@dalmia/calibrate-mcp 0.0.14 → 0.0.16

package/src/models/judgeresult.ts

@@ -18,18 +18,18 @@ export const JudgeResult$zodSchema: z.ZodType<JudgeResult> = z.object({
     "ID of the evaluator that produced this verdict",
   ),
   match: z.boolean().nullable().optional().describe(
-    "Pass/fail verdict, for binary evaluators",
+    "Pass/fail verdict, set for binary evaluators",
   ),
   reasoning: z.string().nullable().optional().describe(
-    "Judge's rationale for this verdict",
+    "The judge's rationale for this verdict",
   ),
   score: z.number().nullable().optional().describe(
-    "Numeric score, for rating evaluators",
+    "Numeric score, set for rating evaluators",
   ),
   value_name: z.string().nullable().optional().describe(
-    "Readable label for the verdict, from the run's rubric",
+    "Readable label for the verdict, taken from the run's rubric",
   ),
   variable_values: z.record(z.string(), z.any()).nullable().optional().describe(
-    "`{{\"{{\"}}var}}` substitutions used for this evaluator on this test case",
+    "Values filled into the evaluator prompt's `{{\"{{\"}}variable}}` placeholders for this test case, keyed by variable name",
   ),
 });

package/src/models/linkevaluatortotaskannotationtaskstaskuuidevaluatorspostop.ts

@@ -0,0 +1,42 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ */
+
+import * as z from "zod";
+import {
+  EvaluatorLinkRequest,
+  EvaluatorLinkRequest$zodSchema,
+} from "./evaluatorlinkrequest.js";
+import {
+  EvaluatorLinkResponse,
+  EvaluatorLinkResponse$zodSchema,
+} from "./evaluatorlinkresponse.js";
+import {
+  HTTPValidationError,
+  HTTPValidationError$zodSchema,
+} from "./httpvalidationerror.js";
+
+export type LinkEvaluatorToTaskAnnotationTasksTaskUuidEvaluatorsPostRequest = {
+  task_uuid: string;
+  xAPIKey?: string | null | undefined;
+  body: EvaluatorLinkRequest;
+};
+
+export const LinkEvaluatorToTaskAnnotationTasksTaskUuidEvaluatorsPostRequest$zodSchema:
+  z.ZodType<LinkEvaluatorToTaskAnnotationTasksTaskUuidEvaluatorsPostRequest> = z
+    .object({
+      body: EvaluatorLinkRequest$zodSchema,
+      task_uuid: z.string().describe("Annotation task to act on"),
+      xAPIKey: z.string().nullable().optional(),
+    });
+
+export type LinkEvaluatorToTaskAnnotationTasksTaskUuidEvaluatorsPostResponse =
+  | EvaluatorLinkResponse
+  | HTTPValidationError;
+
+export const LinkEvaluatorToTaskAnnotationTasksTaskUuidEvaluatorsPostResponse$zodSchema:
+  z.ZodType<LinkEvaluatorToTaskAnnotationTasksTaskUuidEvaluatorsPostResponse> =
+    z.union([
+      EvaluatorLinkResponse$zodSchema,
+      HTTPValidationError$zodSchema,
+    ]);

package/src/models/listagentsagentsgetop.ts

@@ -14,14 +14,12 @@ import {
 
 export type ListAgentsAgentsGetRequest = {
   xAPIKey?: string | null | undefined;
-  xOrgUUID?: string | null | undefined;
 };
 
 export const ListAgentsAgentsGetRequest$zodSchema: z.ZodType<
   ListAgentsAgentsGetRequest
 > = z.object({
   xAPIKey: z.string().nullable().optional(),
-  xOrgUUID: z.string().nullable().optional(),
 });
 
 export type ListAgentsAgentsGetResponse =

package/src/models/listannotationtasksannotationtasksgetop.ts

@@ -0,0 +1,34 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ */
+
+import * as z from "zod";
+import {
+  AnnotationTaskResponse,
+  AnnotationTaskResponse$zodSchema,
+} from "./annotationtaskresponse.js";
+import {
+  HTTPValidationError,
+  HTTPValidationError$zodSchema,
+} from "./httpvalidationerror.js";
+
+export type ListAnnotationTasksAnnotationTasksGetRequest = {
+  xAPIKey?: string | null | undefined;
+};
+
+export const ListAnnotationTasksAnnotationTasksGetRequest$zodSchema: z.ZodType<
+  ListAnnotationTasksAnnotationTasksGetRequest
+> = z.object({
+  xAPIKey: z.string().nullable().optional(),
+});
+
+export type ListAnnotationTasksAnnotationTasksGetResponse =
+  | Array<AnnotationTaskResponse>
+  | HTTPValidationError;
+
+export const ListAnnotationTasksAnnotationTasksGetResponse$zodSchema: z.ZodType<
+  ListAnnotationTasksAnnotationTasksGetResponse
+> = z.union([
+  z.array(AnnotationTaskResponse$zodSchema).describe("Successful Response"),
+  HTTPValidationError$zodSchema,
+]);

package/src/models/listevaluatorsevaluatorsgetop.ts

@@ -0,0 +1,93 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ */
+
+import * as z from "zod";
+import { ClosedEnum } from "../types/enums.js";
+import {
+  HTTPValidationError,
+  HTTPValidationError$zodSchema,
+} from "./httpvalidationerror.js";
+import {
+  RoutersEvaluatorsEvaluatorResponse,
+  RoutersEvaluatorsEvaluatorResponse$zodSchema,
+} from "./routersevaluatorsevaluatorresponse.js";
+
+/**
+ * Filter by what the evaluator judges. Omit for all types
+ */
+export const ListEvaluatorsEvaluatorsGetEvaluatorType = {
+  Tts: "tts",
+  Stt: "stt",
+  Llm: "llm",
+  LlmGeneral: "llm-general",
+  Conversation: "conversation",
+} as const;
+/**
+ * Filter by what the evaluator judges. Omit for all types
+ */
+export type ListEvaluatorsEvaluatorsGetEvaluatorType = ClosedEnum<
+  typeof ListEvaluatorsEvaluatorsGetEvaluatorType
+>;
+
+export const ListEvaluatorsEvaluatorsGetEvaluatorType$zodSchema = z.enum([
+  "tts",
+  "stt",
+  "llm",
+  "llm-general",
+  "conversation",
+]).describe("Filter by what the evaluator judges. Omit for all types");
+
+/**
+ * Filter by modality. Omit for all
+ */
+export const ListEvaluatorsEvaluatorsGetDataType = {
+  Text: "text",
+  Audio: "audio",
+} as const;
+/**
+ * Filter by modality. Omit for all
+ */
+export type ListEvaluatorsEvaluatorsGetDataType = ClosedEnum<
+  typeof ListEvaluatorsEvaluatorsGetDataType
+>;
+
+export const ListEvaluatorsEvaluatorsGetDataType$zodSchema = z.enum([
+  "text",
+  "audio",
+]).describe("Filter by modality. Omit for all");
+
+export type ListEvaluatorsEvaluatorsGetRequest = {
+  evaluator_type?: ListEvaluatorsEvaluatorsGetEvaluatorType | null | undefined;
+  data_type?: ListEvaluatorsEvaluatorsGetDataType | null | undefined;
+  include_defaults?: boolean | undefined;
+  xAPIKey?: string | null | undefined;
+};
+
+export const ListEvaluatorsEvaluatorsGetRequest$zodSchema: z.ZodType<
+  ListEvaluatorsEvaluatorsGetRequest
+> = z.object({
+  data_type: ListEvaluatorsEvaluatorsGetDataType$zodSchema.nullable().optional()
+    .describe("Filter by modality. Omit for all"),
+  evaluator_type: ListEvaluatorsEvaluatorsGetEvaluatorType$zodSchema.nullable()
+    .optional().describe(
+      "Filter by what the evaluator judges. Omit for all types",
+    ),
+  include_defaults: z.boolean().default(true).describe(
+    "When `true`, include the built-in default evaluators alongside the ones you created",
+  ),
+  xAPIKey: z.string().nullable().optional(),
+});
+
+export type ListEvaluatorsEvaluatorsGetResponse =
+  | Array<RoutersEvaluatorsEvaluatorResponse>
+  | HTTPValidationError;
+
+export const ListEvaluatorsEvaluatorsGetResponse$zodSchema: z.ZodType<
+  ListEvaluatorsEvaluatorsGetResponse
+> = z.union([
+  z.array(RoutersEvaluatorsEvaluatorResponse$zodSchema).describe(
+    "Successful Response",
+  ),
+  HTTPValidationError$zodSchema,
+]);

package/src/models/listteststestsgetop.ts

@@ -12,16 +12,12 @@ import {
   RoutersTestsTestResponse$zodSchema,
 } from "./routersteststestresponse.js";
 
-export type ListTestsTestsGetRequest = {
-  xAPIKey?: string | null | undefined;
-  xOrgUUID?: string | null | undefined;
-};
+export type ListTestsTestsGetRequest = { xAPIKey?: string | null | undefined };
 
 export const ListTestsTestsGetRequest$zodSchema: z.ZodType<
   ListTestsTestsGetRequest
 > = z.object({
   xAPIKey: z.string().nullable().optional(),
-  xOrgUUID: z.string().nullable().optional(),
 });
 
 export type ListTestsTestsGetResponse =

package/src/models/modelresult.ts

@@ -0,0 +1,50 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ */
+
+import * as z from "zod";
+import { TestCaseResult, TestCaseResult$zodSchema } from "./testcaseresult.js";
+
+export type ModelResult = {
+  model: string;
+  success?: boolean | null | undefined;
+  message: string;
+  total_tests?: number | null | undefined;
+  passed?: number | null | undefined;
+  failed?: number | null | undefined;
+  evaluator_summary?: Array<{ [k: string]: any }> | null | undefined;
+  test_results?: Array<TestCaseResult> | null | undefined;
+  latency_ms?: { [k: string]: any } | null | undefined;
+  cost?: { [k: string]: any } | null | undefined;
+  total_tokens?: { [k: string]: any } | null | undefined;
+};
+
+export const ModelResult$zodSchema: z.ZodType<ModelResult> = z.object({
+  cost: z.record(z.string(), z.any()).nullable().optional().describe(
+    "Aggregated cost as `{mean, min, max, count}` (USD)",
+  ),
+  evaluator_summary: z.array(z.record(z.string(), z.any())).nullable()
+    .optional().describe("Aggregate summary for each evaluator for this model"),
+  failed: z.int().nullable().optional().describe(
+    "Number of test cases that failed",
+  ),
+  latency_ms: z.record(z.string(), z.any()).nullable().optional().describe(
+    "Aggregated latency in milliseconds, as `{p50, p95, p99, count}`",
+  ),
+  message: z.string().describe("Status or result message for this model"),
+  model: z.string().describe("Model name these results are for"),
+  passed: z.int().nullable().optional().describe(
+    "Number of test cases that passed",
+  ),
+  success: z.boolean().nullable().optional().describe(
+    "Whether this model's run succeeded",
+  ),
+  test_results: z.array(TestCaseResult$zodSchema).nullable().optional()
+    .describe("Results for each test case for this model"),
+  total_tests: z.int().nullable().optional().describe(
+    "Total test cases for this model",
+  ),
+  total_tokens: z.record(z.string(), z.any()).nullable().optional().describe(
+    "Aggregated token usage as `{mean, min, max, count}`",
+  ),
+});

package/src/models/outputconfig.ts

@@ -0,0 +1,19 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ */
+
+import * as z from "zod";
+import {
+  OutputScaleEntry,
+  OutputScaleEntry$zodSchema,
+} from "./outputscaleentry.js";
+
+export type OutputConfig = {
+  scale?: Array<OutputScaleEntry> | null | undefined;
+};
+
+export const OutputConfig$zodSchema: z.ZodType<OutputConfig> = z.object({
+  scale: z.array(OutputScaleEntry$zodSchema).nullable().optional().describe(
+    "The ordered scale points that make up the rubric, each with its label",
+  ),
+});

package/src/models/outputscaleentry.ts

@@ -0,0 +1,35 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ */
+
+import * as z from "zod";
+
+/**
+ * One entry in an evaluator's output_config.scale. `value` is bool|number|string
+ *
+ * @remarks
+ * depending on output_type.
+ */
+export type OutputScaleEntry = {
+  value?: any | undefined;
+  name: string;
+  description?: string | null | undefined;
+  color?: string | null | undefined;
+};
+
+export const OutputScaleEntry$zodSchema: z.ZodType<OutputScaleEntry> = z.object(
+  {
+    color: z.string().nullable().optional().describe(
+      "Color to show for this level. Omit for the default",
+    ),
+    description: z.string().nullable().optional().describe(
+      "Rubric text for this level, added to the judge prompt. Omit to leave this level undescribed",
+    ),
+    name: z.string().describe("Short label for this scale point"),
+    value: z.any().optional().describe(
+      "The value for this scale point. Use a boolean for a `binary` evaluator, a number for a `rating` one",
+    ),
+  },
+).describe(
+  "One entry in an evaluator's output_config.scale. `value` is bool|number|string\ndepending on output_type.",
+);

package/src/models/paginationmeta.ts

@@ -0,0 +1,13 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ */
+
+import * as z from "zod";
+
+export type PaginationMeta = { total: number; limit: number; offset: number };
+
+export const PaginationMeta$zodSchema: z.ZodType<PaginationMeta> = z.object({
+  limit: z.int().describe("Maximum number of items returned in this page"),
+  offset: z.int().describe("Number of items skipped before this page"),
+  total: z.int().describe("Total number of items that match, across all pages"),
+});

package/src/models/resolveagentnamesagentsresolvepostop.ts

@@ -18,7 +18,6 @@ import {
 
 export type ResolveAgentNamesAgentsResolvePostRequest = {
   xAPIKey?: string | null | undefined;
-  xOrgUUID?: string | null | undefined;
   body: ResolveAgentNamesRequest;
 };
 
@@ -27,7 +26,6 @@ export const ResolveAgentNamesAgentsResolvePostRequest$zodSchema: z.ZodType<
 > = z.object({
   body: ResolveAgentNamesRequest$zodSchema,
   xAPIKey: z.string().nullable().optional(),
-  xOrgUUID: z.string().nullable().optional(),
 });
 
 export type ResolveAgentNamesAgentsResolvePostResponse =

package/src/models/routersagentteststestresponse.ts

@@ -0,0 +1,65 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ */
+
+import * as z from "zod";
+import { ClosedEnum } from "../types/enums.js";
+
+/**
+ * What the test checks:
+ *
+ * @remarks
+ * - `response`: judges the reply the agent generates
+ * - `tool_call`: checks the tool calls the agent makes
+ * - `conversation`: judges the full conversation after the agent replies
+ */
+export const RoutersAgentTestsTestResponseType = {
+  Response: "response",
+  ToolCall: "tool_call",
+  Conversation: "conversation",
+} as const;
+/**
+ * What the test checks:
+ *
+ * @remarks
+ * - `response`: judges the reply the agent generates
+ * - `tool_call`: checks the tool calls the agent makes
+ * - `conversation`: judges the full conversation after the agent replies
+ */
+export type RoutersAgentTestsTestResponseType = ClosedEnum<
+  typeof RoutersAgentTestsTestResponseType
+>;
+
+export const RoutersAgentTestsTestResponseType$zodSchema = z.enum([
+  "response",
+  "tool_call",
+  "conversation",
+]).describe(
+  "What the test checks:\n- `response`: judges the reply the agent generates\n- `tool_call`: checks the tool calls the agent makes\n- `conversation`: judges the full conversation after the agent replies",
+);
+
+export type RoutersAgentTestsTestResponse = {
+  uuid: string;
+  name: string;
+  type: RoutersAgentTestsTestResponseType;
+  config?: { [k: string]: any } | null | undefined;
+  created_at: string;
+  updated_at: string;
+};
+
+export const RoutersAgentTestsTestResponse$zodSchema: z.ZodType<
+  RoutersAgentTestsTestResponse
+> = z.object({
+  config: z.record(z.string(), z.any()).nullable().optional().describe(
+    "Test configuration",
+  ),
+  created_at: z.string().describe("When the test was created (ISO 8601 UTC)"),
+  name: z.string().describe("Name of the test"),
+  type: RoutersAgentTestsTestResponseType$zodSchema.describe(
+    "What the test checks:\n- `response`: judges the reply the agent generates\n- `tool_call`: checks the tool calls the agent makes\n- `conversation`: judges the full conversation after the agent replies",
+  ),
+  updated_at: z.string().describe(
+    "When the test was last updated (ISO 8601 UTC)",
+  ),
+  uuid: z.string().describe("Test ID"),
+});

package/src/models/routersevaluatorsevaluatorresponse.ts

@@ -0,0 +1,167 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ */
+
+import * as z from "zod";
+import { ClosedEnum } from "../types/enums.js";
+import {
+  EvaluatorVersionResponse,
+  EvaluatorVersionResponse$zodSchema,
+} from "./evaluatorversionresponse.js";
+
+/**
+ * What the evaluator judges:
+ *
+ * @remarks
+ *
+ * - `tts`: TTS audio
+ * - `stt`: one transcript
+ * - `llm`: a reply with its conversation history
+ * - `llm-general`: a standalone input and output pair
+ * - `conversation`: a full conversation
+ */
+export const RoutersEvaluatorsEvaluatorResponseEvaluatorType = {
+  Tts: "tts",
+  Stt: "stt",
+  Llm: "llm",
+  LlmGeneral: "llm-general",
+  Conversation: "conversation",
+} as const;
+/**
+ * What the evaluator judges:
+ *
+ * @remarks
+ *
+ * - `tts`: TTS audio
+ * - `stt`: one transcript
+ * - `llm`: a reply with its conversation history
+ * - `llm-general`: a standalone input and output pair
+ * - `conversation`: a full conversation
+ */
+export type RoutersEvaluatorsEvaluatorResponseEvaluatorType = ClosedEnum<
+  typeof RoutersEvaluatorsEvaluatorResponseEvaluatorType
+>;
+
+export const RoutersEvaluatorsEvaluatorResponseEvaluatorType$zodSchema = z.enum(
+  [
+    "tts",
+    "stt",
+    "llm",
+    "llm-general",
+    "conversation",
+  ],
+).describe(
+  "What the evaluator judges:\n\n- `tts`: TTS audio\n- `stt`: one transcript\n- `llm`: a reply with its conversation history\n- `llm-general`: a standalone input and output pair\n- `conversation`: a full conversation\n",
+);
+
+/**
+ * The modality the judge reads:
+ *
+ * @remarks
+ *
+ * - `text`
+ * - `audio`
+ */
+export const RoutersEvaluatorsEvaluatorResponseDataType = {
+  Text: "text",
+  Audio: "audio",
+} as const;
+/**
+ * The modality the judge reads:
+ *
+ * @remarks
+ *
+ * - `text`
+ * - `audio`
+ */
+export type RoutersEvaluatorsEvaluatorResponseDataType = ClosedEnum<
+  typeof RoutersEvaluatorsEvaluatorResponseDataType
+>;
+
+export const RoutersEvaluatorsEvaluatorResponseDataType$zodSchema = z.enum([
+  "text",
+  "audio",
+]).describe("The modality the judge reads:\n\n- `text`\n- `audio`\n");
+
+/**
+ * How the evaluator scores:
+ *
+ * @remarks
+ *
+ * - `binary`: pass or fail
+ * - `rating`: a numeric score, using the scale in `output_config`
+ */
+export const RoutersEvaluatorsEvaluatorResponseOutputType = {
+  Binary: "binary",
+  Rating: "rating",
+} as const;
+/**
+ * How the evaluator scores:
+ *
+ * @remarks
+ *
+ * - `binary`: pass or fail
+ * - `rating`: a numeric score, using the scale in `output_config`
+ */
+export type RoutersEvaluatorsEvaluatorResponseOutputType = ClosedEnum<
+  typeof RoutersEvaluatorsEvaluatorResponseOutputType
+>;
+
+export const RoutersEvaluatorsEvaluatorResponseOutputType$zodSchema = z.enum([
+  "binary",
+  "rating",
+]).describe(
+  "How the evaluator scores:\n\n- `binary`: pass or fail\n- `rating`: a numeric score, using the scale in `output_config`\n",
+);
+
+export type RoutersEvaluatorsEvaluatorResponse = {
+  uuid: string;
+  name: string;
+  description?: string | null | undefined;
+  evaluator_type: RoutersEvaluatorsEvaluatorResponseEvaluatorType;
+  data_type: RoutersEvaluatorsEvaluatorResponseDataType;
+  output_type: RoutersEvaluatorsEvaluatorResponseOutputType;
+  is_default: boolean;
+  slug?: string | null | undefined;
+  live_version_id?: string | null | undefined;
+  created_at: string;
+  updated_at: string;
+  live_version?: EvaluatorVersionResponse | null | undefined;
+};
+
+export const RoutersEvaluatorsEvaluatorResponse$zodSchema: z.ZodType<
+  RoutersEvaluatorsEvaluatorResponse
+> = z.object({
+  created_at: z.string().describe(
+    "When the evaluator was created (ISO 8601 UTC)",
+  ),
+  data_type: RoutersEvaluatorsEvaluatorResponseDataType$zodSchema.describe(
+    "The modality the judge reads:\n\n- `text`\n- `audio`\n",
+  ),
+  description: z.string().nullable().optional().describe(
+    "What the evaluator checks",
+  ),
+  evaluator_type: RoutersEvaluatorsEvaluatorResponseEvaluatorType$zodSchema
+    .describe(
+      "What the evaluator judges:\n\n- `tts`: TTS audio\n- `stt`: one transcript\n- `llm`: a reply with its conversation history\n- `llm-general`: a standalone input and output pair\n- `conversation`: a full conversation\n",
+    ),
+  is_default: z.boolean().describe(
+    "True for a built-in default evaluator, which you can't edit. False for an evaluator you created, which you can edit and add versions to",
+  ),
+  live_version: EvaluatorVersionResponse$zodSchema.nullable().optional()
+    .describe("The version that is currently live"),
+  live_version_id: z.string().nullable().optional().describe(
+    "ID of the version that is currently live",
+  ),
+  name: z.string().describe("Evaluator name"),
+  output_type: RoutersEvaluatorsEvaluatorResponseOutputType$zodSchema.describe(
+    "How the evaluator scores:\n\n- `binary`: pass or fail\n- `rating`: a numeric score, using the scale in `output_config`\n",
+  ),
+  slug: z.string().nullable().optional().describe(
+    "Stable slug for a built-in default evaluator",
+  ),
+  updated_at: z.string().describe(
+    "When the evaluator was last updated (ISO 8601 UTC)",
+  ),
+  uuid: z.string().describe("Evaluator ID"),
+});

package/src/models/routerstestsevaluatorref.ts

@@ -14,6 +14,6 @@ export const RoutersTestsEvaluatorRef$zodSchema: z.ZodType<
 > = z.object({
   evaluator_uuid: z.string().describe("Evaluator to attach to the test"),
   variable_values: z.record(z.string(), z.any()).nullable().optional().describe(
-    "Values for the evaluator's `{{\"{{\"}}placeholder}}` variables, pinned per test. Omit to inherit the evaluator version's defaults",
+    "Values for the evaluator's `{{\"{{\"}}placeholder}}` variables, pinned on this test. Omit to inherit the evaluator version's defaults",
   ),
 });

package/src/models/routersteststestresponse.ts

@@ -37,7 +37,7 @@ export const RoutersTestsTestResponseType$zodSchema = z.enum([
   "tool_call",
   "conversation",
 ]).describe(
-  "What the test judges:\n\n- `response`: judges the generated reply\n- `tool_call`: diffs the generated tool calls\n- `conversation`: judges the full conversation",
+  "What the test judges:\n\n- `response`: judges the generated reply\n- `tool_call`: diffs the generated tool calls\n- `conversation`: judges the full conversation\n",
 );
 
 export type RoutersTestsTestResponse = {
@@ -54,20 +54,18 @@ export const RoutersTestsTestResponse$zodSchema: z.ZodType<
   RoutersTestsTestResponse
 > = z.object({
   config: z.record(z.string(), z.any()).nullable().optional().describe(
-    "Config for the test (`history`, `evaluation`, optional `settings`)",
-  ),
-  created_at: z.string().describe(
-    "Timestamp when the test was created (ISO 8601 UTC)",
+    "The stored config: `history`, `evaluation`, and an optional `settings`",
   ),
+  created_at: z.string().describe("When the test was created (ISO 8601 UTC)"),
   evaluators: z.array(z.record(z.string(), z.any())).optional().describe(
     "Linked evaluators, resolved to their current live version at read time",
   ),
   name: z.string().describe("Name of the test"),
   type: RoutersTestsTestResponseType$zodSchema.describe(
-    "What the test judges:\n\n- `response`: judges the generated reply\n- `tool_call`: diffs the generated tool calls\n- `conversation`: judges the full conversation",
+    "What the test judges:\n\n- `response`: judges the generated reply\n- `tool_call`: diffs the generated tool calls\n- `conversation`: judges the full conversation\n",
   ),
   updated_at: z.string().describe(
-    "Timestamp when the test was last updated (ISO 8601 UTC)",
+    "When the test was last updated (ISO 8601 UTC)",
   ),
   uuid: z.string().describe("Unique ID for the test"),
 });