" | "evaluation" | "product" | "harness" | "retrieval";
244
+ layer: "model" | "tools" | "evaluation" | "runtime" | "product" | "harness" | "retrieval";
245
245
  status: "resolved" | "not_applicable" | "unresolved" | "blocked";
246
246
  reason: string;
247
247
  evidenceRefs: {
@@ -294,7 +294,7 @@ export declare const harnessEvaluationReviewConformance: {
294
294
  id: string;
295
295
  contentHash: string;
296
296
  };
297
- kind: "validation" | "observation" | "taskset" | "trigger" | "route_decision" | "refiner_outcome" | "proposal" | "apply_receipt" | "harness_advance" | "rollback" | "work_outcome" | "evaluation" | "training_qualification" | "model_candidate";
297
+ kind: "validation" | "evaluation" | "observation" | "taskset" | "trigger" | "route_decision" | "refiner_outcome" | "proposal" | "apply_receipt" | "harness_advance" | "rollback" | "work_outcome" | "training_qualification" | "model_candidate";
298
298
  sourceRef: string;
299
299
  sourcePolicy: {
300
300
  policy: {
@@ -331,7 +331,7 @@ export declare const harnessEvaluationReviewConformance: {
331
331
  } | null;
332
332
  classification: "runtime" | "taskset" | "no_action" | "harness_maintenance" | "product" | "model_improvement";
333
333
  triage: {
334
- layer: "model" | "tools" | "runtime" | "evaluation" | "product" | "harness" | "retrieval";
334
+ layer: "model" | "tools" | "evaluation" | "runtime" | "product" | "harness" | "retrieval";
335
335
  status: "resolved" | "not_applicable" | "unresolved" | "blocked";
336
336
  reason: string;
337
337
  evidenceRefs: {
@@ -538,7 +538,7 @@ export declare const harnessEvaluationReviewConformance: {
538
538
  id: string;
539
539
  contentHash: string;
540
540
  };
541
- kind: "validation" | "observation" | "taskset" | "trigger" | "route_decision" | "refiner_outcome" | "proposal" | "apply_receipt" | "harness_advance" | "rollback" | "work_outcome" | "evaluation" | "training_qualification" | "model_candidate";
541
+ kind: "validation" | "evaluation" | "observation" | "taskset" | "trigger" | "route_decision" | "refiner_outcome" | "proposal" | "apply_receipt" | "harness_advance" | "rollback" | "work_outcome" | "training_qualification" | "model_candidate";
542
542
  sourceRef: string;
543
543
  sourcePolicy: {
544
544
  policy: {
@@ -575,7 +575,7 @@ export declare const harnessEvaluationReviewConformance: {
575
575
  } | null;
576
576
  classification: "runtime" | "taskset" | "no_action" | "harness_maintenance" | "product" | "model_improvement";
577
577
  triage: {
578
- layer: "model" | "tools" | "runtime" | "evaluation" | "product" | "harness" | "retrieval";
578
+ layer: "model" | "tools" | "evaluation" | "runtime" | "product" | "harness" | "retrieval";
579
579
  status: "resolved" | "not_applicable" | "unresolved" | "blocked";
580
580
  reason: string;
581
581
  evidenceRefs: {