@runtypelabs/sdk 11.2.0 → 11.3.1
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.cjs +82 -1
- package/dist/index.d.cts +557 -3
- package/dist/index.d.ts +557 -3
- package/dist/index.mjs +82 -2
- package/package.json +2 -2
package/dist/index.d.cts
CHANGED
|
@@ -1122,6 +1122,8 @@ interface paths {
|
|
|
1122
1122
|
timeoutMs?: number;
|
|
1123
1123
|
toolPool: string[];
|
|
1124
1124
|
};
|
|
1125
|
+
/** @enum {string} */
|
|
1126
|
+
intent?: "mandatory" | "best-effort";
|
|
1125
1127
|
maxToolCalls?: number;
|
|
1126
1128
|
mcpServers?: {
|
|
1127
1129
|
allowedTools?: string[];
|
|
@@ -1675,6 +1677,8 @@ interface paths {
|
|
|
1675
1677
|
timeoutMs?: number;
|
|
1676
1678
|
toolPool: string[];
|
|
1677
1679
|
};
|
|
1680
|
+
/** @enum {string} */
|
|
1681
|
+
intent?: "mandatory" | "best-effort";
|
|
1678
1682
|
maxToolCalls?: number;
|
|
1679
1683
|
mcpServers?: {
|
|
1680
1684
|
allowedTools?: string[];
|
|
@@ -2518,6 +2522,8 @@ interface paths {
|
|
|
2518
2522
|
timeoutMs?: number;
|
|
2519
2523
|
toolPool: string[];
|
|
2520
2524
|
} | null;
|
|
2525
|
+
/** @enum {string|null} */
|
|
2526
|
+
intent?: "mandatory" | "best-effort" | null;
|
|
2521
2527
|
maxToolCalls?: number;
|
|
2522
2528
|
mcpServers?: {
|
|
2523
2529
|
allowedTools?: string[];
|
|
@@ -14158,6 +14164,8 @@ interface paths {
|
|
|
14158
14164
|
args?: unknown;
|
|
14159
14165
|
at: string;
|
|
14160
14166
|
callId: string;
|
|
14167
|
+
/** @enum {boolean} */
|
|
14168
|
+
scriptRead?: true;
|
|
14161
14169
|
seq: number;
|
|
14162
14170
|
tool: string;
|
|
14163
14171
|
/** @enum {string} */
|
|
@@ -14174,6 +14182,8 @@ interface paths {
|
|
|
14174
14182
|
name?: string;
|
|
14175
14183
|
};
|
|
14176
14184
|
ok: boolean;
|
|
14185
|
+
/** @enum {boolean} */
|
|
14186
|
+
scriptRead?: true;
|
|
14177
14187
|
seq: number;
|
|
14178
14188
|
summary: string;
|
|
14179
14189
|
tool: string;
|
|
@@ -14479,7 +14489,7 @@ interface paths {
|
|
|
14479
14489
|
};
|
|
14480
14490
|
/**
|
|
14481
14491
|
* Stream a delegate task
|
|
14482
|
-
* @description Unified execution Server-Sent Events, replayed then tailed until the task settles. Each frame carries its seq as the SSE id; reconnect with `after=<last id>` or Last-Event-ID. Stream cursors differ from JSON journal cursors.
|
|
14492
|
+
* @description Unified execution Server-Sent Events, replayed then tailed until the task settles. Each frame carries its seq as the SSE id; reconnect with `after=<last id>` or Last-Event-ID. Stream cursors differ from JSON journal cursors. Reads a Code Mode script makes are not streamed; the JSON journal keeps them as events marked `scriptRead: true`.
|
|
14483
14493
|
*/
|
|
14484
14494
|
get: {
|
|
14485
14495
|
parameters: {
|
|
@@ -15472,6 +15482,8 @@ interface paths {
|
|
|
15472
15482
|
timeoutMs?: number;
|
|
15473
15483
|
toolPool: string[];
|
|
15474
15484
|
};
|
|
15485
|
+
/** @enum {string} */
|
|
15486
|
+
intent?: "mandatory" | "best-effort";
|
|
15475
15487
|
maxToolCalls?: number;
|
|
15476
15488
|
mcpServers?: unknown[];
|
|
15477
15489
|
perToolLimits?: {
|
|
@@ -15943,6 +15955,8 @@ interface paths {
|
|
|
15943
15955
|
timeoutMs?: number;
|
|
15944
15956
|
toolPool: string[];
|
|
15945
15957
|
};
|
|
15958
|
+
/** @enum {string} */
|
|
15959
|
+
intent?: "mandatory" | "best-effort";
|
|
15946
15960
|
maxToolCalls?: number;
|
|
15947
15961
|
mcpServers?: unknown[];
|
|
15948
15962
|
perToolLimits?: {
|
|
@@ -21076,6 +21090,8 @@ interface paths {
|
|
|
21076
21090
|
suggestion?: string;
|
|
21077
21091
|
}[];
|
|
21078
21092
|
error: string;
|
|
21093
|
+
/** @description On `insufficient_scope`: every API key scope this request needs. */
|
|
21094
|
+
requiredScopes?: string[];
|
|
21079
21095
|
};
|
|
21080
21096
|
};
|
|
21081
21097
|
};
|
|
@@ -58305,6 +58321,8 @@ interface components {
|
|
|
58305
58321
|
timeoutMs?: number;
|
|
58306
58322
|
toolPool: string[];
|
|
58307
58323
|
};
|
|
58324
|
+
/** @enum {string} */
|
|
58325
|
+
intent?: "mandatory" | "best-effort";
|
|
58308
58326
|
maxToolCalls?: number;
|
|
58309
58327
|
mcpServers?: {
|
|
58310
58328
|
allowedTools?: string[];
|
|
@@ -58898,9 +58916,11 @@ interface components {
|
|
|
58898
58916
|
metadata?: {
|
|
58899
58917
|
[key: string]: unknown;
|
|
58900
58918
|
};
|
|
58919
|
+
/** @description The model that wrote this assistant message. Informational on import: only a message the agent engine stored itself (marked server-side) replays its reasoning parts and provider options, and only to this same model. */
|
|
58920
|
+
model?: string;
|
|
58901
58921
|
/** @enum {string} */
|
|
58902
58922
|
role: "system" | "user" | "assistant" | "tool";
|
|
58903
|
-
/** @description Assistant messages only. Every entry must be answered by the next tool message. */
|
|
58923
|
+
/** @description Assistant messages only. Every entry must be answered by the next tool message. An agent turn stores one assistant and tool message pair per model round, in order. */
|
|
58904
58924
|
toolCalls?: {
|
|
58905
58925
|
/** @default {} */
|
|
58906
58926
|
args: {
|
|
@@ -59137,6 +59157,8 @@ interface components {
|
|
|
59137
59157
|
timeoutMs?: number;
|
|
59138
59158
|
toolPool: string[];
|
|
59139
59159
|
} | null;
|
|
59160
|
+
/** @enum {string|null} */
|
|
59161
|
+
intent?: "mandatory" | "best-effort" | null;
|
|
59140
59162
|
maxToolCalls?: number;
|
|
59141
59163
|
mcpServers?: {
|
|
59142
59164
|
allowedTools?: string[];
|
|
@@ -59610,6 +59632,8 @@ interface components {
|
|
|
59610
59632
|
notes?: string;
|
|
59611
59633
|
};
|
|
59612
59634
|
Error: {
|
|
59635
|
+
/** @description Stable error code. `insufficient_scope` marks an API key that lacks a scope this request needs. */
|
|
59636
|
+
code?: string;
|
|
59613
59637
|
details?: {
|
|
59614
59638
|
/** @description Stable Runtype validation code (e.g. MISSING_REQUIRED_FIELD, INVALID_ENUM). */
|
|
59615
59639
|
code?: string;
|
|
@@ -59620,6 +59644,8 @@ interface components {
|
|
|
59620
59644
|
suggestion?: string;
|
|
59621
59645
|
}[];
|
|
59622
59646
|
error: string;
|
|
59647
|
+
/** @description On `insufficient_scope`: every API key scope this request needs. */
|
|
59648
|
+
requiredScopes?: string[];
|
|
59623
59649
|
};
|
|
59624
59650
|
/** @description Version selector for an Agent-target suite. Supply at most one of alias or versionId; omit it to run the saved configuration. */
|
|
59625
59651
|
EvalAgentSelector: {
|
|
@@ -59940,6 +59966,32 @@ interface components {
|
|
|
59940
59966
|
* @enum {string|null}
|
|
59941
59967
|
*/
|
|
59942
59968
|
humanVerdict: "agree" | "disagree" | null;
|
|
59969
|
+
judgeDiagnostics?: {
|
|
59970
|
+
attempts: {
|
|
59971
|
+
attempt: 1 | 2;
|
|
59972
|
+
cacheReadTokens?: number;
|
|
59973
|
+
cacheWriteTokens?: number;
|
|
59974
|
+
completionTokens?: number;
|
|
59975
|
+
durationMs: number;
|
|
59976
|
+
/** @enum {string} */
|
|
59977
|
+
finishReason?: "stop" | "length" | "content-filter" | "tool-calls" | "error" | "other" | "unknown";
|
|
59978
|
+
/** @enum {string} */
|
|
59979
|
+
modelError?: "empty_completion" | "model_call_error";
|
|
59980
|
+
/** @enum {string} */
|
|
59981
|
+
parseError?: "empty_completion" | "invalid_json" | "invalid_contract" | "missing_access_assessment" | "invalid_access_assessment" | "output_truncated";
|
|
59982
|
+
promptTokens?: number;
|
|
59983
|
+
provider?: string;
|
|
59984
|
+
reasoningTokens?: number;
|
|
59985
|
+
requestedModel: string;
|
|
59986
|
+
resolvedModel?: string;
|
|
59987
|
+
responseChars?: number;
|
|
59988
|
+
/** @enum {string} */
|
|
59989
|
+
status: "valid" | "invalid_response" | "model_error";
|
|
59990
|
+
usageObserved: boolean;
|
|
59991
|
+
}[];
|
|
59992
|
+
/** @enum {string} */
|
|
59993
|
+
flowFailure?: "deadline_exhausted" | "execution_failed" | "invalid_response";
|
|
59994
|
+
};
|
|
59943
59995
|
kind: string;
|
|
59944
59996
|
passed: boolean;
|
|
59945
59997
|
reasoning?: string;
|
|
@@ -61525,6 +61577,32 @@ interface components {
|
|
|
61525
61577
|
kind: "message" | "tool_call";
|
|
61526
61578
|
}[];
|
|
61527
61579
|
graderIndex: number;
|
|
61580
|
+
judgeDiagnostics?: {
|
|
61581
|
+
attempts: {
|
|
61582
|
+
attempt: 1 | 2;
|
|
61583
|
+
cacheReadTokens?: number;
|
|
61584
|
+
cacheWriteTokens?: number;
|
|
61585
|
+
completionTokens?: number;
|
|
61586
|
+
durationMs: number;
|
|
61587
|
+
/** @enum {string} */
|
|
61588
|
+
finishReason?: "stop" | "length" | "content-filter" | "tool-calls" | "error" | "other" | "unknown";
|
|
61589
|
+
/** @enum {string} */
|
|
61590
|
+
modelError?: "empty_completion" | "model_call_error";
|
|
61591
|
+
/** @enum {string} */
|
|
61592
|
+
parseError?: "empty_completion" | "invalid_json" | "invalid_contract" | "missing_access_assessment" | "invalid_access_assessment" | "output_truncated";
|
|
61593
|
+
promptTokens?: number;
|
|
61594
|
+
provider?: string;
|
|
61595
|
+
reasoningTokens?: number;
|
|
61596
|
+
requestedModel: string;
|
|
61597
|
+
resolvedModel?: string;
|
|
61598
|
+
responseChars?: number;
|
|
61599
|
+
/** @enum {string} */
|
|
61600
|
+
status: "valid" | "invalid_response" | "model_error";
|
|
61601
|
+
usageObserved: boolean;
|
|
61602
|
+
}[];
|
|
61603
|
+
/** @enum {string} */
|
|
61604
|
+
flowFailure?: "deadline_exhausted" | "execution_failed" | "invalid_response";
|
|
61605
|
+
};
|
|
61528
61606
|
kind: string;
|
|
61529
61607
|
passed: boolean;
|
|
61530
61608
|
reasoning?: string;
|
|
@@ -70164,6 +70242,34 @@ export declare class SandboxDeploymentsNamespace {
|
|
|
70164
70242
|
update(id: string, data: ManagedDeploymentWriteRequest, options: ManagedDeploymentUpdateOptions): Promise<ManagedDeploymentAccepted>;
|
|
70165
70243
|
}
|
|
70166
70244
|
//#endregion
|
|
70245
|
+
//#region src/judge-diagnostics.d.ts
|
|
70246
|
+
/** The platform judge's response classification, without response text or schema error values. */
|
|
70247
|
+
type JudgeParseError = 'empty_completion' | 'invalid_json' | 'invalid_contract' | 'missing_access_assessment' | 'invalid_access_assessment' | 'output_truncated';
|
|
70248
|
+
/** Metadata from one platform judge attempt. Mirrors the shared contract without a runtime dependency. */
|
|
70249
|
+
interface JudgeAttemptDiagnostic {
|
|
70250
|
+
attempt: 1 | 2;
|
|
70251
|
+
requestedModel: string;
|
|
70252
|
+
resolvedModel?: string;
|
|
70253
|
+
provider?: string;
|
|
70254
|
+
durationMs: number;
|
|
70255
|
+
status: 'valid' | 'invalid_response' | 'model_error';
|
|
70256
|
+
parseError?: JudgeParseError;
|
|
70257
|
+
modelError?: 'empty_completion' | 'model_call_error';
|
|
70258
|
+
finishReason?: 'stop' | 'length' | 'content-filter' | 'tool-calls' | 'error' | 'other' | 'unknown';
|
|
70259
|
+
responseChars?: number;
|
|
70260
|
+
usageObserved: boolean;
|
|
70261
|
+
promptTokens?: number;
|
|
70262
|
+
completionTokens?: number;
|
|
70263
|
+
reasoningTokens?: number;
|
|
70264
|
+
cacheReadTokens?: number;
|
|
70265
|
+
cacheWriteTokens?: number;
|
|
70266
|
+
}
|
|
70267
|
+
/** Bounded judge metadata. Empty attempts require an explicit custom flow failure. */
|
|
70268
|
+
interface JudgeDiagnostics {
|
|
70269
|
+
attempts: JudgeAttemptDiagnostic[];
|
|
70270
|
+
flowFailure?: 'deadline_exhausted' | 'execution_failed' | 'invalid_response';
|
|
70271
|
+
}
|
|
70272
|
+
//#endregion
|
|
70167
70273
|
//#region src/evals-ensure.d.ts
|
|
70168
70274
|
/**
|
|
70169
70275
|
* Per-grader severity (mirror of `@runtypelabs/shared`'s `GraderSeverity`). A
|
|
@@ -70618,6 +70724,8 @@ interface EvalPullResult {
|
|
|
70618
70724
|
* array (`case.graders ?? suite.graders`).
|
|
70619
70725
|
*/
|
|
70620
70726
|
interface GraderOutcome {
|
|
70727
|
+
/** Platform judge attempt metadata; absent for older results, check graders, and custom judge flows. */
|
|
70728
|
+
judgeDiagnostics?: JudgeDiagnostics;
|
|
70621
70729
|
graderIndex: number;
|
|
70622
70730
|
kind: string;
|
|
70623
70731
|
passed: boolean;
|
|
@@ -73036,6 +73144,8 @@ interface AgentDefinitionConfig {
|
|
|
73036
73144
|
threshold?: number;
|
|
73037
73145
|
alwaysLoaded?: string[];
|
|
73038
73146
|
};
|
|
73147
|
+
/** `mandatory`: a non-idempotent tool runs only after its intent is recorded (default `best-effort`). */
|
|
73148
|
+
intent?: 'mandatory' | 'best-effort';
|
|
73039
73149
|
};
|
|
73040
73150
|
artifacts?: {
|
|
73041
73151
|
enabled: true;
|
|
@@ -75006,6 +75116,449 @@ interface DecisionModelCatalog {
|
|
|
75006
75116
|
models: DecisionModelCatalogEntry[];
|
|
75007
75117
|
}
|
|
75008
75118
|
//#endregion
|
|
75119
|
+
//#region src/simulation-tool-memory-types.d.ts
|
|
75120
|
+
/** How one tool call was answered: the rung of the response ladder that served it. */
|
|
75121
|
+
type SimulationServedBy = 'live' | 'recorded' | 'transformed' | 'composite' | 'synthesized' | 'stub' | 'fault';
|
|
75122
|
+
/** One recorded `tools/call` of a target; `arguments` and `result` are null on a shape-only row. */
|
|
75123
|
+
interface SimulationToolExchange {
|
|
75124
|
+
id: string;
|
|
75125
|
+
targetId: string;
|
|
75126
|
+
grant: string | null;
|
|
75127
|
+
serverTool: string;
|
|
75128
|
+
toolsHash: string | null;
|
|
75129
|
+
argsFingerprint: string;
|
|
75130
|
+
arguments: Record<string, unknown> | null;
|
|
75131
|
+
result: unknown;
|
|
75132
|
+
isError: boolean;
|
|
75133
|
+
errorCode: string | null;
|
|
75134
|
+
errorKey: string | null;
|
|
75135
|
+
durationMs: number;
|
|
75136
|
+
world: string;
|
|
75137
|
+
source: 'simulation' | 'transcript_backfill' | 'tool_executions' | 'uploaded';
|
|
75138
|
+
servedBy: SimulationServedBy;
|
|
75139
|
+
variantHash: string | null;
|
|
75140
|
+
ledgerHash: string | null;
|
|
75141
|
+
promptVersion?: number | null;
|
|
75142
|
+
rationale?: string | null;
|
|
75143
|
+
plausible?: boolean | null;
|
|
75144
|
+
shapeOnly: boolean;
|
|
75145
|
+
runId: string | null;
|
|
75146
|
+
episodeId: string | null;
|
|
75147
|
+
callIndex: number | null;
|
|
75148
|
+
createdAt: string;
|
|
75149
|
+
expiresAt: string | null;
|
|
75150
|
+
}
|
|
75151
|
+
interface ListSimulationToolExchangesParams {
|
|
75152
|
+
tool?: string;
|
|
75153
|
+
/** `true` lists only published recordings that kept their payload. */
|
|
75154
|
+
published?: boolean;
|
|
75155
|
+
limit?: number;
|
|
75156
|
+
cursor?: string;
|
|
75157
|
+
}
|
|
75158
|
+
interface ListSimulationToolExchangesResponse {
|
|
75159
|
+
exchanges: SimulationToolExchange[];
|
|
75160
|
+
nextCursor: string | null;
|
|
75161
|
+
}
|
|
75162
|
+
/** One call a server answered outside a simulation. */
|
|
75163
|
+
interface UploadedSimulationToolExchange {
|
|
75164
|
+
tool: string;
|
|
75165
|
+
arguments: Record<string, unknown>;
|
|
75166
|
+
result: unknown;
|
|
75167
|
+
isError?: boolean;
|
|
75168
|
+
}
|
|
75169
|
+
/** Uploaded exchanges, or a backfill from a run's transcripts or a Runtype MCP surface's tool executions. */
|
|
75170
|
+
type UploadSimulationToolExchangesRequest = {
|
|
75171
|
+
exchanges: UploadedSimulationToolExchange[];
|
|
75172
|
+
} | {
|
|
75173
|
+
backfill: 'transcripts' | 'tool_executions';
|
|
75174
|
+
runId?: string;
|
|
75175
|
+
};
|
|
75176
|
+
/** What Runtype learned about one tool of one target from its recorded exchanges. */
|
|
75177
|
+
interface SimulationToolProfile {
|
|
75178
|
+
targetId: string;
|
|
75179
|
+
tool: string;
|
|
75180
|
+
exchangesSeen: number;
|
|
75181
|
+
updatedAt: string;
|
|
75182
|
+
declaredOutputSchema: Record<string, unknown> | null;
|
|
75183
|
+
observedOutputSchema: unknown;
|
|
75184
|
+
samples: string[];
|
|
75185
|
+
errors: Record<string, {
|
|
75186
|
+
count: number;
|
|
75187
|
+
exemplarExchangeId: string | null;
|
|
75188
|
+
isError: boolean;
|
|
75189
|
+
code: string | null;
|
|
75190
|
+
recoveredShare: number | null;
|
|
75191
|
+
}>;
|
|
75192
|
+
fields: Record<string, {
|
|
75193
|
+
alwaysSet: boolean;
|
|
75194
|
+
types: string[];
|
|
75195
|
+
values: unknown[];
|
|
75196
|
+
valuesTruncated: boolean;
|
|
75197
|
+
}>;
|
|
75198
|
+
latencyMs: {
|
|
75199
|
+
p50: number | null;
|
|
75200
|
+
p95: number | null;
|
|
75201
|
+
};
|
|
75202
|
+
writesObserved: boolean;
|
|
75203
|
+
shapeFindings: Array<{
|
|
75204
|
+
kind: string;
|
|
75205
|
+
checklistRow: number;
|
|
75206
|
+
exemplarExchangeId: string | null;
|
|
75207
|
+
detail: string;
|
|
75208
|
+
exchangeIds: string[];
|
|
75209
|
+
suggestedOp?: Record<string, unknown>;
|
|
75210
|
+
}>;
|
|
75211
|
+
adoption: Array<Record<string, unknown>>;
|
|
75212
|
+
recordedAnswerAgeDays: number | null;
|
|
75213
|
+
latestToolsHash: string | null;
|
|
75214
|
+
synthesis?: {
|
|
75215
|
+
answers: number;
|
|
75216
|
+
reviewed: number;
|
|
75217
|
+
plausible: number;
|
|
75218
|
+
};
|
|
75219
|
+
}
|
|
75220
|
+
interface PreviewSimulationVariantRequest {
|
|
75221
|
+
/** The shown or server tool name. */
|
|
75222
|
+
tool: string;
|
|
75223
|
+
exchangeId: string;
|
|
75224
|
+
/** Answer once with the metered synthesizer and freeze the answer for the design's next run. */
|
|
75225
|
+
synthesize?: boolean;
|
|
75226
|
+
}
|
|
75227
|
+
interface PreviewSimulationVariantResponse {
|
|
75228
|
+
before: unknown;
|
|
75229
|
+
after: unknown;
|
|
75230
|
+
servedBy: SimulationServedBy;
|
|
75231
|
+
rationale: string | null;
|
|
75232
|
+
exchangeId: string;
|
|
75233
|
+
}
|
|
75234
|
+
/** The fields a variant update may change; `validate` checks the design as run admission would first. */
|
|
75235
|
+
interface UpdateSimulationVariantRequest {
|
|
75236
|
+
label?: string;
|
|
75237
|
+
role?: 'baseline' | 'candidate' | 'comparison';
|
|
75238
|
+
changes?: Array<Record<string, unknown>>;
|
|
75239
|
+
clientModel?: string | null;
|
|
75240
|
+
toolPolicy?: Record<string, 'live' | 'recorded' | 'simulated'> | null;
|
|
75241
|
+
replay?: {
|
|
75242
|
+
answerNeutralKeys?: string[];
|
|
75243
|
+
synthesize?: boolean;
|
|
75244
|
+
} | null;
|
|
75245
|
+
validate?: boolean;
|
|
75246
|
+
}
|
|
75247
|
+
interface UpdateSimulationTargetRequest {
|
|
75248
|
+
name?: string;
|
|
75249
|
+
config?: Record<string, unknown>;
|
|
75250
|
+
recording?: {
|
|
75251
|
+
enabled: boolean;
|
|
75252
|
+
retentionDays: number;
|
|
75253
|
+
};
|
|
75254
|
+
}
|
|
75255
|
+
interface ExportSimulationDesignBriefResponse {
|
|
75256
|
+
briefId: string;
|
|
75257
|
+
exportedAt: string;
|
|
75258
|
+
/** The brief as Markdown: headline, investment card, operations, before and after definitions, caveats. */
|
|
75259
|
+
markdown: string;
|
|
75260
|
+
}
|
|
75261
|
+
interface SimulationDesignBrief {
|
|
75262
|
+
briefId: string;
|
|
75263
|
+
exportedAt: string;
|
|
75264
|
+
runId: string;
|
|
75265
|
+
variantKey: string;
|
|
75266
|
+
targetId: string;
|
|
75267
|
+
prediction: {
|
|
75268
|
+
baseline: number;
|
|
75269
|
+
candidate: number;
|
|
75270
|
+
delta: number;
|
|
75271
|
+
low: number;
|
|
75272
|
+
high: number;
|
|
75273
|
+
headline: 'recommended' | 'indicative';
|
|
75274
|
+
} | null;
|
|
75275
|
+
operations: Array<{
|
|
75276
|
+
op: string;
|
|
75277
|
+
tool: string;
|
|
75278
|
+
state: 'adopted' | 'partial' | 'not_seen';
|
|
75279
|
+
detail: string | null;
|
|
75280
|
+
}>;
|
|
75281
|
+
observedAt: string | null;
|
|
75282
|
+
baselineAtExport: number | null;
|
|
75283
|
+
trend: Array<{
|
|
75284
|
+
runId: string;
|
|
75285
|
+
completedAt: string;
|
|
75286
|
+
passRate: number | null;
|
|
75287
|
+
live: boolean;
|
|
75288
|
+
}>;
|
|
75289
|
+
}
|
|
75290
|
+
//#endregion
|
|
75291
|
+
//#region src/simulation-episode-types.d.ts
|
|
75292
|
+
type SimulationEpisodeStatus = 'queued' | 'running' | 'completed' | 'failed' | 'skipped' | 'cancelled';
|
|
75293
|
+
/** Why an episode ended. */
|
|
75294
|
+
type SimulationTermination = 'goal_met' | 'gave_up' | 'max_turns' | 'agent_error' | 'simulator_error' | 'stalled' | 'deadline' | 'target_unavailable' | 'client_limit' | 'recording_miss' | 'spend_cap' | 'engine_error';
|
|
75295
|
+
/** Calls per response-ladder rung; a rung no call reached is absent. */
|
|
75296
|
+
type SimulationServedByCounts = Partial<Record<SimulationServedBy, number>>;
|
|
75297
|
+
/** One episode's measurements; target and client-agent cost are what a success costs, the rest is overhead. */
|
|
75298
|
+
interface SimulationEpisodeMetrics {
|
|
75299
|
+
interruptedMidTurn?: true;
|
|
75300
|
+
scoringFailed?: true;
|
|
75301
|
+
turns: number;
|
|
75302
|
+
targetCostUsd: number | null;
|
|
75303
|
+
clientAgentCostUsd: number | null;
|
|
75304
|
+
simulatorCostUsd: number;
|
|
75305
|
+
synthesizerCostUsd?: number;
|
|
75306
|
+
judgeCostUsd: number;
|
|
75307
|
+
inputTokens: number;
|
|
75308
|
+
outputTokens: number;
|
|
75309
|
+
totalLatencyMs: number;
|
|
75310
|
+
medianTurnLatencyMs: number | null;
|
|
75311
|
+
toolCalls: number;
|
|
75312
|
+
toolErrors: number;
|
|
75313
|
+
errors: number;
|
|
75314
|
+
unpricedModels: string[];
|
|
75315
|
+
egress?: {
|
|
75316
|
+
source: 'egress_proxy' | 'candidate_output' | 'unread';
|
|
75317
|
+
refusals: Array<{
|
|
75318
|
+
host: string;
|
|
75319
|
+
reason: string | null;
|
|
75320
|
+
count: number;
|
|
75321
|
+
}>;
|
|
75322
|
+
overflow: number;
|
|
75323
|
+
};
|
|
75324
|
+
/** False when an engine failure measured nothing: every number is a placeholder. */
|
|
75325
|
+
measured?: false;
|
|
75326
|
+
}
|
|
75327
|
+
/** A saved episode, without its transcript. */
|
|
75328
|
+
interface SimulationEpisode {
|
|
75329
|
+
id: string;
|
|
75330
|
+
runId: string;
|
|
75331
|
+
variantId: string;
|
|
75332
|
+
scenarioId: string | null;
|
|
75333
|
+
simulatedUserId: string | null;
|
|
75334
|
+
taskId: string | null;
|
|
75335
|
+
trialIndex: number;
|
|
75336
|
+
round: number;
|
|
75337
|
+
status: SimulationEpisodeStatus;
|
|
75338
|
+
termination: SimulationTermination | null;
|
|
75339
|
+
turns: number;
|
|
75340
|
+
passed: boolean | null;
|
|
75341
|
+
score: number | null;
|
|
75342
|
+
/** Why the episode is left out of pass rates; null when it counts. */
|
|
75343
|
+
leftOutReason: 'simulated_user_broke' | 'spend_cap' | 'engine_error' | null;
|
|
75344
|
+
isolation: 'guaranteed' | 'verified' | 'unverifiable' | 'live_tools' | null;
|
|
75345
|
+
overlay: 'proxy' | 'materialized' | 'shim' | null;
|
|
75346
|
+
ranOn: 'runtype_worker' | 'runtype_durable' | 'sandbox' | 'external_endpoint' | 'private_runner' | 'vendor' | null;
|
|
75347
|
+
fidelity: 'runtype' | 'real_code' | 'endpoint' | 'vendor_hosted' | 'simulated_agent' | null;
|
|
75348
|
+
worldKey: string | null;
|
|
75349
|
+
metrics: SimulationEpisodeMetrics | null;
|
|
75350
|
+
servedBy?: SimulationServedByCounts;
|
|
75351
|
+
executionIds: string[];
|
|
75352
|
+
conversationId: string | null;
|
|
75353
|
+
startedAt: string | null;
|
|
75354
|
+
completedAt: string | null;
|
|
75355
|
+
error: string | null;
|
|
75356
|
+
}
|
|
75357
|
+
interface ListSimulationEpisodesParams {
|
|
75358
|
+
status?: SimulationEpisodeStatus;
|
|
75359
|
+
/** `true` keeps episodes that passed grading, `false` the ones that failed it. */
|
|
75360
|
+
passed?: boolean;
|
|
75361
|
+
variantId?: string;
|
|
75362
|
+
simulatedUserId?: string;
|
|
75363
|
+
taskId?: string;
|
|
75364
|
+
limit?: number;
|
|
75365
|
+
cursor?: string;
|
|
75366
|
+
}
|
|
75367
|
+
interface ListSimulationEpisodesResponse {
|
|
75368
|
+
data: SimulationEpisode[];
|
|
75369
|
+
pagination: {
|
|
75370
|
+
nextCursor: string | null;
|
|
75371
|
+
hasMore: boolean;
|
|
75372
|
+
limit: number;
|
|
75373
|
+
};
|
|
75374
|
+
}
|
|
75375
|
+
/** One message of an episode transcript, with the tool calls and results of its turn. */
|
|
75376
|
+
interface SimulationTranscriptMessage {
|
|
75377
|
+
role: 'user' | 'assistant' | 'tool';
|
|
75378
|
+
content: string;
|
|
75379
|
+
toolCalls?: Array<{
|
|
75380
|
+
toolCallId: string;
|
|
75381
|
+
toolName: string;
|
|
75382
|
+
args: Record<string, unknown>;
|
|
75383
|
+
}>;
|
|
75384
|
+
toolResults?: Array<{
|
|
75385
|
+
toolCallId: string;
|
|
75386
|
+
toolName: string;
|
|
75387
|
+
result?: unknown;
|
|
75388
|
+
}>;
|
|
75389
|
+
}
|
|
75390
|
+
/** A redacted episode transcript with the simulated-user facts it was produced under. */
|
|
75391
|
+
interface SimulationEpisodeTranscript {
|
|
75392
|
+
messages: SimulationTranscriptMessage[];
|
|
75393
|
+
revealedKeys: Record<string, number>;
|
|
75394
|
+
simulatedUserModel: string | null;
|
|
75395
|
+
simulatedUserPromptVersion: number | null;
|
|
75396
|
+
judgePromptVersion: number | null;
|
|
75397
|
+
expiresAt: string | null;
|
|
75398
|
+
}
|
|
75399
|
+
/** One grader's outcome on one episode. */
|
|
75400
|
+
interface SimulationEpisodeScore {
|
|
75401
|
+
id: string;
|
|
75402
|
+
episodeId: string;
|
|
75403
|
+
runId: string;
|
|
75404
|
+
graderIndex: number;
|
|
75405
|
+
grader: Record<string, unknown>;
|
|
75406
|
+
passed: boolean;
|
|
75407
|
+
score: number | null;
|
|
75408
|
+
reasoning: string | null;
|
|
75409
|
+
verdict: string | null;
|
|
75410
|
+
evidence: Array<{
|
|
75411
|
+
kind: 'message' | 'tool_call';
|
|
75412
|
+
index: number;
|
|
75413
|
+
}> | null;
|
|
75414
|
+
outputExcerpt: string | null;
|
|
75415
|
+
error: string | null;
|
|
75416
|
+
costUsd: number | null;
|
|
75417
|
+
durationMs: number | null;
|
|
75418
|
+
judgeModel: string | null;
|
|
75419
|
+
humanVerdict: 'agree' | 'disagree' | null;
|
|
75420
|
+
}
|
|
75421
|
+
/** One tool call of a turn, as the layer facts keep it after the transcript expires. */
|
|
75422
|
+
interface SimulationToolCallFacts {
|
|
75423
|
+
callIndex: number;
|
|
75424
|
+
shownName: string;
|
|
75425
|
+
serverTool: string | null;
|
|
75426
|
+
argumentErrors: Array<{
|
|
75427
|
+
path: string;
|
|
75428
|
+
code: string;
|
|
75429
|
+
}>;
|
|
75430
|
+
errorKey: string | null;
|
|
75431
|
+
errorOrigin: 'server' | 'overlay' | 'transport' | null;
|
|
75432
|
+
retryOf: number | null;
|
|
75433
|
+
intendedTool: string | null;
|
|
75434
|
+
intendedSource: 'hint' | 'judge' | null;
|
|
75435
|
+
durationMs: number;
|
|
75436
|
+
stubbed?: boolean;
|
|
75437
|
+
readOnly?: boolean;
|
|
75438
|
+
errorText?: string | null;
|
|
75439
|
+
errorCode?: string | null;
|
|
75440
|
+
recoveredLater?: boolean;
|
|
75441
|
+
outcome?: 'ok' | 'denied' | 'not_found' | 'invalid' | 'failed';
|
|
75442
|
+
servedBy?: SimulationServedBy;
|
|
75443
|
+
exchangeId?: string | null;
|
|
75444
|
+
replayed?: boolean;
|
|
75445
|
+
designFailed?: boolean;
|
|
75446
|
+
/** The future-state operation that wrote this call's error envelope; absent for the upstream's own error. */
|
|
75447
|
+
designedBy?: 'shape' | 'paginate' | 'error_design' | 'accept_identifier' | 'add_tool';
|
|
75448
|
+
/** The synthesizer's answer failed its contract twice, so the call was answered with a stub. */
|
|
75449
|
+
synthesisFailed?: boolean;
|
|
75450
|
+
/** Why the synthesizer answered as it did, on a synthesized call. */
|
|
75451
|
+
rationale?: string;
|
|
75452
|
+
/** The frozen row of the suggestions filled into this call's designed error; `exchangeId` stays the answer's. */
|
|
75453
|
+
suggestionsExchangeId?: string | null;
|
|
75454
|
+
}
|
|
75455
|
+
/** What one turn crossed: surface, capability, implementation, models and tool calls. */
|
|
75456
|
+
interface SimulationTurnLayerFacts {
|
|
75457
|
+
turnIndex: number;
|
|
75458
|
+
surface: {
|
|
75459
|
+
surfaceId: string | null;
|
|
75460
|
+
surfaceType: string | null;
|
|
75461
|
+
routedCapabilityId: string | null;
|
|
75462
|
+
routerModel: string | null;
|
|
75463
|
+
routingMs: number | null;
|
|
75464
|
+
} | null;
|
|
75465
|
+
capabilityId: string | null;
|
|
75466
|
+
implementation: {
|
|
75467
|
+
kind: 'agent' | 'flow' | 'external';
|
|
75468
|
+
id: string | null;
|
|
75469
|
+
name?: string;
|
|
75470
|
+
} | null;
|
|
75471
|
+
tools: Array<{
|
|
75472
|
+
name: string;
|
|
75473
|
+
calls: number;
|
|
75474
|
+
errors: number;
|
|
75475
|
+
}>;
|
|
75476
|
+
models: Array<{
|
|
75477
|
+
model: string;
|
|
75478
|
+
calls: number;
|
|
75479
|
+
inputTokens: number;
|
|
75480
|
+
outputTokens: number;
|
|
75481
|
+
costUsd: number | null;
|
|
75482
|
+
}>;
|
|
75483
|
+
framework: string | null;
|
|
75484
|
+
clientModel: string | null;
|
|
75485
|
+
latencyMs: number;
|
|
75486
|
+
errored: boolean;
|
|
75487
|
+
toolCalls: SimulationToolCallFacts[];
|
|
75488
|
+
accessEvidence?: Record<string, unknown>;
|
|
75489
|
+
servedBy?: SimulationServedByCounts;
|
|
75490
|
+
synthesis?: {
|
|
75491
|
+
calls: number;
|
|
75492
|
+
inputTokens: number;
|
|
75493
|
+
outputTokens: number;
|
|
75494
|
+
costUsd: number;
|
|
75495
|
+
};
|
|
75496
|
+
}
|
|
75497
|
+
/** One episode with its transcript, grader scores and per-turn layer facts. */
|
|
75498
|
+
interface GetSimulationEpisodeResponse {
|
|
75499
|
+
episode: SimulationEpisode;
|
|
75500
|
+
/** Null when no transcript is kept (it expires); `layerStats` keeps the derived facts. */
|
|
75501
|
+
transcript: SimulationEpisodeTranscript | null;
|
|
75502
|
+
scores: SimulationEpisodeScore[];
|
|
75503
|
+
layerStats: SimulationTurnLayerFacts[];
|
|
75504
|
+
/** Each synthesized answer's review by its exchange id; null when unreviewed. */
|
|
75505
|
+
synthesisReviews?: Record<string, boolean | null>;
|
|
75506
|
+
}
|
|
75507
|
+
//#endregion
|
|
75508
|
+
//#region src/simulations-namespace.d.ts
|
|
75509
|
+
/**
|
|
75510
|
+
* Simulations tool memory and future designs (`/v1/simulations`): recorded exchanges, tool profiles,
|
|
75511
|
+
* design previews, synthesis review, design briefs and run episodes. Pre-GA: the API answers 404 `SIMULATIONS_DISABLED`
|
|
75512
|
+
* where simulations are not available (production today).
|
|
75513
|
+
*/
|
|
75514
|
+
export declare class SimulationsEndpoint {
|
|
75515
|
+
private client;
|
|
75516
|
+
constructor(client: ApiClient);
|
|
75517
|
+
/** Change a target's name, config or recording consent (`{ enabled, retentionDays }`). */
|
|
75518
|
+
updateTarget(targetId: string, data: UpdateSimulationTargetRequest): Promise<{
|
|
75519
|
+
target: Record<string, unknown>;
|
|
75520
|
+
}>;
|
|
75521
|
+
/** Save a variant's design; `validate: true` refuses it with 422 as run admission would. */
|
|
75522
|
+
updateVariant(simulationId: string, variantId: string, data: UpdateSimulationVariantRequest): Promise<{
|
|
75523
|
+
variant: Record<string, unknown>;
|
|
75524
|
+
}>;
|
|
75525
|
+
/** What Runtype learned about one tool of a target from its recorded answers. */
|
|
75526
|
+
getToolProfile(targetId: string, tool: string): Promise<SimulationToolProfile>;
|
|
75527
|
+
/** A target's recorded exchanges, newest first. */
|
|
75528
|
+
listExchanges(targetId: string, params?: ListSimulationToolExchangesParams): Promise<ListSimulationToolExchangesResponse>;
|
|
75529
|
+
/** Record calls your server answered outside a simulation, or backfill from transcripts or tool executions. */
|
|
75530
|
+
uploadExchanges(targetId: string, data: UploadSimulationToolExchangesRequest): Promise<{
|
|
75531
|
+
recorded: number;
|
|
75532
|
+
}>;
|
|
75533
|
+
/** Delete a target's recorded exchanges, or one tool's. */
|
|
75534
|
+
purgeExchanges(targetId: string, options?: {
|
|
75535
|
+
tool?: string;
|
|
75536
|
+
}): Promise<{
|
|
75537
|
+
deleted: number;
|
|
75538
|
+
}>;
|
|
75539
|
+
/** Answer one recorded exchange as published and as the variant's design answers it. */
|
|
75540
|
+
previewVariant(simulationId: string, variantId: string, data: PreviewSimulationVariantRequest): Promise<PreviewSimulationVariantResponse>;
|
|
75541
|
+
/** Discard a run variant's frozen synthesized answers so its next run synthesizes again. */
|
|
75542
|
+
resynthesizeVariant(runId: string, variantId: string): Promise<{
|
|
75543
|
+
discarded: number;
|
|
75544
|
+
}>;
|
|
75545
|
+
/** Mark a synthesized answer plausible or implausible; `null` clears the review. */
|
|
75546
|
+
reviewSynthesizedAnswer(runId: string, exchangeId: string, plausible: boolean | null): Promise<{
|
|
75547
|
+
exchangeId: string;
|
|
75548
|
+
plausible: boolean | null;
|
|
75549
|
+
}>;
|
|
75550
|
+
/** Export a run's future design as a Markdown brief; exporting again returns the same brief. */
|
|
75551
|
+
exportDesignBrief(runId: string): Promise<ExportSimulationDesignBriefResponse>;
|
|
75552
|
+
/** A simulation's exported briefs, with each operation's adoption state and the baseline since. */
|
|
75553
|
+
listDesignBriefs(simulationId: string): Promise<{
|
|
75554
|
+
briefs: SimulationDesignBrief[];
|
|
75555
|
+
}>;
|
|
75556
|
+
/** A run's episodes with status, error, turns, tool calls and served-by; filter `status: 'failed'` to debug. */
|
|
75557
|
+
listEpisodes(runId: string, params?: ListSimulationEpisodesParams): Promise<ListSimulationEpisodesResponse>;
|
|
75558
|
+
/** One episode with its transcript, grader scores and per-turn layer facts: why it failed or surprised. */
|
|
75559
|
+
getEpisode(runId: string, episodeId: string): Promise<GetSimulationEpisodeResponse>;
|
|
75560
|
+
}
|
|
75561
|
+
//#endregion
|
|
75009
75562
|
//#region src/client.d.ts
|
|
75010
75563
|
type LocalToolHandler = (args: unknown) => Promise<unknown>;
|
|
75011
75564
|
/**
|
|
@@ -75111,6 +75664,7 @@ export declare class RuntypeClient implements ApiClient {
|
|
|
75111
75664
|
integrations: IntegrationsEndpoint;
|
|
75112
75665
|
billing: BillingEndpoint;
|
|
75113
75666
|
toolApprovalGrants: ToolApprovalGrantsEndpoint;
|
|
75667
|
+
simulations: SimulationsEndpoint;
|
|
75114
75668
|
constructor(config?: ClientConfig);
|
|
75115
75669
|
/**
|
|
75116
75670
|
* Set the API key for authentication
|
|
@@ -76602,5 +77156,5 @@ export declare function getLikelySupportingCandidatePaths(bestCandidatePath: str
|
|
|
76602
77156
|
export declare function getDefaultPlanPath(taskName: string): string;
|
|
76603
77157
|
export declare function sanitizeTaskSlug(taskName: string): string;
|
|
76604
77158
|
//#endregion
|
|
76605
|
-
export type { A2AHandlerOptions, A2ARunInput, A2ATextMessage, AIGrader, ActivateAgentAliasInput, ActivateAgentPromotionInput, AgentAlias, AgentAliasActivation, AgentAliasArchiveFailure, AgentAliasArchived, AgentAliasList, AgentAliasTransport, AgentCapabilityChange, AgentCapabilityDeclaration, AgentDefinition, AgentDefinitionConfig, AgentDeploymentList, AgentDeploymentPromotion, AgentDeploymentReceipt, AgentPromotionManifest, AgentPromotionTransport, AgentPromotionValidation, AgentPullResult, AgentSkillBinding, AgentSkillBindingChange, AgentStreamEvent, ApiKeyRequest, ApiKeyRequestDelivery, ApiKeyRequestEnvironment, ApiKeyRequestHandoff, ApiKeyRequestListParams, ApiKeyRequestRequester, ApiKeyRequestStatus, ApplyGeneratedProposalOptions, ApplyGeneratedProposalResult, ArchiveAgentAliasEverywhereResult, ArchiveAgentAliasInput, AttachRuntimeToolsOptions, BatchClient, BatchListParams, BatchOptions, BatchRequest, BatchResult, BatchScheduleConfig, BatchStatus, BindSkillInput, BuiltInGraderId, CaseExpected, CheckGrader, ClaimApiKeyRequestInput, ClaimApiKeyRequestResponse, ClientConversation, ClientToken, ClientTokenConfig, ClientTokenDefinition, ClientTokenEnvironment, ClientTokenVersionPin, ClientWidgetTheme, CollectionPullResult, CollectionSchemaChange, ConditionalGetResult, ConditionalStepConfig, ContextErrorHandling, ContextFallback, CreateApiKeyRequestInput, CreateApiKeyRequestResponse, CreateClientTokenRequest, CreateClientTokenResponse, CreateEvalSuiteInput, CreatePromptData, DecisionModelCatalog, DecisionModelCatalogEntry, DecomposeCriteriaResult, DefineAgentInput, DefineEvalCaseInput, DefineEvalInput, DefineFlowInput, DefineProductInput, DefineScheduleInput, DefineSkillInput, DefineSurfaceInput, DefineToolInput, DetachedReattach, DetachedReconnectOptions, DispatchClient, DispatchEvent, DispatchOptions, EnsureAgentConverged, EnsureAgentOptions, EnsureAgentPlan, EnsureAgentResult, EnsureClientTokenConverged, EnsureClientTokenOptions, EnsureClientTokenPlan, EnsureClientTokenResult, EnsureCollectionConverged, EnsureCollectionOptions, EnsureCollectionPlan, EnsureCollectionResult, EnsureEvalApplyOptions, EnsureEvalOptions, EnsureEvalPlan, EnsureEvalResult, EnsureFlowConverged, EnsureFlowOptions, EnsureFlowPlan, EnsureFlowResult, EnsureFpoOptions, EnsureFpoResult, EnsureProductConverged, EnsureProductOptions, EnsureProductPlan, EnsureProductResult, EnsureQuestionSetConverged, EnsureQuestionSetOptions, EnsureQuestionSetPlan, EnsureQuestionSetResult, EnsureScheduleConverged, EnsureScheduleOptions, EnsureSchedulePlan, EnsureScheduleResult, EnsureSkillConverged, EnsureSkillOptions, EnsureSkillPlan, EnsureSkillResult, EnsureSurfaceConverged, EnsureSurfaceOptions, EnsureSurfacePlan, EnsureSurfaceResult, EnsureToolConverged, EnsureToolOptions, EnsureToolPlan, EnsureToolResult, ErrorHandlingMode, EvalAgentSelector, EvalAgentTargetResolution, EvalCaseDefinition, EvalCaseInput, EvalCaseProposal, EvalCaseProposalAccepted, EvalCaseProposalListResult, EvalCasesGenerated, EvalClient, EvalDefinition, EvalEndUserIdentity, EvalIdentityDefaults, EvalListParams, EvalMessage, EvalOptions, EvalProposalSource, EvalProposalStatus, EvalProposedCase, EvalPullResult, EvalRecord, EvalRequest, EvalResult, EvalRunCaseScores, EvalRunConfig, EvalRunEvidence, EvalRunScores, EvalStatus, EvalSuiteCase, EvalSuiteCaseInput, EvalSuiteCoverage, EvalSuiteDetail, EvalSuiteLatestRun, EvalSuiteListResult, EvalSuiteRunQueued, EvalSuiteRunResult, EvalSuiteSummary, EvalTarget, EvalTenantIdentity, ExecutionStreamEvent, ExecutionTerminalOutcome, FallbackFailEvent, FallbackStartEvent, FallbackSuccessEvent, FallbackTrigger, FallbackTriggerType, FallbacksExhaustedEvent, FallbacksInitiatedEvent, FetchUrlStepConfig, FieldFormat, FlowCompleteEvent, FlowConfig, FlowDefinition, FlowDefinitionStep, FlowErrorEvent, FlowFallback, FlowInlineEvalInput, FlowPausedEvent, FlowPullResult, FlowStartEvent, FlowStreamEvent, FlowSummary, FlowValidationClient, FlowValidationIssue, FlowValidationResult, FpoEntityOutcome, FpoInput, GenerateEmbeddingStepConfig, GenerateEvalCasesInput, GeneratedRuntimeToolGateDecision, GeneratedRuntimeToolGateOptions, GetApiKeyRequestResponse, GetRecordStepConfig, Gradeable, GraderConfig, GraderOutcome, GraderSeverity, HumanVerdict, ListAgentAliasesOptions, ListAgentDeploymentsOptions, ListConversationsResponse, ListRecordsStepConfig, LoopStepConfig, ManagedDeploymentAccepted, ManagedDeploymentCapability, ManagedDeploymentConfig, ManagedDeploymentCredentialAuthority, ManagedDeploymentDefinition, ManagedDeploymentDesiredSpec, ManagedDeploymentHealth, ManagedDeploymentLanguage, ManagedDeploymentProfileIdentity, ManagedDeploymentResource, ManagedDeploymentRetention, ManagedDeploymentRuntimeProfile, ManagedDeploymentSleepPolicy, ManagedDeploymentUpdateOptions, ManagedDeploymentWriteOptions, ManagedDeploymentWriteRequest, Message, MessageFallback, ModelFallback, ModelOverride, OrganizationAgentAlias, OrganizationAgentAliasList, PersistedGraderOutcome, PrepareAgentPromotionInput, ProductDefinition, ProductPullResult, PromoteAgentInput, PromoteAgentResult, PromptErrorHandling, PromptFallback, PromptListParams, PromptStepConfig, PullFpoResult, QuestionSetDefinition, QuestionSetPullResult, RecordCollectionDefinition, RecordCollectionValidationMode, RecordConfig, RecordFilter, RecordFilterCondition, RecordFilterGroup, RecordFilterOperator, RetrieveRecordStepConfig, RetryFallback, RollbackAgentAliasInput, RunEvalCaseResult, RunEvalInput, RunEvalResult, RunListParams, RunListResponse, RunListRow, AgentSkillBinding$1 as RuntypeAgentSkillBinding, ConditionalStepConfig$1 as RuntypeConditionalStepConfig, RuntypeConfig, FetchUrlStepConfig$1 as RuntypeFetchUrlStepConfig, FlowConfig$1 as RuntypeFlowConfig, GenerateEmbeddingStepConfig$1 as RuntypeGenerateEmbeddingStepConfig, GetRecordStepConfig$1 as RuntypeGetRecordStepConfig, ListRecordsStepConfig$1 as RuntypeListRecordsStepConfig, LoopStepConfig$1 as RuntypeLoopStepConfig, Message$1 as RuntypeMessage, ModelOverride$1 as RuntypeModelOverride, Prompt$1 as RuntypePrompt, PromptStepConfig$1 as RuntypePromptStepConfig, RecordConfig$1 as RuntypeRecordConfig, RetrieveRecordStepConfig$1 as RuntypeRetrieveRecordStepConfig, SearchStepConfig as RuntypeSearchStepConfig, SendEmailStepConfig as RuntypeSendEmailStepConfig, SendEventStepConfig as RuntypeSendEventStepConfig, SendStreamStepConfig as RuntypeSendStreamStepConfig, SetVariableStepConfig as RuntypeSetVariableStepConfig, Skill as RuntypeSkill, SkillBindingWithAgent as RuntypeSkillBindingWithAgent, SkillCapabilities as RuntypeSkillCapabilities, SkillFrontmatter as RuntypeSkillFrontmatter, SkillManifest as RuntypeSkillManifest, SkillProposal as RuntypeSkillProposal, SkillRuntypeExtensions as RuntypeSkillRuntypeExtensions, SkillVersion as RuntypeSkillVersion, TransformDataStepConfig as RuntypeTransformDataStepConfig, UpsertFlowConfig as RuntypeUpsertFlowConfig, UpsertRecordStepConfig as RuntypeUpsertRecordStepConfig, VectorSearchStepConfig as RuntypeVectorSearchStepConfig, WaitUntilStepConfig as RuntypeWaitUntilStepConfig, ScheduleDefinition, ScheduleDefinitionMessage, ScheduleEnsureTransport, SchedulePullResult, SearchStepConfig$1 as SearchStepConfig, SendEmailStepConfig$1 as SendEmailStepConfig, SendEventStepConfig$1 as SendEventStepConfig, SendStreamStepConfig$1 as SendStreamStepConfig, SetVariableStepConfig$1 as SetVariableStepConfig, SkillDefinition, SkillListPage, SkillListPagination, SkillListParams, SkillManifestInput, SkillMarkdownInput, SkillOrigin, SkillProposalStatus, SkillPullResult, SkillStatus, SkillTrustLevel, SkillVersionStatus, SkillWithVersion, SkillWriteInput, StepCompleteEvent, StepDeltaEvent, StepFallback, StepFieldMeta, StepStartEvent, StepWaitingLocalEvent, StreamCallbacks, StreamConsumeOptions, StreamEvent, StreamEventOf, SurfaceDefinition, SurfaceDefinitionStatus, SurfaceDefinitionType, SurfacePullResult, ToolDefinition, ToolDefinitionType, ToolPullResult, TransformDataStepConfig$1 as TransformDataStepConfig, UpdateClientTokenRequest, UpdateEvalCaseInput, UpdateEvalSuiteInput, UpdatePromptData, UpsertFlowConfig$1 as UpsertFlowConfig, UpsertRecordStepConfig$1 as UpsertRecordStepConfig, ValidateAgentPromotionInput, VectorSearchStepConfig$1 as VectorSearchStepConfig, WaitUntilStepConfig$1 as WaitUntilStepConfig, WorkflowCompileDeps, WorkflowCompletionCriteriaConfig, WorkflowConfig, WorkflowConfigFactory, WorkflowHookEntry, WorkflowHookKind, WorkflowHookRef, WorkflowHookSignatures, WorkflowMilestoneConfig, WorkflowPolicyConfig, WorkflowRecoveryConfig, WorkflowSlot };
|
|
77159
|
+
export type { A2AHandlerOptions, A2ARunInput, A2ATextMessage, AIGrader, ActivateAgentAliasInput, ActivateAgentPromotionInput, AgentAlias, AgentAliasActivation, AgentAliasArchiveFailure, AgentAliasArchived, AgentAliasList, AgentAliasTransport, AgentCapabilityChange, AgentCapabilityDeclaration, AgentDefinition, AgentDefinitionConfig, AgentDeploymentList, AgentDeploymentPromotion, AgentDeploymentReceipt, AgentPromotionManifest, AgentPromotionTransport, AgentPromotionValidation, AgentPullResult, AgentSkillBinding, AgentSkillBindingChange, AgentStreamEvent, ApiKeyRequest, ApiKeyRequestDelivery, ApiKeyRequestEnvironment, ApiKeyRequestHandoff, ApiKeyRequestListParams, ApiKeyRequestRequester, ApiKeyRequestStatus, ApplyGeneratedProposalOptions, ApplyGeneratedProposalResult, ArchiveAgentAliasEverywhereResult, ArchiveAgentAliasInput, AttachRuntimeToolsOptions, BatchClient, BatchListParams, BatchOptions, BatchRequest, BatchResult, BatchScheduleConfig, BatchStatus, BindSkillInput, BuiltInGraderId, CaseExpected, CheckGrader, ClaimApiKeyRequestInput, ClaimApiKeyRequestResponse, ClientConversation, ClientToken, ClientTokenConfig, ClientTokenDefinition, ClientTokenEnvironment, ClientTokenVersionPin, ClientWidgetTheme, CollectionPullResult, CollectionSchemaChange, ConditionalGetResult, ConditionalStepConfig, ContextErrorHandling, ContextFallback, CreateApiKeyRequestInput, CreateApiKeyRequestResponse, CreateClientTokenRequest, CreateClientTokenResponse, CreateEvalSuiteInput, CreatePromptData, DecisionModelCatalog, DecisionModelCatalogEntry, DecomposeCriteriaResult, DefineAgentInput, DefineEvalCaseInput, DefineEvalInput, DefineFlowInput, DefineProductInput, DefineScheduleInput, DefineSkillInput, DefineSurfaceInput, DefineToolInput, DetachedReattach, DetachedReconnectOptions, DispatchClient, DispatchEvent, DispatchOptions, EnsureAgentConverged, EnsureAgentOptions, EnsureAgentPlan, EnsureAgentResult, EnsureClientTokenConverged, EnsureClientTokenOptions, EnsureClientTokenPlan, EnsureClientTokenResult, EnsureCollectionConverged, EnsureCollectionOptions, EnsureCollectionPlan, EnsureCollectionResult, EnsureEvalApplyOptions, EnsureEvalOptions, EnsureEvalPlan, EnsureEvalResult, EnsureFlowConverged, EnsureFlowOptions, EnsureFlowPlan, EnsureFlowResult, EnsureFpoOptions, EnsureFpoResult, EnsureProductConverged, EnsureProductOptions, EnsureProductPlan, EnsureProductResult, EnsureQuestionSetConverged, EnsureQuestionSetOptions, EnsureQuestionSetPlan, EnsureQuestionSetResult, EnsureScheduleConverged, EnsureScheduleOptions, EnsureSchedulePlan, EnsureScheduleResult, EnsureSkillConverged, EnsureSkillOptions, EnsureSkillPlan, EnsureSkillResult, EnsureSurfaceConverged, EnsureSurfaceOptions, EnsureSurfacePlan, EnsureSurfaceResult, EnsureToolConverged, EnsureToolOptions, EnsureToolPlan, EnsureToolResult, ErrorHandlingMode, EvalAgentSelector, EvalAgentTargetResolution, EvalCaseDefinition, EvalCaseInput, EvalCaseProposal, EvalCaseProposalAccepted, EvalCaseProposalListResult, EvalCasesGenerated, EvalClient, EvalDefinition, EvalEndUserIdentity, EvalIdentityDefaults, EvalListParams, EvalMessage, EvalOptions, EvalProposalSource, EvalProposalStatus, EvalProposedCase, EvalPullResult, EvalRecord, EvalRequest, EvalResult, EvalRunCaseScores, EvalRunConfig, EvalRunEvidence, EvalRunScores, EvalStatus, EvalSuiteCase, EvalSuiteCaseInput, EvalSuiteCoverage, EvalSuiteDetail, EvalSuiteLatestRun, EvalSuiteListResult, EvalSuiteRunQueued, EvalSuiteRunResult, EvalSuiteSummary, EvalTarget, EvalTenantIdentity, ExecutionStreamEvent, ExecutionTerminalOutcome, ExportSimulationDesignBriefResponse, FallbackFailEvent, FallbackStartEvent, FallbackSuccessEvent, FallbackTrigger, FallbackTriggerType, FallbacksExhaustedEvent, FallbacksInitiatedEvent, FetchUrlStepConfig, FieldFormat, FlowCompleteEvent, FlowConfig, FlowDefinition, FlowDefinitionStep, FlowErrorEvent, FlowFallback, FlowInlineEvalInput, FlowPausedEvent, FlowPullResult, FlowStartEvent, FlowStreamEvent, FlowSummary, FlowValidationClient, FlowValidationIssue, FlowValidationResult, FpoEntityOutcome, FpoInput, GenerateEmbeddingStepConfig, GenerateEvalCasesInput, GeneratedRuntimeToolGateDecision, GeneratedRuntimeToolGateOptions, GetApiKeyRequestResponse, GetRecordStepConfig, GetSimulationEpisodeResponse, Gradeable, GraderConfig, GraderOutcome, GraderSeverity, HumanVerdict, JudgeAttemptDiagnostic, JudgeDiagnostics, JudgeParseError, ListAgentAliasesOptions, ListAgentDeploymentsOptions, ListConversationsResponse, ListRecordsStepConfig, ListSimulationEpisodesParams, ListSimulationEpisodesResponse, ListSimulationToolExchangesParams, ListSimulationToolExchangesResponse, LoopStepConfig, ManagedDeploymentAccepted, ManagedDeploymentCapability, ManagedDeploymentConfig, ManagedDeploymentCredentialAuthority, ManagedDeploymentDefinition, ManagedDeploymentDesiredSpec, ManagedDeploymentHealth, ManagedDeploymentLanguage, ManagedDeploymentProfileIdentity, ManagedDeploymentResource, ManagedDeploymentRetention, ManagedDeploymentRuntimeProfile, ManagedDeploymentSleepPolicy, ManagedDeploymentUpdateOptions, ManagedDeploymentWriteOptions, ManagedDeploymentWriteRequest, Message, MessageFallback, ModelFallback, ModelOverride, OrganizationAgentAlias, OrganizationAgentAliasList, PersistedGraderOutcome, PrepareAgentPromotionInput, PreviewSimulationVariantRequest, PreviewSimulationVariantResponse, ProductDefinition, ProductPullResult, PromoteAgentInput, PromoteAgentResult, PromptErrorHandling, PromptFallback, PromptListParams, PromptStepConfig, PullFpoResult, QuestionSetDefinition, QuestionSetPullResult, RecordCollectionDefinition, RecordCollectionValidationMode, RecordConfig, RecordFilter, RecordFilterCondition, RecordFilterGroup, RecordFilterOperator, RetrieveRecordStepConfig, RetryFallback, RollbackAgentAliasInput, RunEvalCaseResult, RunEvalInput, RunEvalResult, RunListParams, RunListResponse, RunListRow, AgentSkillBinding$1 as RuntypeAgentSkillBinding, ConditionalStepConfig$1 as RuntypeConditionalStepConfig, RuntypeConfig, FetchUrlStepConfig$1 as RuntypeFetchUrlStepConfig, FlowConfig$1 as RuntypeFlowConfig, GenerateEmbeddingStepConfig$1 as RuntypeGenerateEmbeddingStepConfig, GetRecordStepConfig$1 as RuntypeGetRecordStepConfig, ListRecordsStepConfig$1 as RuntypeListRecordsStepConfig, LoopStepConfig$1 as RuntypeLoopStepConfig, Message$1 as RuntypeMessage, ModelOverride$1 as RuntypeModelOverride, Prompt$1 as RuntypePrompt, PromptStepConfig$1 as RuntypePromptStepConfig, RecordConfig$1 as RuntypeRecordConfig, RetrieveRecordStepConfig$1 as RuntypeRetrieveRecordStepConfig, SearchStepConfig as RuntypeSearchStepConfig, SendEmailStepConfig as RuntypeSendEmailStepConfig, SendEventStepConfig as RuntypeSendEventStepConfig, SendStreamStepConfig as RuntypeSendStreamStepConfig, SetVariableStepConfig as RuntypeSetVariableStepConfig, Skill as RuntypeSkill, SkillBindingWithAgent as RuntypeSkillBindingWithAgent, SkillCapabilities as RuntypeSkillCapabilities, SkillFrontmatter as RuntypeSkillFrontmatter, SkillManifest as RuntypeSkillManifest, SkillProposal as RuntypeSkillProposal, SkillRuntypeExtensions as RuntypeSkillRuntypeExtensions, SkillVersion as RuntypeSkillVersion, TransformDataStepConfig as RuntypeTransformDataStepConfig, UpsertFlowConfig as RuntypeUpsertFlowConfig, UpsertRecordStepConfig as RuntypeUpsertRecordStepConfig, VectorSearchStepConfig as RuntypeVectorSearchStepConfig, WaitUntilStepConfig as RuntypeWaitUntilStepConfig, ScheduleDefinition, ScheduleDefinitionMessage, ScheduleEnsureTransport, SchedulePullResult, SearchStepConfig$1 as SearchStepConfig, SendEmailStepConfig$1 as SendEmailStepConfig, SendEventStepConfig$1 as SendEventStepConfig, SendStreamStepConfig$1 as SendStreamStepConfig, SetVariableStepConfig$1 as SetVariableStepConfig, SimulationDesignBrief, SimulationEpisode, SimulationEpisodeMetrics, SimulationEpisodeScore, SimulationEpisodeStatus, SimulationEpisodeTranscript, SimulationServedBy, SimulationServedByCounts, SimulationTermination, SimulationToolCallFacts, SimulationToolExchange, SimulationToolProfile, SimulationTranscriptMessage, SimulationTurnLayerFacts, SkillDefinition, SkillListPage, SkillListPagination, SkillListParams, SkillManifestInput, SkillMarkdownInput, SkillOrigin, SkillProposalStatus, SkillPullResult, SkillStatus, SkillTrustLevel, SkillVersionStatus, SkillWithVersion, SkillWriteInput, StepCompleteEvent, StepDeltaEvent, StepFallback, StepFieldMeta, StepStartEvent, StepWaitingLocalEvent, StreamCallbacks, StreamConsumeOptions, StreamEvent, StreamEventOf, SurfaceDefinition, SurfaceDefinitionStatus, SurfaceDefinitionType, SurfacePullResult, ToolDefinition, ToolDefinitionType, ToolPullResult, TransformDataStepConfig$1 as TransformDataStepConfig, UpdateClientTokenRequest, UpdateEvalCaseInput, UpdateEvalSuiteInput, UpdatePromptData, UpdateSimulationTargetRequest, UpdateSimulationVariantRequest, UploadSimulationToolExchangesRequest, UploadedSimulationToolExchange, UpsertFlowConfig$1 as UpsertFlowConfig, UpsertRecordStepConfig$1 as UpsertRecordStepConfig, ValidateAgentPromotionInput, VectorSearchStepConfig$1 as VectorSearchStepConfig, WaitUntilStepConfig$1 as WaitUntilStepConfig, WorkflowCompileDeps, WorkflowCompletionCriteriaConfig, WorkflowConfig, WorkflowConfigFactory, WorkflowHookEntry, WorkflowHookKind, WorkflowHookRef, WorkflowHookSignatures, WorkflowMilestoneConfig, WorkflowPolicyConfig, WorkflowRecoveryConfig, WorkflowSlot };
|
|
76606
77160
|
export {}
|