@runtypelabs/sdk 11.2.0 → 11.3.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/dist/index.cjs +82 -1
- package/dist/index.d.cts +530 -2
- package/dist/index.d.ts +530 -2
- package/dist/index.mjs +82 -2
- package/package.json +2 -2
package/dist/index.d.ts
CHANGED
|
@@ -58898,9 +58898,11 @@ interface components {
|
|
|
58898
58898
|
metadata?: {
|
|
58899
58899
|
[key: string]: unknown;
|
|
58900
58900
|
};
|
|
58901
|
+
/** @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. */
|
|
58902
|
+
model?: string;
|
|
58901
58903
|
/** @enum {string} */
|
|
58902
58904
|
role: "system" | "user" | "assistant" | "tool";
|
|
58903
|
-
/** @description Assistant messages only. Every entry must be answered by the next tool message. */
|
|
58905
|
+
/** @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
58906
|
toolCalls?: {
|
|
58905
58907
|
/** @default {} */
|
|
58906
58908
|
args: {
|
|
@@ -59940,6 +59942,32 @@ interface components {
|
|
|
59940
59942
|
* @enum {string|null}
|
|
59941
59943
|
*/
|
|
59942
59944
|
humanVerdict: "agree" | "disagree" | null;
|
|
59945
|
+
judgeDiagnostics?: {
|
|
59946
|
+
attempts: {
|
|
59947
|
+
attempt: 1 | 2;
|
|
59948
|
+
cacheReadTokens?: number;
|
|
59949
|
+
cacheWriteTokens?: number;
|
|
59950
|
+
completionTokens?: number;
|
|
59951
|
+
durationMs: number;
|
|
59952
|
+
/** @enum {string} */
|
|
59953
|
+
finishReason?: "stop" | "length" | "content-filter" | "tool-calls" | "error" | "other" | "unknown";
|
|
59954
|
+
/** @enum {string} */
|
|
59955
|
+
modelError?: "empty_completion" | "model_call_error";
|
|
59956
|
+
/** @enum {string} */
|
|
59957
|
+
parseError?: "empty_completion" | "invalid_json" | "invalid_contract" | "missing_access_assessment" | "invalid_access_assessment" | "output_truncated";
|
|
59958
|
+
promptTokens?: number;
|
|
59959
|
+
provider?: string;
|
|
59960
|
+
reasoningTokens?: number;
|
|
59961
|
+
requestedModel: string;
|
|
59962
|
+
resolvedModel?: string;
|
|
59963
|
+
responseChars?: number;
|
|
59964
|
+
/** @enum {string} */
|
|
59965
|
+
status: "valid" | "invalid_response" | "model_error";
|
|
59966
|
+
usageObserved: boolean;
|
|
59967
|
+
}[];
|
|
59968
|
+
/** @enum {string} */
|
|
59969
|
+
flowFailure?: "deadline_exhausted" | "execution_failed" | "invalid_response";
|
|
59970
|
+
};
|
|
59943
59971
|
kind: string;
|
|
59944
59972
|
passed: boolean;
|
|
59945
59973
|
reasoning?: string;
|
|
@@ -61525,6 +61553,32 @@ interface components {
|
|
|
61525
61553
|
kind: "message" | "tool_call";
|
|
61526
61554
|
}[];
|
|
61527
61555
|
graderIndex: number;
|
|
61556
|
+
judgeDiagnostics?: {
|
|
61557
|
+
attempts: {
|
|
61558
|
+
attempt: 1 | 2;
|
|
61559
|
+
cacheReadTokens?: number;
|
|
61560
|
+
cacheWriteTokens?: number;
|
|
61561
|
+
completionTokens?: number;
|
|
61562
|
+
durationMs: number;
|
|
61563
|
+
/** @enum {string} */
|
|
61564
|
+
finishReason?: "stop" | "length" | "content-filter" | "tool-calls" | "error" | "other" | "unknown";
|
|
61565
|
+
/** @enum {string} */
|
|
61566
|
+
modelError?: "empty_completion" | "model_call_error";
|
|
61567
|
+
/** @enum {string} */
|
|
61568
|
+
parseError?: "empty_completion" | "invalid_json" | "invalid_contract" | "missing_access_assessment" | "invalid_access_assessment" | "output_truncated";
|
|
61569
|
+
promptTokens?: number;
|
|
61570
|
+
provider?: string;
|
|
61571
|
+
reasoningTokens?: number;
|
|
61572
|
+
requestedModel: string;
|
|
61573
|
+
resolvedModel?: string;
|
|
61574
|
+
responseChars?: number;
|
|
61575
|
+
/** @enum {string} */
|
|
61576
|
+
status: "valid" | "invalid_response" | "model_error";
|
|
61577
|
+
usageObserved: boolean;
|
|
61578
|
+
}[];
|
|
61579
|
+
/** @enum {string} */
|
|
61580
|
+
flowFailure?: "deadline_exhausted" | "execution_failed" | "invalid_response";
|
|
61581
|
+
};
|
|
61528
61582
|
kind: string;
|
|
61529
61583
|
passed: boolean;
|
|
61530
61584
|
reasoning?: string;
|
|
@@ -70164,6 +70218,34 @@ export declare class SandboxDeploymentsNamespace {
|
|
|
70164
70218
|
update(id: string, data: ManagedDeploymentWriteRequest, options: ManagedDeploymentUpdateOptions): Promise<ManagedDeploymentAccepted>;
|
|
70165
70219
|
}
|
|
70166
70220
|
//#endregion
|
|
70221
|
+
//#region src/judge-diagnostics.d.ts
|
|
70222
|
+
/** The platform judge's response classification, without response text or schema error values. */
|
|
70223
|
+
type JudgeParseError = 'empty_completion' | 'invalid_json' | 'invalid_contract' | 'missing_access_assessment' | 'invalid_access_assessment' | 'output_truncated';
|
|
70224
|
+
/** Metadata from one platform judge attempt. Mirrors the shared contract without a runtime dependency. */
|
|
70225
|
+
interface JudgeAttemptDiagnostic {
|
|
70226
|
+
attempt: 1 | 2;
|
|
70227
|
+
requestedModel: string;
|
|
70228
|
+
resolvedModel?: string;
|
|
70229
|
+
provider?: string;
|
|
70230
|
+
durationMs: number;
|
|
70231
|
+
status: 'valid' | 'invalid_response' | 'model_error';
|
|
70232
|
+
parseError?: JudgeParseError;
|
|
70233
|
+
modelError?: 'empty_completion' | 'model_call_error';
|
|
70234
|
+
finishReason?: 'stop' | 'length' | 'content-filter' | 'tool-calls' | 'error' | 'other' | 'unknown';
|
|
70235
|
+
responseChars?: number;
|
|
70236
|
+
usageObserved: boolean;
|
|
70237
|
+
promptTokens?: number;
|
|
70238
|
+
completionTokens?: number;
|
|
70239
|
+
reasoningTokens?: number;
|
|
70240
|
+
cacheReadTokens?: number;
|
|
70241
|
+
cacheWriteTokens?: number;
|
|
70242
|
+
}
|
|
70243
|
+
/** Bounded judge metadata. Empty attempts require an explicit custom flow failure. */
|
|
70244
|
+
interface JudgeDiagnostics {
|
|
70245
|
+
attempts: JudgeAttemptDiagnostic[];
|
|
70246
|
+
flowFailure?: 'deadline_exhausted' | 'execution_failed' | 'invalid_response';
|
|
70247
|
+
}
|
|
70248
|
+
//#endregion
|
|
70167
70249
|
//#region src/evals-ensure.d.ts
|
|
70168
70250
|
/**
|
|
70169
70251
|
* Per-grader severity (mirror of `@runtypelabs/shared`'s `GraderSeverity`). A
|
|
@@ -70618,6 +70700,8 @@ interface EvalPullResult {
|
|
|
70618
70700
|
* array (`case.graders ?? suite.graders`).
|
|
70619
70701
|
*/
|
|
70620
70702
|
interface GraderOutcome {
|
|
70703
|
+
/** Platform judge attempt metadata; absent for older results, check graders, and custom judge flows. */
|
|
70704
|
+
judgeDiagnostics?: JudgeDiagnostics;
|
|
70621
70705
|
graderIndex: number;
|
|
70622
70706
|
kind: string;
|
|
70623
70707
|
passed: boolean;
|
|
@@ -75006,6 +75090,449 @@ interface DecisionModelCatalog {
|
|
|
75006
75090
|
models: DecisionModelCatalogEntry[];
|
|
75007
75091
|
}
|
|
75008
75092
|
//#endregion
|
|
75093
|
+
//#region src/simulation-tool-memory-types.d.ts
|
|
75094
|
+
/** How one tool call was answered: the rung of the response ladder that served it. */
|
|
75095
|
+
type SimulationServedBy = 'live' | 'recorded' | 'transformed' | 'composite' | 'synthesized' | 'stub' | 'fault';
|
|
75096
|
+
/** One recorded `tools/call` of a target; `arguments` and `result` are null on a shape-only row. */
|
|
75097
|
+
interface SimulationToolExchange {
|
|
75098
|
+
id: string;
|
|
75099
|
+
targetId: string;
|
|
75100
|
+
grant: string | null;
|
|
75101
|
+
serverTool: string;
|
|
75102
|
+
toolsHash: string | null;
|
|
75103
|
+
argsFingerprint: string;
|
|
75104
|
+
arguments: Record<string, unknown> | null;
|
|
75105
|
+
result: unknown;
|
|
75106
|
+
isError: boolean;
|
|
75107
|
+
errorCode: string | null;
|
|
75108
|
+
errorKey: string | null;
|
|
75109
|
+
durationMs: number;
|
|
75110
|
+
world: string;
|
|
75111
|
+
source: 'simulation' | 'transcript_backfill' | 'tool_executions' | 'uploaded';
|
|
75112
|
+
servedBy: SimulationServedBy;
|
|
75113
|
+
variantHash: string | null;
|
|
75114
|
+
ledgerHash: string | null;
|
|
75115
|
+
promptVersion?: number | null;
|
|
75116
|
+
rationale?: string | null;
|
|
75117
|
+
plausible?: boolean | null;
|
|
75118
|
+
shapeOnly: boolean;
|
|
75119
|
+
runId: string | null;
|
|
75120
|
+
episodeId: string | null;
|
|
75121
|
+
callIndex: number | null;
|
|
75122
|
+
createdAt: string;
|
|
75123
|
+
expiresAt: string | null;
|
|
75124
|
+
}
|
|
75125
|
+
interface ListSimulationToolExchangesParams {
|
|
75126
|
+
tool?: string;
|
|
75127
|
+
/** `true` lists only published recordings that kept their payload. */
|
|
75128
|
+
published?: boolean;
|
|
75129
|
+
limit?: number;
|
|
75130
|
+
cursor?: string;
|
|
75131
|
+
}
|
|
75132
|
+
interface ListSimulationToolExchangesResponse {
|
|
75133
|
+
exchanges: SimulationToolExchange[];
|
|
75134
|
+
nextCursor: string | null;
|
|
75135
|
+
}
|
|
75136
|
+
/** One call a server answered outside a simulation. */
|
|
75137
|
+
interface UploadedSimulationToolExchange {
|
|
75138
|
+
tool: string;
|
|
75139
|
+
arguments: Record<string, unknown>;
|
|
75140
|
+
result: unknown;
|
|
75141
|
+
isError?: boolean;
|
|
75142
|
+
}
|
|
75143
|
+
/** Uploaded exchanges, or a backfill from a run's transcripts or a Runtype MCP surface's tool executions. */
|
|
75144
|
+
type UploadSimulationToolExchangesRequest = {
|
|
75145
|
+
exchanges: UploadedSimulationToolExchange[];
|
|
75146
|
+
} | {
|
|
75147
|
+
backfill: 'transcripts' | 'tool_executions';
|
|
75148
|
+
runId?: string;
|
|
75149
|
+
};
|
|
75150
|
+
/** What Runtype learned about one tool of one target from its recorded exchanges. */
|
|
75151
|
+
interface SimulationToolProfile {
|
|
75152
|
+
targetId: string;
|
|
75153
|
+
tool: string;
|
|
75154
|
+
exchangesSeen: number;
|
|
75155
|
+
updatedAt: string;
|
|
75156
|
+
declaredOutputSchema: Record<string, unknown> | null;
|
|
75157
|
+
observedOutputSchema: unknown;
|
|
75158
|
+
samples: string[];
|
|
75159
|
+
errors: Record<string, {
|
|
75160
|
+
count: number;
|
|
75161
|
+
exemplarExchangeId: string | null;
|
|
75162
|
+
isError: boolean;
|
|
75163
|
+
code: string | null;
|
|
75164
|
+
recoveredShare: number | null;
|
|
75165
|
+
}>;
|
|
75166
|
+
fields: Record<string, {
|
|
75167
|
+
alwaysSet: boolean;
|
|
75168
|
+
types: string[];
|
|
75169
|
+
values: unknown[];
|
|
75170
|
+
valuesTruncated: boolean;
|
|
75171
|
+
}>;
|
|
75172
|
+
latencyMs: {
|
|
75173
|
+
p50: number | null;
|
|
75174
|
+
p95: number | null;
|
|
75175
|
+
};
|
|
75176
|
+
writesObserved: boolean;
|
|
75177
|
+
shapeFindings: Array<{
|
|
75178
|
+
kind: string;
|
|
75179
|
+
checklistRow: number;
|
|
75180
|
+
exemplarExchangeId: string | null;
|
|
75181
|
+
detail: string;
|
|
75182
|
+
exchangeIds: string[];
|
|
75183
|
+
suggestedOp?: Record<string, unknown>;
|
|
75184
|
+
}>;
|
|
75185
|
+
adoption: Array<Record<string, unknown>>;
|
|
75186
|
+
recordedAnswerAgeDays: number | null;
|
|
75187
|
+
latestToolsHash: string | null;
|
|
75188
|
+
synthesis?: {
|
|
75189
|
+
answers: number;
|
|
75190
|
+
reviewed: number;
|
|
75191
|
+
plausible: number;
|
|
75192
|
+
};
|
|
75193
|
+
}
|
|
75194
|
+
interface PreviewSimulationVariantRequest {
|
|
75195
|
+
/** The shown or server tool name. */
|
|
75196
|
+
tool: string;
|
|
75197
|
+
exchangeId: string;
|
|
75198
|
+
/** Answer once with the metered synthesizer and freeze the answer for the design's next run. */
|
|
75199
|
+
synthesize?: boolean;
|
|
75200
|
+
}
|
|
75201
|
+
interface PreviewSimulationVariantResponse {
|
|
75202
|
+
before: unknown;
|
|
75203
|
+
after: unknown;
|
|
75204
|
+
servedBy: SimulationServedBy;
|
|
75205
|
+
rationale: string | null;
|
|
75206
|
+
exchangeId: string;
|
|
75207
|
+
}
|
|
75208
|
+
/** The fields a variant update may change; `validate` checks the design as run admission would first. */
|
|
75209
|
+
interface UpdateSimulationVariantRequest {
|
|
75210
|
+
label?: string;
|
|
75211
|
+
role?: 'baseline' | 'candidate' | 'comparison';
|
|
75212
|
+
changes?: Array<Record<string, unknown>>;
|
|
75213
|
+
clientModel?: string | null;
|
|
75214
|
+
toolPolicy?: Record<string, 'live' | 'recorded' | 'simulated'> | null;
|
|
75215
|
+
replay?: {
|
|
75216
|
+
answerNeutralKeys?: string[];
|
|
75217
|
+
synthesize?: boolean;
|
|
75218
|
+
} | null;
|
|
75219
|
+
validate?: boolean;
|
|
75220
|
+
}
|
|
75221
|
+
interface UpdateSimulationTargetRequest {
|
|
75222
|
+
name?: string;
|
|
75223
|
+
config?: Record<string, unknown>;
|
|
75224
|
+
recording?: {
|
|
75225
|
+
enabled: boolean;
|
|
75226
|
+
retentionDays: number;
|
|
75227
|
+
};
|
|
75228
|
+
}
|
|
75229
|
+
interface ExportSimulationDesignBriefResponse {
|
|
75230
|
+
briefId: string;
|
|
75231
|
+
exportedAt: string;
|
|
75232
|
+
/** The brief as Markdown: headline, investment card, operations, before and after definitions, caveats. */
|
|
75233
|
+
markdown: string;
|
|
75234
|
+
}
|
|
75235
|
+
interface SimulationDesignBrief {
|
|
75236
|
+
briefId: string;
|
|
75237
|
+
exportedAt: string;
|
|
75238
|
+
runId: string;
|
|
75239
|
+
variantKey: string;
|
|
75240
|
+
targetId: string;
|
|
75241
|
+
prediction: {
|
|
75242
|
+
baseline: number;
|
|
75243
|
+
candidate: number;
|
|
75244
|
+
delta: number;
|
|
75245
|
+
low: number;
|
|
75246
|
+
high: number;
|
|
75247
|
+
headline: 'recommended' | 'indicative';
|
|
75248
|
+
} | null;
|
|
75249
|
+
operations: Array<{
|
|
75250
|
+
op: string;
|
|
75251
|
+
tool: string;
|
|
75252
|
+
state: 'adopted' | 'partial' | 'not_seen';
|
|
75253
|
+
detail: string | null;
|
|
75254
|
+
}>;
|
|
75255
|
+
observedAt: string | null;
|
|
75256
|
+
baselineAtExport: number | null;
|
|
75257
|
+
trend: Array<{
|
|
75258
|
+
runId: string;
|
|
75259
|
+
completedAt: string;
|
|
75260
|
+
passRate: number | null;
|
|
75261
|
+
live: boolean;
|
|
75262
|
+
}>;
|
|
75263
|
+
}
|
|
75264
|
+
//#endregion
|
|
75265
|
+
//#region src/simulation-episode-types.d.ts
|
|
75266
|
+
type SimulationEpisodeStatus = 'queued' | 'running' | 'completed' | 'failed' | 'skipped' | 'cancelled';
|
|
75267
|
+
/** Why an episode ended. */
|
|
75268
|
+
type SimulationTermination = 'goal_met' | 'gave_up' | 'max_turns' | 'agent_error' | 'simulator_error' | 'stalled' | 'deadline' | 'target_unavailable' | 'client_limit' | 'recording_miss' | 'spend_cap' | 'engine_error';
|
|
75269
|
+
/** Calls per response-ladder rung; a rung no call reached is absent. */
|
|
75270
|
+
type SimulationServedByCounts = Partial<Record<SimulationServedBy, number>>;
|
|
75271
|
+
/** One episode's measurements; target and client-agent cost are what a success costs, the rest is overhead. */
|
|
75272
|
+
interface SimulationEpisodeMetrics {
|
|
75273
|
+
interruptedMidTurn?: true;
|
|
75274
|
+
scoringFailed?: true;
|
|
75275
|
+
turns: number;
|
|
75276
|
+
targetCostUsd: number | null;
|
|
75277
|
+
clientAgentCostUsd: number | null;
|
|
75278
|
+
simulatorCostUsd: number;
|
|
75279
|
+
synthesizerCostUsd?: number;
|
|
75280
|
+
judgeCostUsd: number;
|
|
75281
|
+
inputTokens: number;
|
|
75282
|
+
outputTokens: number;
|
|
75283
|
+
totalLatencyMs: number;
|
|
75284
|
+
medianTurnLatencyMs: number | null;
|
|
75285
|
+
toolCalls: number;
|
|
75286
|
+
toolErrors: number;
|
|
75287
|
+
errors: number;
|
|
75288
|
+
unpricedModels: string[];
|
|
75289
|
+
egress?: {
|
|
75290
|
+
source: 'egress_proxy' | 'candidate_output' | 'unread';
|
|
75291
|
+
refusals: Array<{
|
|
75292
|
+
host: string;
|
|
75293
|
+
reason: string | null;
|
|
75294
|
+
count: number;
|
|
75295
|
+
}>;
|
|
75296
|
+
overflow: number;
|
|
75297
|
+
};
|
|
75298
|
+
/** False when an engine failure measured nothing: every number is a placeholder. */
|
|
75299
|
+
measured?: false;
|
|
75300
|
+
}
|
|
75301
|
+
/** A saved episode, without its transcript. */
|
|
75302
|
+
interface SimulationEpisode {
|
|
75303
|
+
id: string;
|
|
75304
|
+
runId: string;
|
|
75305
|
+
variantId: string;
|
|
75306
|
+
scenarioId: string | null;
|
|
75307
|
+
simulatedUserId: string | null;
|
|
75308
|
+
taskId: string | null;
|
|
75309
|
+
trialIndex: number;
|
|
75310
|
+
round: number;
|
|
75311
|
+
status: SimulationEpisodeStatus;
|
|
75312
|
+
termination: SimulationTermination | null;
|
|
75313
|
+
turns: number;
|
|
75314
|
+
passed: boolean | null;
|
|
75315
|
+
score: number | null;
|
|
75316
|
+
/** Why the episode is left out of pass rates; null when it counts. */
|
|
75317
|
+
leftOutReason: 'simulated_user_broke' | 'spend_cap' | 'engine_error' | null;
|
|
75318
|
+
isolation: 'guaranteed' | 'verified' | 'unverifiable' | 'live_tools' | null;
|
|
75319
|
+
overlay: 'proxy' | 'materialized' | 'shim' | null;
|
|
75320
|
+
ranOn: 'runtype_worker' | 'runtype_durable' | 'sandbox' | 'external_endpoint' | 'private_runner' | 'vendor' | null;
|
|
75321
|
+
fidelity: 'runtype' | 'real_code' | 'endpoint' | 'vendor_hosted' | 'simulated_agent' | null;
|
|
75322
|
+
worldKey: string | null;
|
|
75323
|
+
metrics: SimulationEpisodeMetrics | null;
|
|
75324
|
+
servedBy?: SimulationServedByCounts;
|
|
75325
|
+
executionIds: string[];
|
|
75326
|
+
conversationId: string | null;
|
|
75327
|
+
startedAt: string | null;
|
|
75328
|
+
completedAt: string | null;
|
|
75329
|
+
error: string | null;
|
|
75330
|
+
}
|
|
75331
|
+
interface ListSimulationEpisodesParams {
|
|
75332
|
+
status?: SimulationEpisodeStatus;
|
|
75333
|
+
/** `true` keeps episodes that passed grading, `false` the ones that failed it. */
|
|
75334
|
+
passed?: boolean;
|
|
75335
|
+
variantId?: string;
|
|
75336
|
+
simulatedUserId?: string;
|
|
75337
|
+
taskId?: string;
|
|
75338
|
+
limit?: number;
|
|
75339
|
+
cursor?: string;
|
|
75340
|
+
}
|
|
75341
|
+
interface ListSimulationEpisodesResponse {
|
|
75342
|
+
data: SimulationEpisode[];
|
|
75343
|
+
pagination: {
|
|
75344
|
+
nextCursor: string | null;
|
|
75345
|
+
hasMore: boolean;
|
|
75346
|
+
limit: number;
|
|
75347
|
+
};
|
|
75348
|
+
}
|
|
75349
|
+
/** One message of an episode transcript, with the tool calls and results of its turn. */
|
|
75350
|
+
interface SimulationTranscriptMessage {
|
|
75351
|
+
role: 'user' | 'assistant' | 'tool';
|
|
75352
|
+
content: string;
|
|
75353
|
+
toolCalls?: Array<{
|
|
75354
|
+
toolCallId: string;
|
|
75355
|
+
toolName: string;
|
|
75356
|
+
args: Record<string, unknown>;
|
|
75357
|
+
}>;
|
|
75358
|
+
toolResults?: Array<{
|
|
75359
|
+
toolCallId: string;
|
|
75360
|
+
toolName: string;
|
|
75361
|
+
result?: unknown;
|
|
75362
|
+
}>;
|
|
75363
|
+
}
|
|
75364
|
+
/** A redacted episode transcript with the simulated-user facts it was produced under. */
|
|
75365
|
+
interface SimulationEpisodeTranscript {
|
|
75366
|
+
messages: SimulationTranscriptMessage[];
|
|
75367
|
+
revealedKeys: Record<string, number>;
|
|
75368
|
+
simulatedUserModel: string | null;
|
|
75369
|
+
simulatedUserPromptVersion: number | null;
|
|
75370
|
+
judgePromptVersion: number | null;
|
|
75371
|
+
expiresAt: string | null;
|
|
75372
|
+
}
|
|
75373
|
+
/** One grader's outcome on one episode. */
|
|
75374
|
+
interface SimulationEpisodeScore {
|
|
75375
|
+
id: string;
|
|
75376
|
+
episodeId: string;
|
|
75377
|
+
runId: string;
|
|
75378
|
+
graderIndex: number;
|
|
75379
|
+
grader: Record<string, unknown>;
|
|
75380
|
+
passed: boolean;
|
|
75381
|
+
score: number | null;
|
|
75382
|
+
reasoning: string | null;
|
|
75383
|
+
verdict: string | null;
|
|
75384
|
+
evidence: Array<{
|
|
75385
|
+
kind: 'message' | 'tool_call';
|
|
75386
|
+
index: number;
|
|
75387
|
+
}> | null;
|
|
75388
|
+
outputExcerpt: string | null;
|
|
75389
|
+
error: string | null;
|
|
75390
|
+
costUsd: number | null;
|
|
75391
|
+
durationMs: number | null;
|
|
75392
|
+
judgeModel: string | null;
|
|
75393
|
+
humanVerdict: 'agree' | 'disagree' | null;
|
|
75394
|
+
}
|
|
75395
|
+
/** One tool call of a turn, as the layer facts keep it after the transcript expires. */
|
|
75396
|
+
interface SimulationToolCallFacts {
|
|
75397
|
+
callIndex: number;
|
|
75398
|
+
shownName: string;
|
|
75399
|
+
serverTool: string | null;
|
|
75400
|
+
argumentErrors: Array<{
|
|
75401
|
+
path: string;
|
|
75402
|
+
code: string;
|
|
75403
|
+
}>;
|
|
75404
|
+
errorKey: string | null;
|
|
75405
|
+
errorOrigin: 'server' | 'overlay' | 'transport' | null;
|
|
75406
|
+
retryOf: number | null;
|
|
75407
|
+
intendedTool: string | null;
|
|
75408
|
+
intendedSource: 'hint' | 'judge' | null;
|
|
75409
|
+
durationMs: number;
|
|
75410
|
+
stubbed?: boolean;
|
|
75411
|
+
readOnly?: boolean;
|
|
75412
|
+
errorText?: string | null;
|
|
75413
|
+
errorCode?: string | null;
|
|
75414
|
+
recoveredLater?: boolean;
|
|
75415
|
+
outcome?: 'ok' | 'denied' | 'not_found' | 'invalid' | 'failed';
|
|
75416
|
+
servedBy?: SimulationServedBy;
|
|
75417
|
+
exchangeId?: string | null;
|
|
75418
|
+
replayed?: boolean;
|
|
75419
|
+
designFailed?: boolean;
|
|
75420
|
+
/** The future-state operation that wrote this call's error envelope; absent for the upstream's own error. */
|
|
75421
|
+
designedBy?: 'shape' | 'paginate' | 'error_design' | 'accept_identifier' | 'add_tool';
|
|
75422
|
+
/** The synthesizer's answer failed its contract twice, so the call was answered with a stub. */
|
|
75423
|
+
synthesisFailed?: boolean;
|
|
75424
|
+
/** Why the synthesizer answered as it did, on a synthesized call. */
|
|
75425
|
+
rationale?: string;
|
|
75426
|
+
/** The frozen row of the suggestions filled into this call's designed error; `exchangeId` stays the answer's. */
|
|
75427
|
+
suggestionsExchangeId?: string | null;
|
|
75428
|
+
}
|
|
75429
|
+
/** What one turn crossed: surface, capability, implementation, models and tool calls. */
|
|
75430
|
+
interface SimulationTurnLayerFacts {
|
|
75431
|
+
turnIndex: number;
|
|
75432
|
+
surface: {
|
|
75433
|
+
surfaceId: string | null;
|
|
75434
|
+
surfaceType: string | null;
|
|
75435
|
+
routedCapabilityId: string | null;
|
|
75436
|
+
routerModel: string | null;
|
|
75437
|
+
routingMs: number | null;
|
|
75438
|
+
} | null;
|
|
75439
|
+
capabilityId: string | null;
|
|
75440
|
+
implementation: {
|
|
75441
|
+
kind: 'agent' | 'flow' | 'external';
|
|
75442
|
+
id: string | null;
|
|
75443
|
+
name?: string;
|
|
75444
|
+
} | null;
|
|
75445
|
+
tools: Array<{
|
|
75446
|
+
name: string;
|
|
75447
|
+
calls: number;
|
|
75448
|
+
errors: number;
|
|
75449
|
+
}>;
|
|
75450
|
+
models: Array<{
|
|
75451
|
+
model: string;
|
|
75452
|
+
calls: number;
|
|
75453
|
+
inputTokens: number;
|
|
75454
|
+
outputTokens: number;
|
|
75455
|
+
costUsd: number | null;
|
|
75456
|
+
}>;
|
|
75457
|
+
framework: string | null;
|
|
75458
|
+
clientModel: string | null;
|
|
75459
|
+
latencyMs: number;
|
|
75460
|
+
errored: boolean;
|
|
75461
|
+
toolCalls: SimulationToolCallFacts[];
|
|
75462
|
+
accessEvidence?: Record<string, unknown>;
|
|
75463
|
+
servedBy?: SimulationServedByCounts;
|
|
75464
|
+
synthesis?: {
|
|
75465
|
+
calls: number;
|
|
75466
|
+
inputTokens: number;
|
|
75467
|
+
outputTokens: number;
|
|
75468
|
+
costUsd: number;
|
|
75469
|
+
};
|
|
75470
|
+
}
|
|
75471
|
+
/** One episode with its transcript, grader scores and per-turn layer facts. */
|
|
75472
|
+
interface GetSimulationEpisodeResponse {
|
|
75473
|
+
episode: SimulationEpisode;
|
|
75474
|
+
/** Null when no transcript is kept (it expires); `layerStats` keeps the derived facts. */
|
|
75475
|
+
transcript: SimulationEpisodeTranscript | null;
|
|
75476
|
+
scores: SimulationEpisodeScore[];
|
|
75477
|
+
layerStats: SimulationTurnLayerFacts[];
|
|
75478
|
+
/** Each synthesized answer's review by its exchange id; null when unreviewed. */
|
|
75479
|
+
synthesisReviews?: Record<string, boolean | null>;
|
|
75480
|
+
}
|
|
75481
|
+
//#endregion
|
|
75482
|
+
//#region src/simulations-namespace.d.ts
|
|
75483
|
+
/**
|
|
75484
|
+
* Simulations tool memory and future designs (`/v1/simulations`): recorded exchanges, tool profiles,
|
|
75485
|
+
* design previews, synthesis review, design briefs and run episodes. Pre-GA: the API answers 404 `SIMULATIONS_DISABLED`
|
|
75486
|
+
* where simulations are not available (production today).
|
|
75487
|
+
*/
|
|
75488
|
+
export declare class SimulationsEndpoint {
|
|
75489
|
+
private client;
|
|
75490
|
+
constructor(client: ApiClient);
|
|
75491
|
+
/** Change a target's name, config or recording consent (`{ enabled, retentionDays }`). */
|
|
75492
|
+
updateTarget(targetId: string, data: UpdateSimulationTargetRequest): Promise<{
|
|
75493
|
+
target: Record<string, unknown>;
|
|
75494
|
+
}>;
|
|
75495
|
+
/** Save a variant's design; `validate: true` refuses it with 422 as run admission would. */
|
|
75496
|
+
updateVariant(simulationId: string, variantId: string, data: UpdateSimulationVariantRequest): Promise<{
|
|
75497
|
+
variant: Record<string, unknown>;
|
|
75498
|
+
}>;
|
|
75499
|
+
/** What Runtype learned about one tool of a target from its recorded answers. */
|
|
75500
|
+
getToolProfile(targetId: string, tool: string): Promise<SimulationToolProfile>;
|
|
75501
|
+
/** A target's recorded exchanges, newest first. */
|
|
75502
|
+
listExchanges(targetId: string, params?: ListSimulationToolExchangesParams): Promise<ListSimulationToolExchangesResponse>;
|
|
75503
|
+
/** Record calls your server answered outside a simulation, or backfill from transcripts or tool executions. */
|
|
75504
|
+
uploadExchanges(targetId: string, data: UploadSimulationToolExchangesRequest): Promise<{
|
|
75505
|
+
recorded: number;
|
|
75506
|
+
}>;
|
|
75507
|
+
/** Delete a target's recorded exchanges, or one tool's. */
|
|
75508
|
+
purgeExchanges(targetId: string, options?: {
|
|
75509
|
+
tool?: string;
|
|
75510
|
+
}): Promise<{
|
|
75511
|
+
deleted: number;
|
|
75512
|
+
}>;
|
|
75513
|
+
/** Answer one recorded exchange as published and as the variant's design answers it. */
|
|
75514
|
+
previewVariant(simulationId: string, variantId: string, data: PreviewSimulationVariantRequest): Promise<PreviewSimulationVariantResponse>;
|
|
75515
|
+
/** Discard a run variant's frozen synthesized answers so its next run synthesizes again. */
|
|
75516
|
+
resynthesizeVariant(runId: string, variantId: string): Promise<{
|
|
75517
|
+
discarded: number;
|
|
75518
|
+
}>;
|
|
75519
|
+
/** Mark a synthesized answer plausible or implausible; `null` clears the review. */
|
|
75520
|
+
reviewSynthesizedAnswer(runId: string, exchangeId: string, plausible: boolean | null): Promise<{
|
|
75521
|
+
exchangeId: string;
|
|
75522
|
+
plausible: boolean | null;
|
|
75523
|
+
}>;
|
|
75524
|
+
/** Export a run's future design as a Markdown brief; exporting again returns the same brief. */
|
|
75525
|
+
exportDesignBrief(runId: string): Promise<ExportSimulationDesignBriefResponse>;
|
|
75526
|
+
/** A simulation's exported briefs, with each operation's adoption state and the baseline since. */
|
|
75527
|
+
listDesignBriefs(simulationId: string): Promise<{
|
|
75528
|
+
briefs: SimulationDesignBrief[];
|
|
75529
|
+
}>;
|
|
75530
|
+
/** A run's episodes with status, error, turns, tool calls and served-by; filter `status: 'failed'` to debug. */
|
|
75531
|
+
listEpisodes(runId: string, params?: ListSimulationEpisodesParams): Promise<ListSimulationEpisodesResponse>;
|
|
75532
|
+
/** One episode with its transcript, grader scores and per-turn layer facts: why it failed or surprised. */
|
|
75533
|
+
getEpisode(runId: string, episodeId: string): Promise<GetSimulationEpisodeResponse>;
|
|
75534
|
+
}
|
|
75535
|
+
//#endregion
|
|
75009
75536
|
//#region src/client.d.ts
|
|
75010
75537
|
type LocalToolHandler = (args: unknown) => Promise<unknown>;
|
|
75011
75538
|
/**
|
|
@@ -75111,6 +75638,7 @@ export declare class RuntypeClient implements ApiClient {
|
|
|
75111
75638
|
integrations: IntegrationsEndpoint;
|
|
75112
75639
|
billing: BillingEndpoint;
|
|
75113
75640
|
toolApprovalGrants: ToolApprovalGrantsEndpoint;
|
|
75641
|
+
simulations: SimulationsEndpoint;
|
|
75114
75642
|
constructor(config?: ClientConfig);
|
|
75115
75643
|
/**
|
|
75116
75644
|
* Set the API key for authentication
|
|
@@ -76602,5 +77130,5 @@ export declare function getLikelySupportingCandidatePaths(bestCandidatePath: str
|
|
|
76602
77130
|
export declare function getDefaultPlanPath(taskName: string): string;
|
|
76603
77131
|
export declare function sanitizeTaskSlug(taskName: string): string;
|
|
76604
77132
|
//#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 };
|
|
77133
|
+
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
77134
|
export {}
|