@mutagent/sdk 0.2.111 → 0.2.115

package/src/models/errors/ws-error.ts

@@ -0,0 +1,69 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ */
+
+import * as z from "zod/v4-mini";
+import * as types from "../../types/primitives.js";
+import { MutagentError } from "./mutagent-error.js";
+
+/**
+ * Standard error response
+ */
+export type WsErrorData = {
+  /**
+   * Error type or code
+   */
+  error: string;
+  /**
+   * Human-readable error message
+   */
+  message: string;
+  /**
+   * HTTP status code
+   */
+  statusCode?: number | undefined;
+};
+
+/**
+ * Standard error response
+ */
+export class WsError extends MutagentError {
+  /**
+   * Error type or code
+   */
+  error: string;
+
+  /** The original data that was passed to this error instance. */
+  data$: WsErrorData;
+
+  constructor(
+    err: WsErrorData,
+    httpMeta: { response: Response; request: Request; body: string },
+  ) {
+    const message = err.message || `API error occurred: ${JSON.stringify(err)}`;
+    super(message, httpMeta);
+    this.data$ = err;
+    this.error = err.error;
+
+    this.name = "WsError";
+  }
+}
+
+/** @internal */
+export const WsError$inboundSchema: z.ZodMiniType<WsError, unknown> = z.pipe(
+  z.object({
+    error: types.string(),
+    message: types.string(),
+    statusCode: types.optional(types.number()),
+    request$: z.custom<Request>(x => x instanceof Request),
+    response$: z.custom<Response>(x => x instanceof Response),
+    body$: z.string(),
+  }),
+  z.transform((v) => {
+    return new WsError(v, {
+      request: v.request$,
+      response: v.response$,
+      body: v.body$,
+    });
+  }),
+);

package/src/models/evaluation-history.ts

@@ -8,6 +8,26 @@ import { Result as SafeParseResult } from "../types/fp.js";
 import * as types from "../types/primitives.js";
 import { SDKValidationError } from "./errors/sdk-validation-error.js";
 
+export type EvaluationHistoryCriterion = {
+  id?: string | undefined;
+  /**
+   * Rubric text describing how to score this criterion
+   */
+  criteria: string;
+  /**
+   * Schema field this criterion scores against
+   */
+  evaluationParameter: string;
+  evaluationSteps?: Array<string> | undefined;
+  name?: string | undefined;
+};
+
+export type EvaluationHistoryEvalConfig = {
+  criteria?: Array<EvaluationHistoryCriterion> | undefined;
+  threshold?: number | undefined;
+  model?: string | undefined;
+};
+
 /**
  * Evaluation definition object. This specifies HOW to evaluate, not the results (see PromptEvaluationResult for execution outcomes)
  */
@@ -20,6 +40,10 @@ export type EvaluationHistory = {
    * UUID of the prompt family being evaluated (cache this for performance)
    */
   promptGroupId: string;
+  /**
+   * ID of the test dataset used for evaluation (null if not yet linked)
+   */
+  datasetId: number | null;
   /**
    * Evaluation name
    */
@@ -28,10 +52,7 @@ export type EvaluationHistory = {
    * Evaluation description
    */
   description: string | null;
-  /**
-   * Evaluation configuration with criteria array
-   */
-  evalConfig: any;
+  evalConfig: EvaluationHistoryEvalConfig;
   /**
    * LLM execution configuration
    */
@@ -59,25 +80,65 @@ export type EvaluationHistory = {
   /**
    * UUID grouping all versions of this evaluation together
    */
-  evaluationGroupId?: string | undefined;
+  evaluationGroupId: string;
   /**
    * Semantic version string (e.g., "1.0.0", "1.1.0")
    */
-  version?: string | undefined;
+  version: string;
   /**
    * ID of the previous version in the version chain (null for v1.0.0)
    */
-  ancestorId?: number | null | undefined;
+  ancestorId: number | null;
   /**
    * Whether this is the latest version in the evaluation group
    */
-  isLatest?: boolean | undefined;
-  /**
-   * ID of the test dataset used for evaluation
-   */
-  datasetId?: number | undefined;
+  isLatest: boolean;
 };
 
+/** @internal */
+export const EvaluationHistoryCriterion$inboundSchema: z.ZodMiniType<
+  EvaluationHistoryCriterion,
+  unknown
+> = z.object({
+  id: types.optional(types.string()),
+  criteria: types.string(),
+  evaluationParameter: types.string(),
+  evaluationSteps: types.optional(z.array(types.string())),
+  name: types.optional(types.string()),
+});
+
+export function evaluationHistoryCriterionFromJSON(
+  jsonString: string,
+): SafeParseResult<EvaluationHistoryCriterion, SDKValidationError> {
+  return safeParse(
+    jsonString,
+    (x) => EvaluationHistoryCriterion$inboundSchema.parse(JSON.parse(x)),
+    `Failed to parse 'EvaluationHistoryCriterion' from JSON`,
+  );
+}
+
+/** @internal */
+export const EvaluationHistoryEvalConfig$inboundSchema: z.ZodMiniType<
+  EvaluationHistoryEvalConfig,
+  unknown
+> = z.object({
+  criteria: types.optional(
+    z.array(z.lazy(() => EvaluationHistoryCriterion$inboundSchema)),
+  ),
+  threshold: types.optional(types.number()),
+  model: types.optional(types.string()),
+});
+
+export function evaluationHistoryEvalConfigFromJSON(
+  jsonString: string,
+): SafeParseResult<EvaluationHistoryEvalConfig, SDKValidationError> {
+  return safeParse(
+    jsonString,
+    (x) => EvaluationHistoryEvalConfig$inboundSchema.parse(JSON.parse(x)),
+    `Failed to parse 'EvaluationHistoryEvalConfig' from JSON`,
+  );
+}
+
 /** @internal */
 export const EvaluationHistory$inboundSchema: z.ZodMiniType<
   EvaluationHistory,
@@ -85,20 +146,20 @@ export const EvaluationHistory$inboundSchema: z.ZodMiniType<
 > = z.object({
   id: types.number(),
   promptGroupId: types.string(),
+  datasetId: types.nullable(types.number()),
   name: types.string(),
   description: types.nullable(types.string()),
-  evalConfig: z.any(),
+  evalConfig: z.lazy(() => EvaluationHistoryEvalConfig$inboundSchema),
   llmConfig: z.any(),
   tags: types.nullable(z.array(types.string())),
   metadata: z.any(),
   createdAt: types.nullable(types.date()),
   updatedAt: types.nullable(types.date()),
   createdBy: types.nullable(types.string()),
-  evaluationGroupId: types.optional(types.string()),
-  version: types.optional(types.string()),
-  ancestorId: z.optional(z.nullable(types.number())),
-  isLatest: types.optional(types.boolean()),
-  datasetId: types.optional(types.number()),
+  evaluationGroupId: types.string(),
+  version: types.string(),
+  ancestorId: types.nullable(types.number()),
+  isLatest: types.boolean(),
 });
 
 export function evaluationHistoryFromJSON(

package/src/models/evaluation-list.ts

@@ -8,6 +8,26 @@ import { Result as SafeParseResult } from "../types/fp.js";
 import * as types from "../types/primitives.js";
 import { SDKValidationError } from "./errors/sdk-validation-error.js";
 
+export type EvaluationListCriterion = {
+  id?: string | undefined;
+  /**
+   * Rubric text describing how to score this criterion
+   */
+  criteria: string;
+  /**
+   * Schema field this criterion scores against
+   */
+  evaluationParameter: string;
+  evaluationSteps?: Array<string> | undefined;
+  name?: string | undefined;
+};
+
+export type EvaluationListEvalConfig = {
+  criteria?: Array<EvaluationListCriterion> | undefined;
+  threshold?: number | undefined;
+  model?: string | undefined;
+};
+
 /**
  * Evaluation definition object. This specifies HOW to evaluate, not the results (see PromptEvaluationResult for execution outcomes)
  */
@@ -20,6 +40,10 @@ export type EvaluationListData = {
    * UUID of the prompt family being evaluated (cache this for performance)
    */
   promptGroupId: string;
+  /**
+   * ID of the test dataset used for evaluation (null if not yet linked)
+   */
+  datasetId: number | null;
   /**
    * Evaluation name
    */
@@ -28,10 +52,7 @@ export type EvaluationListData = {
    * Evaluation description
    */
   description: string | null;
-  /**
-   * Evaluation configuration with criteria array
-   */
-  evalConfig: any;
+  evalConfig: EvaluationListEvalConfig;
   /**
    * LLM execution configuration
    */
@@ -59,23 +80,19 @@ export type EvaluationListData = {
   /**
    * UUID grouping all versions of this evaluation together
    */
-  evaluationGroupId?: string | undefined;
+  evaluationGroupId: string;
   /**
    * Semantic version string (e.g., "1.0.0", "1.1.0")
    */
-  version?: string | undefined;
+  version: string;
   /**
    * ID of the previous version in the version chain (null for v1.0.0)
    */
-  ancestorId?: number | null | undefined;
+  ancestorId: number | null;
   /**
    * Whether this is the latest version in the evaluation group
    */
-  isLatest?: boolean | undefined;
-  /**
-   * ID of the test dataset used for evaluation
-   */
-  datasetId?: number | undefined;
+  isLatest: boolean;
 };
 
 /**
@@ -108,6 +125,50 @@ export type EvaluationList = {
   hasPrevious: boolean;
 };
 
+/** @internal */
+export const EvaluationListCriterion$inboundSchema: z.ZodMiniType<
+  EvaluationListCriterion,
+  unknown
+> = z.object({
+  id: types.optional(types.string()),
+  criteria: types.string(),
+  evaluationParameter: types.string(),
+  evaluationSteps: types.optional(z.array(types.string())),
+  name: types.optional(types.string()),
+});
+
+export function evaluationListCriterionFromJSON(
+  jsonString: string,
+): SafeParseResult<EvaluationListCriterion, SDKValidationError> {
+  return safeParse(
+    jsonString,
+    (x) => EvaluationListCriterion$inboundSchema.parse(JSON.parse(x)),
+    `Failed to parse 'EvaluationListCriterion' from JSON`,
+  );
+}
+
+/** @internal */
+export const EvaluationListEvalConfig$inboundSchema: z.ZodMiniType<
+  EvaluationListEvalConfig,
+  unknown
+> = z.object({
+  criteria: types.optional(
+    z.array(z.lazy(() => EvaluationListCriterion$inboundSchema)),
+  ),
+  threshold: types.optional(types.number()),
+  model: types.optional(types.string()),
+});
+
+export function evaluationListEvalConfigFromJSON(
+  jsonString: string,
+): SafeParseResult<EvaluationListEvalConfig, SDKValidationError> {
+  return safeParse(
+    jsonString,
+    (x) => EvaluationListEvalConfig$inboundSchema.parse(JSON.parse(x)),
+    `Failed to parse 'EvaluationListEvalConfig' from JSON`,
+  );
+}
+
 /** @internal */
 export const EvaluationListData$inboundSchema: z.ZodMiniType<
   EvaluationListData,
@@ -115,20 +176,20 @@ export const EvaluationListData$inboundSchema: z.ZodMiniType<
 > = z.object({
   id: types.number(),
   promptGroupId: types.string(),
+  datasetId: types.nullable(types.number()),
   name: types.string(),
   description: types.nullable(types.string()),
-  evalConfig: z.any(),
+  evalConfig: z.lazy(() => EvaluationListEvalConfig$inboundSchema),
   llmConfig: z.any(),
   tags: types.nullable(z.array(types.string())),
   metadata: z.any(),
   createdAt: types.nullable(types.date()),
   updatedAt: types.nullable(types.date()),
   createdBy: types.nullable(types.string()),
-  evaluationGroupId: types.optional(types.string()),
-  version: types.optional(types.string()),
-  ancestorId: z.optional(z.nullable(types.number())),
-  isLatest: types.optional(types.boolean()),
-  datasetId: types.optional(types.number()),
+  evaluationGroupId: types.string(),
+  version: types.string(),
+  ancestorId: types.nullable(types.number()),
+  isLatest: types.boolean(),
 });
 
 export function evaluationListDataFromJSON(

package/src/models/evaluation.ts

@@ -8,6 +8,26 @@ import { Result as SafeParseResult } from "../types/fp.js";
 import * as types from "../types/primitives.js";
 import { SDKValidationError } from "./errors/sdk-validation-error.js";
 
+export type EvaluationCriterion = {
+  id?: string | undefined;
+  /**
+   * Rubric text describing how to score this criterion
+   */
+  criteria: string;
+  /**
+   * Schema field this criterion scores against
+   */
+  evaluationParameter: string;
+  evaluationSteps?: Array<string> | undefined;
+  name?: string | undefined;
+};
+
+export type EvaluationEvalConfig = {
+  criteria?: Array<EvaluationCriterion> | undefined;
+  threshold?: number | undefined;
+  model?: string | undefined;
+};
+
 /**
  * Evaluation definition object. This specifies HOW to evaluate, not the results (see PromptEvaluationResult for execution outcomes)
  */
@@ -20,6 +40,10 @@ export type Evaluation = {
    * UUID of the prompt family being evaluated (cache this for performance)
    */
   promptGroupId: string;
+  /**
+   * ID of the test dataset used for evaluation (null if not yet linked)
+   */
+  datasetId: number | null;
   /**
    * Evaluation name
    */
@@ -28,10 +52,7 @@ export type Evaluation = {
    * Evaluation description
    */
   description: string | null;
-  /**
-   * Evaluation configuration with criteria array
-   */
-  evalConfig: any;
+  evalConfig: EvaluationEvalConfig;
   /**
    * LLM execution configuration
    */
@@ -59,44 +80,84 @@ export type Evaluation = {
   /**
    * UUID grouping all versions of this evaluation together
    */
-  evaluationGroupId?: string | undefined;
+  evaluationGroupId: string;
   /**
    * Semantic version string (e.g., "1.0.0", "1.1.0")
    */
-  version?: string | undefined;
+  version: string;
   /**
    * ID of the previous version in the version chain (null for v1.0.0)
    */
-  ancestorId?: number | null | undefined;
+  ancestorId: number | null;
   /**
    * Whether this is the latest version in the evaluation group
    */
-  isLatest?: boolean | undefined;
-  /**
-   * ID of the test dataset used for evaluation
-   */
-  datasetId?: number | undefined;
+  isLatest: boolean;
 };
 
+/** @internal */
+export const EvaluationCriterion$inboundSchema: z.ZodMiniType<
+  EvaluationCriterion,
+  unknown
+> = z.object({
+  id: types.optional(types.string()),
+  criteria: types.string(),
+  evaluationParameter: types.string(),
+  evaluationSteps: types.optional(z.array(types.string())),
+  name: types.optional(types.string()),
+});
+
+export function evaluationCriterionFromJSON(
+  jsonString: string,
+): SafeParseResult<EvaluationCriterion, SDKValidationError> {
+  return safeParse(
+    jsonString,
+    (x) => EvaluationCriterion$inboundSchema.parse(JSON.parse(x)),
+    `Failed to parse 'EvaluationCriterion' from JSON`,
+  );
+}
+
+/** @internal */
+export const EvaluationEvalConfig$inboundSchema: z.ZodMiniType<
+  EvaluationEvalConfig,
+  unknown
+> = z.object({
+  criteria: types.optional(
+    z.array(z.lazy(() => EvaluationCriterion$inboundSchema)),
+  ),
+  threshold: types.optional(types.number()),
+  model: types.optional(types.string()),
+});
+
+export function evaluationEvalConfigFromJSON(
+  jsonString: string,
+): SafeParseResult<EvaluationEvalConfig, SDKValidationError> {
+  return safeParse(
+    jsonString,
+    (x) => EvaluationEvalConfig$inboundSchema.parse(JSON.parse(x)),
+    `Failed to parse 'EvaluationEvalConfig' from JSON`,
+  );
+}
+
 /** @internal */
 export const Evaluation$inboundSchema: z.ZodMiniType<Evaluation, unknown> = z
   .object({
     id: types.number(),
     promptGroupId: types.string(),
+    datasetId: types.nullable(types.number()),
     name: types.string(),
     description: types.nullable(types.string()),
-    evalConfig: z.any(),
+    evalConfig: z.lazy(() => EvaluationEvalConfig$inboundSchema),
     llmConfig: z.any(),
     tags: types.nullable(z.array(types.string())),
     metadata: z.any(),
     createdAt: types.nullable(types.date()),
     updatedAt: types.nullable(types.date()),
     createdBy: types.nullable(types.string()),
-    evaluationGroupId: types.optional(types.string()),
-    version: types.optional(types.string()),
-    ancestorId: z.optional(z.nullable(types.number())),
-    isLatest: types.optional(types.boolean()),
-    datasetId: types.optional(types.number()),
+    evaluationGroupId: types.string(),
+    version: types.string(),
+    ancestorId: types.nullable(types.number()),
+    isLatest: types.boolean(),
   });
 
 export function evaluationFromJSON(

package/src/models/index.ts

@@ -2,21 +2,15 @@
  * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
  */
 
-export * from "./agent-dataset-export.js";
-export * from "./agent-dataset-item-list.js";
-export * from "./agent-dataset-item.js";
-export * from "./agent-dataset-list.js";
-export * from "./agent-dataset-statistics.js";
-export * from "./agent-dataset.js";
 export * from "./agent-list.js";
+export * from "./agent-success-response.js";
 export * from "./agent.js";
+export * from "./analytics-summary.js";
 export * from "./array-of-id-prompt-id-workspace-id.js";
-export * from "./array-of-path-content-type.js";
-export * from "./array-of-path-data-format.js";
 export * from "./array-of-resource-scope-spans.js";
 export * from "./array-of-role-content.js";
-export * from "./array-of-url-content-type.js";
 export * from "./batch-ingestion-response.js";
+export * from "./criteria-threshold-model.js";
 export * from "./current-password-new-password.js";
 export * from "./dataset-id-evaluation-id-config.js";
 export * from "./dataset-item-id-dataset-item-index-model-column-id.js";
@@ -27,14 +21,11 @@ export * from "./evaluation-list.js";
 export * from "./evaluation-result.js";
 export * from "./evaluation-run.js";
 export * from "./evaluation.js";
-export * from "./files-urls-images.js";
 export * from "./id-name-email.js";
 export * from "./id-organization-id-email.js";
 export * from "./id-prompt-id-workspace-id.js";
 export * from "./id-user-id-role.js";
 export * from "./id-user-id-role2.js";
-export * from "./input-context-expected-output-sequence.js";
-export * from "./input-context-expected-output-sequence2.js";
 export * from "./input-evaluation-id-llm-config.js";
 export * from "./input-llm-config.js";
 export * from "./input-output-total.js";
@@ -45,8 +36,6 @@ export * from "./name-description-is-latest.js";
 export * from "./name-description-metadata.js";
 export * from "./name-description-metadata2.js";
 export * from "./name-description-system-prompt.js";
-export * from "./name-description-tools-definition.js";
-export * from "./name-description-tools-definition2.js";
 export * from "./name-description-version.js";
 export * from "./name-email-avatar-url.js";
 export * from "./name-input-expected-output.js";
@@ -81,16 +70,16 @@ export * from "./prompt-success-response.js";
 export * from "./prompt.js";
 export * from "./provider-config.js";
 export * from "./provider-configs-list.js";
+export * from "./provider-success.js";
 export * from "./schemaany-of.js";
 export * from "./schemaany-of2.js";
-export * from "./schemadefaul12.js";
-export * from "./schemadefaul14.js";
+export * from "./schemadefaul11.js";
+export * from "./schemadefaul13.js";
 export * from "./schemaenum2.js";
-export * from "./schemaitems4.js";
 export * from "./schemanullab4.js";
 export * from "./security.js";
 export * from "./status-summary-completed-at.js";
-export * from "./success-response.js";
+export * from "./success-message.js";
 export * from "./target-prompt-id-target-prompt-group-id-new-name.js";
 export * from "./test-connection-result.js";
 export * from "./trace-id-session-id-name.js";
@@ -101,3 +90,4 @@ export * from "./user-id-role-status.js";
 export * from "./user-id-role.js";
 export * from "./workspace.js";
 export * from "./workspaces-list.js";
+export * from "./ws-success.js";

package/src/models/input-llm-config.ts

@@ -46,7 +46,7 @@ export const InputLlmConfigLlmConfig$outboundSchema: z.ZodMiniType<
   InputLlmConfigLlmConfig$Outbound,
   InputLlmConfigLlmConfig
 > = z.object({
-  model: z._default(z.string(), "gpt-5"),
+  model: z._default(z.string(), "gpt-5.4"),
   temperature: z._default(z.number(), 0.7),
   maxTokens: z._default(z.number(), 4096),
 });

package/src/models/model-temperature-max-tokens.ts

@@ -22,7 +22,7 @@ export const ModelTemperatureMaxTokens$outboundSchema: z.ZodMiniType<
   ModelTemperatureMaxTokens$Outbound,
   ModelTemperatureMaxTokens
 > = z.object({
-  model: z._default(z.string(), "gpt-5"),
+  model: z._default(z.string(), "gpt-5.4"),
   temperature: z._default(z.number(), 0.7),
   maxTokens: z._default(z.number(), 4096),
 });

package/src/models/name-description-eval-config.ts

@@ -3,11 +3,11 @@
  */
 
 import * as z from "zod/v4-mini";
-
-/**
- * Updated evaluation configuration
- */
-export type NameDescriptionEvalConfigEvalConfig = {};
+import {
+  CriteriaThresholdModel,
+  CriteriaThresholdModel$Outbound,
+  CriteriaThresholdModel$outboundSchema,
+} from "./criteria-threshold-model.js";
 
 /**
  * Updated LLM configuration
@@ -28,10 +28,7 @@ export type NameDescriptionEvalConfig = {
    * Updated description
    */
   description?: string | undefined;
-  /**
-   * Updated evaluation configuration
-   */
-  evalConfig?: NameDescriptionEvalConfigEvalConfig | undefined;
+  evalConfig?: CriteriaThresholdModel | undefined;
   /**
    * Updated LLM configuration
    */
@@ -46,25 +43,6 @@ export type NameDescriptionEvalConfig = {
   metadata?: NameDescriptionEvalConfigMetadata | undefined;
 };
 
-/** @internal */
-export type NameDescriptionEvalConfigEvalConfig$Outbound = {};
-
-/** @internal */
-export const NameDescriptionEvalConfigEvalConfig$outboundSchema: z.ZodMiniType<
-  NameDescriptionEvalConfigEvalConfig$Outbound,
-  NameDescriptionEvalConfigEvalConfig
-> = z.object({});
-
-export function nameDescriptionEvalConfigEvalConfigToJSON(
-  nameDescriptionEvalConfigEvalConfig: NameDescriptionEvalConfigEvalConfig,
-): string {
-  return JSON.stringify(
-    NameDescriptionEvalConfigEvalConfig$outboundSchema.parse(
-      nameDescriptionEvalConfigEvalConfig,
-    ),
-  );
-}
-
 /** @internal */
 export type NameDescriptionEvalConfigLlmConfig$Outbound = {};
 
@@ -107,7 +85,7 @@ export function nameDescriptionEvalConfigMetadataToJSON(
 export type NameDescriptionEvalConfig$Outbound = {
   name?: string | undefined;
   description?: string | undefined;
-  evalConfig?: NameDescriptionEvalConfigEvalConfig$Outbound | undefined;
+  evalConfig?: CriteriaThresholdModel$Outbound | undefined;
   llmConfig?: NameDescriptionEvalConfigLlmConfig$Outbound | undefined;
   tags?: Array<string> | undefined;
   metadata?: NameDescriptionEvalConfigMetadata$Outbound | undefined;
@@ -120,9 +98,7 @@ export const NameDescriptionEvalConfig$outboundSchema: z.ZodMiniType<
 > = z.object({
   name: z.optional(z.string()),
   description: z.optional(z.string()),
-  evalConfig: z.optional(
-    z.lazy(() => NameDescriptionEvalConfigEvalConfig$outboundSchema),
-  ),
+  evalConfig: z.optional(CriteriaThresholdModel$outboundSchema),
   llmConfig: z.optional(
     z.lazy(() => NameDescriptionEvalConfigLlmConfig$outboundSchema),
   ),