@lightsage/mcp 0.2.2 → 0.3.1

package/src/models/createv2evalrequest.ts

@@ -4,43 +4,86 @@
  */
 
 import * as z from "zod";
-import { ClosedEnum } from "../types/enums.js";
-import { V2EvalJudge, V2EvalJudge$zodSchema } from "./v2evaljudge.js";
+import {
+  V2InlineCliReference,
+  V2InlineCliReference$zodSchema,
+} from "./v2inlineclireference.js";
+import {
+  V2InlineMcpDefinition,
+  V2InlineMcpDefinition$zodSchema,
+} from "./v2inlinemcpdefinition.js";
+import {
+  V2SavedCliReference,
+  V2SavedCliReference$zodSchema,
+} from "./v2savedclireference.js";
+import {
+  V2SavedMcpReference,
+  V2SavedMcpReference$zodSchema,
+} from "./v2savedmcpreference.js";
 
-export const CreateV2EvalRequestType = {
-  Workflow: "workflow",
-  Mcp: "mcp",
-  Cli: "cli",
-  Skill: "skill",
-} as const;
-export type CreateV2EvalRequestType = ClosedEnum<
-  typeof CreateV2EvalRequestType
->;
+export type CreateV2EvalRequestCli = V2SavedCliReference | V2InlineCliReference;
 
-export const CreateV2EvalRequestType$zodSchema = z.enum([
-  "workflow",
-  "mcp",
-  "cli",
-  "skill",
+export const CreateV2EvalRequestCli$zodSchema: z.ZodType<
+  CreateV2EvalRequestCli
+> = z.union([
+  V2SavedCliReference$zodSchema,
+  V2InlineCliReference$zodSchema,
+]);
+
+export type CreateV2EvalRequestMcp =
+  | V2InlineMcpDefinition
+  | V2SavedMcpReference;
+
+export const CreateV2EvalRequestMcp$zodSchema: z.ZodType<
+  CreateV2EvalRequestMcp
+> = z.union([
+  V2InlineMcpDefinition$zodSchema,
+  V2SavedMcpReference$zodSchema,
 ]);
 
 export type CreateV2EvalRequest = {
   name: string;
-  type: CreateV2EvalRequestType;
-  prompt: string;
-  judge: V2EvalJudge;
-  mcp_server_id?: string | undefined;
-  cli_id?: string | undefined;
-  skill_id?: string | undefined;
+  prompts: Array<string>;
+  judges: Array<string>;
+  agent?: Array<string> | undefined;
+  persona?: string | undefined;
+  clis?: Array<V2SavedCliReference | V2InlineCliReference> | undefined;
+  mcps?: Array<V2InlineMcpDefinition | V2SavedMcpReference> | undefined;
+  runs?: number | undefined;
+  tags?: Array<string> | undefined;
 };
 
 export const CreateV2EvalRequest$zodSchema: z.ZodType<CreateV2EvalRequest> = z
   .object({
-    cli_id: z.string().optional(),
-    judge: V2EvalJudge$zodSchema,
-    mcp_server_id: z.string().optional(),
+    agent: z.array(z.string()).optional().describe(
+      "Runnable agents from `GET /v2/agents`. An empty array saves the eval without default agents. CLI clients expose repeatable `--agent`.",
+    ),
+    clis: z.array(z.union([
+      V2SavedCliReference$zodSchema,
+      V2InlineCliReference$zodSchema,
+    ])).optional().describe(
+      "Saved CLI references or inline installation commands.",
+    ),
+    judges: z.array(z.string()).describe(
+      "Plain-text success criteria. CLI clients expose this as repeatable `--judge`.",
+    ),
+    mcps: z.array(z.union([
+      V2InlineMcpDefinition$zodSchema,
+      V2SavedMcpReference$zodSchema,
+    ])).optional().describe(
+      "Saved MCP references or inline portable MCP definitions.",
+    ),
     name: z.string(),
-    prompt: z.string(),
-    skill_id: z.string().optional(),
-    type: CreateV2EvalRequestType$zodSchema,
+    persona: z.string().optional().describe(
+      "Optional perspective applied to the whole eval. It is stored as metadata and does not change prompts or multiply attempts.",
+    ),
+    prompts: z.array(z.string()).describe(
+      "Task prompts. CLI clients expose this as repeatable `--prompt`.",
+    ),
+    runs: z.int().default(1).describe(
+      "Independent attempts for every prompt-agent pair.",
+    ),
+    tags: z.array(z.string()).optional().describe(
+      "Optional labels for organizing the eval. CLI clients may expose this string array as a repeatable `--tag` flag.",
+    ),
   });

package/src/models/createv2evalrunrequest.ts

@@ -4,48 +4,120 @@
  */
 
 import * as z from "zod";
+import {
+  V2InlineCliReference,
+  V2InlineCliReference$zodSchema,
+} from "./v2inlineclireference.js";
+import {
+  V2InlineMcpDefinition,
+  V2InlineMcpDefinition$zodSchema,
+} from "./v2inlinemcpdefinition.js";
+import {
+  V2SavedCliReference,
+  V2SavedCliReference$zodSchema,
+} from "./v2savedclireference.js";
+import {
+  V2SavedMcpReference,
+  V2SavedMcpReference$zodSchema,
+} from "./v2savedmcpreference.js";
+
+export type CreateV2EvalRunRequestCli =
+  | V2SavedCliReference
+  | V2InlineCliReference;
+
+export const CreateV2EvalRunRequestCli$zodSchema: z.ZodType<
+  CreateV2EvalRunRequestCli
+> = z.union([
+  V2SavedCliReference$zodSchema,
+  V2InlineCliReference$zodSchema,
+]);
+
+export type CreateV2EvalRunRequestMcp =
+  | V2InlineMcpDefinition
+  | V2SavedMcpReference;
+
+export const CreateV2EvalRunRequestMcp$zodSchema: z.ZodType<
+  CreateV2EvalRunRequestMcp
+> = z.union([
+  V2InlineMcpDefinition$zodSchema,
+  V2SavedMcpReference$zodSchema,
+]);
 
 /**
- * Start an eval run using one of two flat request forms. Saved runs require `eval_id`, `configuration_id`, and `agent_id`. Direct create-and-run requests require `prompt`, at least one item in `judges`, and `agent_id`. Do not combine saved-run and direct-run fields; no separate mode field is required.
+ * Start an eval run using one of two flat request forms. Saved runs require `eval_id` and use saved simulation defaults unless the request replaces them. Direct create-and-run requests require `prompts`, `judges`, and `agent`. No separate mode field is required.
  */
 export type CreateV2EvalRunRequest = {
   eval_id?: string | undefined;
   configuration_id?: string | undefined;
-  agent_id: string;
-  prompt?: string | undefined;
-  judges?: Array<string> | undefined;
+  name?: string | undefined;
+  prompts?: Array<string> | null | undefined;
+  judges?: Array<string> | null | undefined;
+  agent?: Array<string> | null | undefined;
+  persona?: string | undefined;
+  clis?: Array<V2SavedCliReference | V2InlineCliReference> | null | undefined;
+  mcps?: Array<V2InlineMcpDefinition | V2SavedMcpReference> | null | undefined;
+  runs?: number | undefined;
+  tags?: Array<string> | undefined;
   repository?: string | undefined;
   env?: { [k: string]: string } | undefined;
   skills?: Array<string> | undefined;
+  skill_ids?: Array<string> | undefined;
 };
 
 export const CreateV2EvalRunRequest$zodSchema: z.ZodType<
   CreateV2EvalRunRequest
 > = z.object({
-  agent_id: z.string().describe(
-    "Required in both request forms. Runnable agent ID from `GET /v2/agents`.",
+  agent: z.array(z.string()).nullable().optional().describe(
+    "Runnable agents from `GET /v2/agents`. Required unless the saved eval already has agents. CLI clients expose repeatable `--agent`.",
+  ),
+  clis: z.array(z.union([
+    V2SavedCliReference$zodSchema,
+    V2InlineCliReference$zodSchema,
+  ])).nullable().optional().describe(
+    "Saved CLI references or inline installation commands. In saved mode, an explicit empty array disables saved CLIs for this launch.",
   ),
   configuration_id: z.string().optional().describe(
-    "Saved-run form only. Saved runtime configuration; required when `eval_id` is provided.",
+    "Saved-run form only. Optional saved repository, environment, skills, and legacy tool configuration.",
   ),
   env: z.record(z.string(), z.string()).optional().describe(
-    "Direct create-and-run form only. Write-only environment variables; values are stored only in the queued job's private configuration.",
+    "Write-only environment variables for this launch. Values are stored only in the queued job's private configuration.",
   ),
   eval_id: z.string().optional().describe(
     "Saved-run form only. Existing eval definition to run. Omit this field to use the direct create-and-run form.",
   ),
-  judges: z.array(z.string()).optional().describe(
-    "Direct create-and-run form only. Provide at least one success criterion. CLI clients may expose this string array as a repeatable `--judge` flag.",
+  judges: z.array(z.string()).nullable().optional().describe(
+    "Success criteria. Required in direct mode; in saved mode, replaces saved judges for this launch. CLI clients expose repeatable `--judge`.",
+  ),
+  mcps: z.array(z.union([
+    V2InlineMcpDefinition$zodSchema,
+    V2SavedMcpReference$zodSchema,
+  ])).nullable().optional().describe(
+    "Saved MCP references or inline portable definitions. In saved mode, an explicit empty array disables saved MCPs for this launch.",
   ),
-  prompt: z.string().optional().describe(
-    "Direct create-and-run form only. Task prompt; required when `eval_id` is omitted.",
+  name: z.string().optional().describe(
+    "Direct create-and-run form only. Display name for the durable eval. Defaults to the first prompt.",
+  ),
+  persona: z.string().optional().describe(
+    "Optional perspective for every attempt in this launch. It does not modify prompts or add a fan-out dimension.",
+  ),
+  prompts: z.array(z.string()).nullable().optional().describe(
+    "Task prompts. Required in direct mode; in saved mode, replaces saved prompts for this launch. CLI clients expose repeatable `--prompt`.",
   ),
   repository: z.string().optional().describe(
     "Direct create-and-run form only. A saved repository ID, built-in alias such as `nextjs-starter`, or a public GitHub HTTPS URL.",
   ),
+  runs: z.int().optional().describe(
+    "Independent attempts for every prompt-agent pair. Defaults to the saved eval value or 1 for a direct run.",
+  ),
+  skill_ids: z.array(z.string()).optional().describe(
+    "Direct create-and-run form only. Reusable skill IDs from `GET /v2/skills` to install before the agent starts.",
+  ),
   skills: z.array(z.string()).optional().describe(
     "Additional public skill packages to install before the agent starts in either request form, using GitHub `owner/repository` references such as `resend/resend-skills`. CLI clients may expose this string array as a repeatable `--skill` flag.",
   ),
+  tags: z.array(z.string()).optional().describe(
+    "Direct create-and-run form only. Optional labels saved on the created eval. CLI clients may expose this string array as a repeatable `--tag` flag.",
+  ),
 }).describe(
-  "Start an eval run using one of two flat request forms. Saved runs require `eval_id`, `configuration_id`, and `agent_id`. Direct create-and-run requests require `prompt`, at least one item in `judges`, and `agent_id`. Do not combine saved-run and direct-run fields; no separate mode field is required.",
+  "Start an eval run using one of two flat request forms. Saved runs require `eval_id` and use saved simulation defaults unless the request replaces them. Direct create-and-run requests require `prompts`, `judges`, and `agent`. No separate mode field is required.",
 );

package/src/models/updatev2evalrequest.ts

@@ -4,43 +4,76 @@
  */
 
 import * as z from "zod";
-import { ClosedEnum } from "../types/enums.js";
-import { V2EvalJudge, V2EvalJudge$zodSchema } from "./v2evaljudge.js";
+import {
+  V2InlineCliReference,
+  V2InlineCliReference$zodSchema,
+} from "./v2inlineclireference.js";
+import {
+  V2InlineMcpDefinition,
+  V2InlineMcpDefinition$zodSchema,
+} from "./v2inlinemcpdefinition.js";
+import {
+  V2SavedCliReference,
+  V2SavedCliReference$zodSchema,
+} from "./v2savedclireference.js";
+import {
+  V2SavedMcpReference,
+  V2SavedMcpReference$zodSchema,
+} from "./v2savedmcpreference.js";
 
-export const UpdateV2EvalRequestType = {
-  Workflow: "workflow",
-  Mcp: "mcp",
-  Cli: "cli",
-  Skill: "skill",
-} as const;
-export type UpdateV2EvalRequestType = ClosedEnum<
-  typeof UpdateV2EvalRequestType
->;
+export type UpdateV2EvalRequestCli = V2SavedCliReference | V2InlineCliReference;
 
-export const UpdateV2EvalRequestType$zodSchema = z.enum([
-  "workflow",
-  "mcp",
-  "cli",
-  "skill",
+export const UpdateV2EvalRequestCli$zodSchema: z.ZodType<
+  UpdateV2EvalRequestCli
+> = z.union([
+  V2SavedCliReference$zodSchema,
+  V2InlineCliReference$zodSchema,
+]);
+
+export type UpdateV2EvalRequestMcp =
+  | V2InlineMcpDefinition
+  | V2SavedMcpReference;
+
+export const UpdateV2EvalRequestMcp$zodSchema: z.ZodType<
+  UpdateV2EvalRequestMcp
+> = z.union([
+  V2InlineMcpDefinition$zodSchema,
+  V2SavedMcpReference$zodSchema,
 ]);
 
 export type UpdateV2EvalRequest = {
   name?: string | undefined;
-  type?: UpdateV2EvalRequestType | undefined;
-  prompt?: string | undefined;
-  judge?: V2EvalJudge | null | undefined;
-  mcp_server_id?: string | undefined;
-  cli_id?: string | undefined;
-  skill_id?: string | undefined;
+  prompts?: Array<string> | null | undefined;
+  judges?: Array<string> | null | undefined;
+  agent?: Array<string> | null | undefined;
+  persona?: string | undefined;
+  clis?: Array<V2SavedCliReference | V2InlineCliReference> | null | undefined;
+  mcps?: Array<V2InlineMcpDefinition | V2SavedMcpReference> | null | undefined;
+  runs?: number | undefined;
+  tags?: Array<string> | null | undefined;
 };
 
 export const UpdateV2EvalRequest$zodSchema: z.ZodType<UpdateV2EvalRequest> = z
   .object({
-    cli_id: z.string().optional(),
-    judge: V2EvalJudge$zodSchema.nullable().optional(),
-    mcp_server_id: z.string().optional(),
+    agent: z.array(z.string()).nullable().optional(),
+    clis: z.array(z.union([
+      V2SavedCliReference$zodSchema,
+      V2InlineCliReference$zodSchema,
+    ])).nullable().optional(),
+    judges: z.array(z.string()).nullable().optional().describe(
+      "Replacement success criteria for the eval.",
+    ),
+    mcps: z.array(z.union([
+      V2InlineMcpDefinition$zodSchema,
+      V2SavedMcpReference$zodSchema,
+    ])).nullable().optional(),
     name: z.string().optional(),
-    prompt: z.string().optional(),
-    skill_id: z.string().optional(),
-    type: UpdateV2EvalRequestType$zodSchema.optional(),
+    persona: z.string().optional().describe(
+      "Replacement eval persona. Pass null to clear it; omit it to leave the saved value unchanged.",
+    ),
+    prompts: z.array(z.string()).nullable().optional(),
+    runs: z.int().optional(),
+    tags: z.array(z.string()).nullable().optional().describe(
+      "Replacement labels for the eval. Pass an empty array to clear all tags.",
+    ),
   });

package/src/models/v2eval.ts

@@ -4,46 +4,69 @@
  */
 
 import * as z from "zod";
-import { ClosedEnum } from "../types/enums.js";
-import { V2EvalJudge, V2EvalJudge$zodSchema } from "./v2evaljudge.js";
+import {
+  V2InlineCliReference,
+  V2InlineCliReference$zodSchema,
+} from "./v2inlineclireference.js";
+import {
+  V2InlineMcpDefinition,
+  V2InlineMcpDefinition$zodSchema,
+} from "./v2inlinemcpdefinition.js";
+import {
+  V2SavedCliReference,
+  V2SavedCliReference$zodSchema,
+} from "./v2savedclireference.js";
+import {
+  V2SavedMcpReference,
+  V2SavedMcpReference$zodSchema,
+} from "./v2savedmcpreference.js";
 
-export const V2EvalType = {
-  Workflow: "workflow",
-  Mcp: "mcp",
-  Cli: "cli",
-  Skill: "skill",
-} as const;
-export type V2EvalType = ClosedEnum<typeof V2EvalType>;
+export type V2EvalCli = V2SavedCliReference | V2InlineCliReference;
 
-export const V2EvalType$zodSchema = z.enum([
-  "workflow",
-  "mcp",
-  "cli",
-  "skill",
+export const V2EvalCli$zodSchema: z.ZodType<V2EvalCli> = z.union([
+  V2SavedCliReference$zodSchema,
+  V2InlineCliReference$zodSchema,
+]);
+
+export type V2EvalMcp = V2InlineMcpDefinition | V2SavedMcpReference;
+
+export const V2EvalMcp$zodSchema: z.ZodType<V2EvalMcp> = z.union([
+  V2InlineMcpDefinition$zodSchema,
+  V2SavedMcpReference$zodSchema,
 ]);
 
 export type V2Eval = {
   id: string;
   name: string;
-  type: V2EvalType;
-  prompt: string;
-  judge: V2EvalJudge;
-  mcp_server_id?: string | null | undefined;
-  cli_id?: string | null | undefined;
-  skill_id?: string | null | undefined;
+  prompts?: Array<string> | undefined;
+  judges?: Array<string> | undefined;
+  agent?: Array<string> | undefined;
+  persona?: string | null | undefined;
+  clis?: Array<V2SavedCliReference | V2InlineCliReference> | undefined;
+  mcps?: Array<V2InlineMcpDefinition | V2SavedMcpReference> | undefined;
+  runs?: number | undefined;
+  tags?: Array<string> | undefined;
   created_at?: string | null | undefined;
   updated_at?: string | null | undefined;
 };
 
 export const V2Eval$zodSchema: z.ZodType<V2Eval> = z.object({
-  cli_id: z.string().nullable().optional(),
+  agent: z.array(z.string()).optional(),
+  clis: z.array(z.union([
+    V2SavedCliReference$zodSchema,
+    V2InlineCliReference$zodSchema,
+  ])).optional(),
   created_at: z.string().nullable().optional(),
   id: z.string(),
-  judge: V2EvalJudge$zodSchema,
-  mcp_server_id: z.string().nullable().optional(),
+  judges: z.array(z.string()).optional(),
+  mcps: z.array(z.union([
+    V2InlineMcpDefinition$zodSchema,
+    V2SavedMcpReference$zodSchema,
+  ])).optional(),
   name: z.string(),
-  prompt: z.string(),
-  skill_id: z.string().nullable().optional(),
-  type: V2EvalType$zodSchema,
+  persona: z.string().nullable().optional(),
+  prompts: z.array(z.string()).optional(),
+  runs: z.int().default(1),
+  tags: z.array(z.string()).optional(),
   updated_at: z.string().nullable().optional(),
 });

package/src/models/v2evalresult.ts

@@ -8,13 +8,15 @@ import * as z from "zod";
 export type V2EvalResult = {
   id: string;
   eval_id?: string | null | undefined;
-  agent_id?: string | null | undefined;
+  agent?: string | null | undefined;
+  prompt?: string | null | undefined;
+  prompt_index?: number | undefined;
+  repeat_index?: number | undefined;
   passed?: boolean | null | undefined;
   score?: number | null | undefined;
   failure_reasons?: Array<string> | null | undefined;
   latency_ms?: number | null | undefined;
   created_at?: string | null | undefined;
-  prompt?: string | null | undefined;
   output?: string | null | undefined;
   execution_stdout?: string | null | undefined;
   execution_stderr?: string | null | undefined;
@@ -23,8 +25,8 @@ export type V2EvalResult = {
 };
 
 export const V2EvalResult$zodSchema: z.ZodType<V2EvalResult> = z.object({
-  agent_id: z.string().nullable().optional().describe(
-    "Runnable agent ID combining the harness and model.",
+  agent: z.string().nullable().optional().describe(
+    "Runnable agent combining the harness and model.",
   ),
   created_at: z.string().nullable().optional(),
   eval_id: z.string().nullable().optional(),
@@ -37,6 +39,8 @@ export const V2EvalResult$zodSchema: z.ZodType<V2EvalResult> = z.object({
   output: z.string().nullable().optional(),
   passed: z.boolean().nullable().optional(),
   prompt: z.string().nullable().optional(),
+  prompt_index: z.int().default(0),
+  repeat_index: z.int().default(0),
   score: z.number().nullable().optional().describe("Score from 0 to 1."),
   tool_calls: z.array(z.record(z.string(), z.any())).nullable().optional(),
 });

package/src/models/v2evalresultsummary.ts

@@ -8,7 +8,10 @@ import * as z from "zod";
 export type V2EvalResultSummary = {
   id: string;
   eval_id?: string | null | undefined;
-  agent_id?: string | null | undefined;
+  agent?: string | null | undefined;
+  prompt?: string | null | undefined;
+  prompt_index?: number | undefined;
+  repeat_index?: number | undefined;
   passed?: boolean | null | undefined;
   score?: number | null | undefined;
   failure_reasons?: Array<string> | null | undefined;
@@ -18,8 +21,8 @@ export type V2EvalResultSummary = {
 
 export const V2EvalResultSummary$zodSchema: z.ZodType<V2EvalResultSummary> = z
   .object({
-    agent_id: z.string().nullable().optional().describe(
-      "Runnable agent ID combining the harness and model.",
+    agent: z.string().nullable().optional().describe(
+      "Runnable agent combining the harness and model.",
     ),
     created_at: z.string().nullable().optional(),
     eval_id: z.string().nullable().optional(),
@@ -27,5 +30,8 @@ export const V2EvalResultSummary$zodSchema: z.ZodType<V2EvalResultSummary> = z
     id: z.string().describe("Result ID, not an eval job ID."),
     latency_ms: z.int().nullable().optional(),
     passed: z.boolean().nullable().optional(),
+    prompt: z.string().nullable().optional(),
+    prompt_index: z.int().default(0),
+    repeat_index: z.int().default(0),
     score: z.number().nullable().optional().describe("Score from 0 to 1."),
   });

package/src/models/v2evalrun.ts

@@ -15,8 +15,11 @@ export type V2EvalRun = {
   eval_id?: string | null | undefined;
   eval_name?: string | null | undefined;
   configuration_id?: string | null | undefined;
-  agent_id?: string | null | undefined;
-  agent_name?: string | null | undefined;
+  persona?: string | null | undefined;
+  agent?: Array<string> | undefined;
+  prompt_count?: number | undefined;
+  runs?: number | undefined;
+  attempt_count?: number | undefined;
   status: string;
   stage?: string | null | undefined;
   dispatch_id?: string | null | undefined;
@@ -29,11 +32,11 @@ export type V2EvalRun = {
 };
 
 export const V2EvalRun$zodSchema: z.ZodType<V2EvalRun> = z.object({
-  agent_id: z.string().nullable().optional().describe(
-    "Runnable agent ID selected for this run.",
+  agent: z.array(z.string()).optional().describe(
+    "Runnable agents selected for this simulation.",
   ),
-  agent_name: z.string().nullable().optional().describe(
-    "Display name captured for the selected agent.",
+  attempt_count: z.int().default(1).describe(
+    "Total concrete attempts after all fan-out dimensions.",
   ),
   completed_at: z.string().nullable().optional(),
   configuration_id: z.string().nullable().optional(),
@@ -49,7 +52,16 @@ export const V2EvalRun$zodSchema: z.ZodType<V2EvalRun> = z.object({
   ),
   id: z.string().describe("Run ID to use when polling eval-run status."),
   job_id: z.string().describe("Alias of `id` for the durable parent job."),
+  persona: z.string().nullable().optional().describe(
+    "Perspective applied to every attempt in this simulation.",
+  ),
   progress: PublicEvalRunProgress$zodSchema.optional(),
+  prompt_count: z.int().default(1).describe(
+    "Number of prompts included in the simulation.",
+  ),
+  runs: z.int().default(1).describe(
+    "Independent attempts requested for every prompt-agent pair.",
+  ),
   stage: z.string().nullable().optional(),
   started_at: z.string().nullable().optional(),
   status: z.string(),

package/src/models/v2evalrunanalysis.ts

@@ -0,0 +1,59 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ * @generated-id: ce9f97cba8e0
+ */
+
+import * as z from "zod";
+
+export type V2EvalRunAnalysis = {
+  id: string;
+  result_id: string;
+  attempt_id: string;
+  eval_id?: string | null | undefined;
+  agent?: string | null | undefined;
+  prompt?: string | null | undefined;
+  prompt_index?: number | undefined;
+  variant_key?: string | undefined;
+  repeat_index?: number | undefined;
+  analysis_version?: string | null | undefined;
+  analysis?: { [k: string]: any } | undefined;
+  created_at?: string | null | undefined;
+  updated_at?: string | null | undefined;
+};
+
+export const V2EvalRunAnalysis$zodSchema: z.ZodType<V2EvalRunAnalysis> = z
+  .object({
+    agent: z.string().nullable().optional().describe(
+      "Runnable agent used for this concrete attempt.",
+    ),
+    analysis: z.record(z.string(), z.any()).optional().describe(
+      "Persisted eval_run_analysis payload, including verdict, outcome, findings, bottleneck, and analysis sections.",
+    ),
+    analysis_version: z.string().nullable().optional().describe(
+      "Version of the analysis pipeline that produced this payload.",
+    ),
+    attempt_id: z.string().describe(
+      "Child eval-attempt ID used to correlate the analysis with trace events.",
+    ),
+    created_at: z.string().nullable().optional(),
+    eval_id: z.string().nullable().optional().describe(
+      "Public eval definition ID, not the backing executable eval ID.",
+    ),
+    id: z.string().describe("Persisted eval analysis ID."),
+    prompt: z.string().nullable().optional().describe(
+      "Exact task prompt used for this concrete attempt.",
+    ),
+    prompt_index: z.int().default(0).describe(
+      "Zero-based prompt index within the submitted simulation.",
+    ),
+    repeat_index: z.int().default(0).describe(
+      "Zero-based repeat index within the submitted run.",
+    ),
+    result_id: z.string().describe(
+      "Concrete eval result ID analyzed for this attempt.",
+    ),
+    updated_at: z.string().nullable().optional(),
+    variant_key: z.string().default("default").describe(
+      "Runtime variant that produced this result.",
+    ),
+  });

package/src/models/v2evalrunanalysisresponse.ts

@@ -0,0 +1,32 @@
+/*
+ * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
+ * @generated-id: 16648f81799b
+ */
+
+import * as z from "zod";
+import {
+  PublicEvalRunProgress,
+  PublicEvalRunProgress$zodSchema,
+} from "./publicevalrunprogress.js";
+import {
+  V2EvalRunAnalysis,
+  V2EvalRunAnalysis$zodSchema,
+} from "./v2evalrunanalysis.js";
+
+export type V2EvalRunAnalysisResponse = {
+  run_id: string;
+  status: string;
+  progress?: PublicEvalRunProgress | undefined;
+  error?: string | null | undefined;
+  data?: Array<V2EvalRunAnalysis> | undefined;
+};
+
+export const V2EvalRunAnalysisResponse$zodSchema: z.ZodType<
+  V2EvalRunAnalysisResponse
+> = z.object({
+  data: z.array(V2EvalRunAnalysis$zodSchema).optional(),
+  error: z.string().nullable().optional(),
+  progress: PublicEvalRunProgress$zodSchema.optional(),
+  run_id: z.string(),
+  status: z.string(),
+});