@wix/evalforge-types 0.91.0 → 0.92.0

package/build/index.mjs

@@ -820,14 +820,65 @@ var EnvironmentSchema = z22.object({
   metaSite: MetaSiteConfigSchema.optional()
 });
 
+// src/scenario/site-setup.ts
+import { z as z23 } from "zod";
+var SiteBootstrapHttpMethodSchema = z23.enum([
+  "get",
+  "post",
+  "put",
+  "patch",
+  "delete"
+]);
+var SiteBootstrapStepSchema = z23.object({
+  label: z23.string().optional(),
+  method: SiteBootstrapHttpMethodSchema,
+  url: z23.string().min(1),
+  body: z23.record(z23.string(), z23.unknown()).optional()
+});
+var SiteBootstrapSchema = z23.object({
+  steps: z23.array(SiteBootstrapStepSchema).default([])
+});
+var SiteBootstrapStepResultSchema = z23.object({
+  label: z23.string().optional(),
+  statusCode: z23.number().int(),
+  ok: z23.boolean(),
+  error: z23.string().optional()
+});
+var SiteBootstrapResultSchema = z23.object({
+  steps: z23.array(SiteBootstrapStepResultSchema)
+});
+var SiteSetupConfigSchema = z23.discriminatedUnion("mode", [
+  z23.object({ mode: z23.literal("none") }),
+  z23.object({
+    mode: z23.literal("clone"),
+    sourceSiteId: z23.string().min(1),
+    bootstrap: SiteBootstrapSchema.optional()
+  }),
+  z23.object({
+    mode: z23.literal("template"),
+    templateId: z23.string().min(1),
+    bootstrap: SiteBootstrapSchema.optional()
+  })
+]);
+var WixSiteSummarySchema = z23.object({
+  id: z23.string(),
+  displayName: z23.string(),
+  url: z23.string().optional()
+});
+var ProvisionedSiteSchema = z23.object({
+  id: z23.string(),
+  url: z23.string().optional(),
+  editorUrl: z23.string().optional()
+});
+
 // src/scenario/test-scenario.ts
-import { z as z25 } from "zod";
+import { z as z26 } from "zod";
 
 // src/assertion/assertion.ts
-import { z as z24 } from "zod";
+import { z as z25 } from "zod";
 
 // src/assertion/build-passed-command.ts
-import { z as z23 } from "zod";
+import { z as z24 } from "zod";
 var ALLOWED_BUILD_COMMANDS = [
   "yarn build",
   "npm run build",
@@ -853,10 +904,10 @@ function parseBuildCommandToArgv(command) {
   return BUILD_COMMAND_ARGV[trimmed];
 }
 var enumTuple = ALLOWED_BUILD_COMMANDS;
-var BuildPassedCommandStringSchema = z23.enum(enumTuple);
+var BuildPassedCommandStringSchema = z24.enum(enumTuple);
 
 // src/assertion/assertion.ts
-var AssertionTypeSchema = z24.enum([
+var AssertionTypeSchema = z25.enum([
   "skill_was_called",
   "tool_called_with_param",
   "build_passed",
@@ -865,61 +916,61 @@ var AssertionTypeSchema = z24.enum([
   "llm_judge",
   "api_call"
 ]);
-var AssertionParameterTypeSchema = z24.enum([
+var AssertionParameterTypeSchema = z25.enum([
   "string",
   "number",
   "boolean"
 ]);
-var AssertionParameterSchema = z24.object({
+var AssertionParameterSchema = z25.object({
   /** Parameter name (used as key in params object) */
-  name: z24.string().min(1),
+  name: z25.string().min(1),
   /** Display label for the parameter */
-  label: z24.string().min(1),
+  label: z25.string().min(1),
   /** Parameter type */
   type: AssertionParameterTypeSchema,
   /** Whether this parameter is required */
-  required: z24.boolean(),
+  required: z25.boolean(),
   /** Default value (optional, used when not provided) */
-  defaultValue: z24.union([z24.string(), z24.number(), z24.boolean()]).optional(),
+  defaultValue: z25.union([z25.string(), z25.number(), z25.boolean()]).optional(),
   /** If true, parameter is hidden by default behind "Show advanced options" */
-  advanced: z24.boolean().optional()
+  advanced: z25.boolean().optional()
 });
-var ScenarioAssertionLinkSchema = z24.object({
+var ScenarioAssertionLinkSchema = z25.object({
   /** ID of the system assertion (e.g., 'system:skill_was_called') */
-  assertionId: z24.string(),
+  assertionId: z25.string(),
   /** Parameter values for this assertion in this scenario */
-  params: z24.record(
-    z24.string(),
-    z24.union([z24.string(), z24.number(), z24.boolean(), z24.null()])
+  params: z25.record(
+    z25.string(),
+    z25.union([z25.string(), z25.number(), z25.boolean(), z25.null()])
   ).optional()
 });
-var SkillWasCalledConfigSchema = z24.object({
+var SkillWasCalledConfigSchema = z25.object({
   /** Names of the skills that must have been called */
-  skillNames: z24.array(z24.string().min(1)).min(1)
+  skillNames: z25.array(z25.string().min(1)).min(1)
 });
-var CostConfigSchema = z24.strictObject({
+var CostConfigSchema = z25.strictObject({
   /** Maximum allowed cost in USD */
-  maxCostUsd: z24.number().positive()
+  maxCostUsd: z25.number().positive()
 });
-var ToolCalledWithParamConfigSchema = z24.strictObject({
+var ToolCalledWithParamConfigSchema = z25.strictObject({
   /** Name of the tool that must have been called */
-  toolName: z24.string().min(1),
+  toolName: z25.string().min(1),
   /** JSON string of key-value pairs for expected parameters (substring match). Optional — when omitted, only checks tool presence. */
-  expectedParams: z24.string().min(1).optional(),
+  expectedParams: z25.string().min(1).optional(),
   /** If true, the matching tool call must also have succeeded (step.success === true) */
-  requireSuccess: z24.boolean().optional()
+  requireSuccess: z25.boolean().optional()
 });
-var BuildPassedConfigSchema = z24.strictObject({
+var BuildPassedConfigSchema = z25.strictObject({
   /** Allowlisted command only (default at runtime: "npm run build") */
   command: BuildPassedCommandStringSchema.optional(),
   /** Expected exit code (default: 0) */
-  expectedExitCode: z24.number().int().optional()
+  expectedExitCode: z25.number().int().optional()
 });
-var TimeConfigSchema = z24.strictObject({
+var TimeConfigSchema = z25.strictObject({
   /** Maximum allowed duration in milliseconds */
-  maxDurationMs: z24.number().int().positive()
+  maxDurationMs: z25.number().int().positive()
 });
-var LlmJudgeConfigSchema = z24.object({
+var LlmJudgeConfigSchema = z25.object({
   /**
    * Prompt template with placeholders:
    * - {{output}}: agent's final output
@@ -930,65 +981,65 @@ var LlmJudgeConfigSchema = z24.object({
    * - {{trace}}: step-by-step trace of tool calls
    * - Custom parameters defined in the parameters array
    */
-  prompt: z24.string().min(1),
+  prompt: z25.string().min(1),
   /** Minimum score to pass (0-10, default 7) */
-  minScore: z24.number().int().min(0).max(10).optional(),
+  minScore: z25.number().int().min(0).max(10).optional(),
   /** Model for the judge (e.g. claude-3-5-haiku-20241022) */
-  model: z24.string().optional(),
+  model: z25.string().optional(),
   /** Max output tokens */
-  maxTokens: z24.number().int().optional(),
+  maxTokens: z25.number().int().optional(),
   /** Temperature (0-1) */
-  temperature: z24.number().min(0).max(1).optional(),
+  temperature: z25.number().min(0).max(1).optional(),
   /** User-defined parameters for this assertion */
-  parameters: z24.array(AssertionParameterSchema).optional()
+  parameters: z25.array(AssertionParameterSchema).optional()
 });
-var ApiCallConfigSchema = z24.strictObject({
+var ApiCallConfigSchema = z25.strictObject({
   /** URL to call */
-  url: z24.string().min(1),
+  url: z25.string().min(1),
   /** HTTP method (default GET) */
-  method: z24.enum(["GET", "POST"]).optional(),
+  method: z25.enum(["GET", "POST"]).optional(),
   /** Request body (JSON string, for POST requests) */
-  requestBody: z24.string().optional(),
+  requestBody: z25.string().optional(),
   /** Expected JSON response to validate against (subset match — extra fields in actual are OK) */
-  expectedResponse: z24.string().min(1),
+  expectedResponse: z25.string().min(1),
   /** Request headers as JSON string of key-value pairs */
-  requestHeaders: z24.string().optional(),
+  requestHeaders: z25.string().optional(),
   /** Request timeout in milliseconds (default 30000) */
-  timeoutMs: z24.number().int().positive().optional()
+  timeoutMs: z25.number().int().positive().optional()
 });
 var AssertionBaseFields = {
   /** When true, the assertion's pass/fail logic is inverted (NOT operator). */
-  negate: z24.boolean().optional()
+  negate: z25.boolean().optional()
 };
 var SkillWasCalledAssertionSchema = SkillWasCalledConfigSchema.extend({
-  type: z24.literal("skill_was_called"),
+  type: z25.literal("skill_was_called"),
   ...AssertionBaseFields
 });
 var ToolCalledWithParamAssertionSchema = ToolCalledWithParamConfigSchema.extend({
-  type: z24.literal("tool_called_with_param"),
+  type: z25.literal("tool_called_with_param"),
   ...AssertionBaseFields
 });
 var BuildPassedAssertionSchema = BuildPassedConfigSchema.extend({
-  type: z24.literal("build_passed"),
+  type: z25.literal("build_passed"),
   ...AssertionBaseFields
 });
 var CostAssertionSchema = CostConfigSchema.extend({
-  type: z24.literal("cost"),
+  type: z25.literal("cost"),
   ...AssertionBaseFields
 });
 var LlmJudgeAssertionSchema = LlmJudgeConfigSchema.extend({
-  type: z24.literal("llm_judge"),
+  type: z25.literal("llm_judge"),
   ...AssertionBaseFields
 });
 var ApiCallAssertionSchema = ApiCallConfigSchema.extend({
-  type: z24.literal("api_call"),
+  type: z25.literal("api_call"),
   ...AssertionBaseFields
 });
 var TimeAssertionSchema = TimeConfigSchema.extend({
-  type: z24.literal("time_limit"),
+  type: z25.literal("time_limit"),
   ...AssertionBaseFields
 });
-var AssertionSchema = z24.union([
+var AssertionSchema = z25.union([
   SkillWasCalledAssertionSchema,
   ToolCalledWithParamAssertionSchema,
   BuildPassedAssertionSchema,
@@ -997,7 +1048,7 @@ var AssertionSchema = z24.union([
   LlmJudgeAssertionSchema,
   ApiCallAssertionSchema
 ]);
-var AssertionConfigSchema = z24.union([
+var AssertionConfigSchema = z25.union([
   LlmJudgeConfigSchema,
   // requires prompt - check first
   SkillWasCalledConfigSchema,
@@ -1012,7 +1063,7 @@ var AssertionConfigSchema = z24.union([
   // requires maxCostUsd, uses strictObject
   BuildPassedConfigSchema,
   // all optional, uses strictObject to reject unknown keys
-  z24.object({})
+  z25.object({})
   // fallback empty config
 ]);
 function validateAssertionConfig(type, config) {
@@ -1258,36 +1309,67 @@ function getSystemAssertion(id) {
 
 // src/scenario/test-scenario.ts
 var MAX_IMAGE_BASE64_LENGTH = 4 * Math.ceil(2 * 1024 * 1024 / 3);
-var TriggerPromptImageSchema = z25.object({
+var TriggerPromptImageSchema = z26.object({
   /** Base64-encoded image data (no data URL prefix) */
-  base64: z25.string().max(MAX_IMAGE_BASE64_LENGTH, "Image exceeds 2 MB size limit"),
+  base64: z26.string().max(MAX_IMAGE_BASE64_LENGTH, "Image exceeds 2 MB size limit"),
   /** MIME type of the image */
-  mediaType: z25.enum(["image/jpeg", "image/png", "image/gif", "image/webp"]),
+  mediaType: z26.enum(["image/jpeg", "image/png", "image/gif", "image/webp"]),
   /** Original filename of the image */
-  name: z25.string()
+  name: z26.string()
 });
-var ExpectedFileSchema = z25.object({
+var ExpectedFileSchema = z26.object({
   /** Relative path where the file should be created */
-  path: z25.string(),
+  path: z26.string(),
   /** Optional expected content */
-  content: z25.string().optional()
+  content: z26.string().optional()
 });
 var TestScenarioSchema = TenantEntitySchema.extend({
   /** The prompt sent to the agent to trigger the task */
-  triggerPrompt: z25.string().min(10),
+  triggerPrompt: z26.string().min(10),
   /** ID of the template to use for this scenario (null = no template) */
-  templateId: z25.string().nullish(),
+  templateId: z26.string().nullish(),
   /** Inline assertions to evaluate for this scenario (legacy) */
-  assertions: z25.array(AssertionSchema).optional(),
+  assertions: z26.array(AssertionSchema).optional(),
   /** IDs of saved assertions to evaluate (from assertions table) - legacy, use assertionLinks */
-  assertionIds: z25.array(z25.string()).optional(),
+  assertionIds: z26.array(z26.string()).optional(),
   /** Linked assertions with per-scenario parameter values */
-  assertionLinks: z25.array(ScenarioAssertionLinkSchema).optional(),
+  assertionLinks: z26.array(ScenarioAssertionLinkSchema).optional(),
   /** Tags for categorisation and filtering */
-  tags: z25.array(z25.string()).optional(),
+  tags: z26.array(z26.string()).optional(),
   /** Base64-encoded images attached to the trigger prompt (max 3) */
-  triggerPromptImages: z25.array(TriggerPromptImageSchema).max(3).optional()
-});
+  triggerPromptImages: z26.array(TriggerPromptImageSchema).max(3).optional(),
+  /** Optional per-scenario Wix site provisioning instructions. Absent ≡ no site. */
+  siteSetup: SiteSetupConfigSchema.optional()
+});
+var SITE_SETUP_EXCLUSIVE_VARIABLES = ["site-id"];
+function extractVariableNamesFromPrompt(prompt) {
+  const names = /* @__PURE__ */ new Set();
+  for (const match of prompt.matchAll(/\{\{([\w-]+)\}\}/g)) {
+    names.add(match[1]);
+  }
+  return [...names];
+}
+function promptUsesSiteSetupExclusiveVariables(prompt) {
+  const names = extractVariableNamesFromPrompt(prompt);
+  return SITE_SETUP_EXCLUSIVE_VARIABLES.some(
+    (exclusive) => names.includes(exclusive)
+  );
+}
+function hasActiveSiteSetup(siteSetup) {
+  return siteSetup?.mode === "clone" || siteSetup?.mode === "template";
+}
+function validateSiteSetupExclusivity(data, ctx) {
+  if (!hasActiveSiteSetup(data.siteSetup)) return;
+  const prompt = data.triggerPrompt;
+  if (prompt === void 0 || !promptUsesSiteSetupExclusiveVariables(prompt)) {
+    return;
+  }
+  ctx.addIssue({
+    code: z26.ZodIssueCode.custom,
+    message: "Site setup and {{site-id}} run variables cannot be used together. Remove {{site-id}} from the trigger prompt or disable site setup.",
+    path: ["triggerPrompt"]
+  });
+}
 function validateBuildPassedParamsInAssertionLinks(links, ctx) {
   if (!links) return;
   for (let i = 0; i < links.length; i++) {
@@ -1297,7 +1379,7 @@ function validateBuildPassedParamsInAssertionLinks(links, ctx) {
     if (cmd === void 0 || cmd === null) continue;
     if (typeof cmd !== "string") {
       ctx.addIssue({
-        code: z25.ZodIssueCode.custom,
+        code: z26.ZodIssueCode.custom,
         message: "build_passed command must be a string",
         path: ["assertionLinks", i, "params", "command"]
       });
@@ -1305,7 +1387,7 @@ function validateBuildPassedParamsInAssertionLinks(links, ctx) {
     }
     if (!isAllowedBuildCommandString(cmd)) {
       ctx.addIssue({
-        code: z25.ZodIssueCode.custom,
+        code: z26.ZodIssueCode.custom,
         message: "Invalid build_passed command. Allowed: yarn build, npm run build, pnpm run build, pnpm build",
         path: ["assertionLinks", i, "params", "command"]
       });
@@ -1320,27 +1402,56 @@ var TestScenarioCreateBaseSchema = TestScenarioSchema.omit({
 });
 var CreateTestScenarioInputSchema = TestScenarioCreateBaseSchema.superRefine((data, ctx) => {
   validateBuildPassedParamsInAssertionLinks(data.assertionLinks, ctx);
+  validateSiteSetupExclusivity(data, ctx);
 });
 var UpdateTestScenarioInputSchema = TestScenarioCreateBaseSchema.partial().superRefine((data, ctx) => {
   if (data.assertionLinks !== void 0) {
     validateBuildPassedParamsInAssertionLinks(data.assertionLinks, ctx);
   }
+  validateSiteSetupExclusivity(data, ctx);
 });
 
+// src/scenario/wix-origin-template-ids.ts
+var WIX_ORIGIN_TEMPLATE_ID_BY_ALIAS = {
+  ecommerce: "e5da13f4-c01e-4b61-a9c7-55dacd961d54",
+  default: "212b41cb-0da6-4401-9c72-7c579e6477a2",
+  blog: "68fc7371-365f-44c6-8467-69d88bfc172e",
+  astrowind: "9e9292c1-1a35-4ba0-8986-d06f2ecb5366",
+  scheduler: "72ade0e3-1871-4c04-ac54-419ca874d9d3",
+  registration: "e5d63bf1-cd06-48eb-ad77-0da9235adcf1",
+  "picasso-studio": "61f05de1-b0ce-4873-b9f5-52241a6fd262",
+  "picasso-ecom": "daa9187d-f010-4eb0-bd49-e658b5a5037a",
+  picasso: "99b9a3c7-82ad-4e1b-9066-e490bb9863af",
+  "ecom-editorless": "738c7c0b-046e-4bf0-87dd-9a06ee5a52c4"
+};
+var GUID_PATTERN = /^[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}$/i;
+function resolveWixOriginTemplateId(templateId) {
+  if (GUID_PATTERN.test(templateId)) {
+    return templateId;
+  }
+  const originTemplateId = WIX_ORIGIN_TEMPLATE_ID_BY_ALIAS[templateId];
+  if (!originTemplateId) {
+    throw new Error(
+      `Unknown Wix site template alias "${templateId}". Use a GUID or one of: ${Object.keys(WIX_ORIGIN_TEMPLATE_ID_BY_ALIAS).join(", ")}`
+    );
+  }
+  return originTemplateId;
+}
+
 // src/scenario/batch-import.ts
-import { z as z26 } from "zod";
+import { z as z27 } from "zod";
 var UUID_REGEX = /^[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}$/i;
-var BatchAssertionLinkSchema = z26.union([
-  z26.string().min(1),
+var BatchAssertionLinkSchema = z27.union([
+  z27.string().min(1),
   ScenarioAssertionLinkSchema
 ]);
-var BatchScenarioEntrySchema = z26.object({
-  name: z26.string().min(1, "name: Required"),
-  description: z26.string().optional().default(""),
-  triggerPrompt: z26.string().min(10, "triggerPrompt: Must be at least 10 characters"),
-  templateId: z26.string().nullish(),
-  tags: z26.array(z26.string()).optional(),
-  assertionLinks: z26.array(BatchAssertionLinkSchema).optional()
+var BatchScenarioEntrySchema = z27.object({
+  name: z27.string().min(1, "name: Required"),
+  description: z27.string().optional().default(""),
+  triggerPrompt: z27.string().min(10, "triggerPrompt: Must be at least 10 characters"),
+  templateId: z27.string().nullish(),
+  tags: z27.array(z27.string()).optional(),
+  assertionLinks: z27.array(BatchAssertionLinkSchema).optional()
 }).superRefine((data, ctx) => {
   if (!data.assertionLinks) return;
   const objectLinks = data.assertionLinks.filter(
@@ -1350,8 +1461,8 @@ var BatchScenarioEntrySchema = z26.object({
     validateBuildPassedParamsInAssertionLinks(objectLinks, ctx);
   }
 });
-var BatchImportPayloadSchema = z26.object({
-  scenarios: z26.array(BatchScenarioEntrySchema).min(1, "scenarios array must contain at least one entry").max(100, "Maximum 100 scenarios per upload")
+var BatchImportPayloadSchema = z27.object({
+  scenarios: z27.array(BatchScenarioEntrySchema).min(1, "scenarios array must contain at least one entry").max(100, "Maximum 100 scenarios per upload")
 });
 var BATCH_IMPORT_LIMITS = {
   MAX_SCENARIOS: 100,
@@ -1373,29 +1484,29 @@ function normalizeBatchAssertionLink(link) {
   }
   return link;
 }
-var BatchResultItemSchema = z26.object({
-  index: z26.number(),
-  name: z26.string(),
-  status: z26.enum(["valid", "invalid"]),
-  id: z26.string().nullable().optional(),
-  errors: z26.array(z26.string()).optional()
-});
-var BatchSummarySchema = z26.object({
-  total: z26.number(),
-  valid: z26.number(),
-  invalid: z26.number(),
-  created: z26.number()
-});
-var BatchImportResponseSchema = z26.object({
+var BatchResultItemSchema = z27.object({
+  index: z27.number(),
+  name: z27.string(),
+  status: z27.enum(["valid", "invalid"]),
+  id: z27.string().nullable().optional(),
+  errors: z27.array(z27.string()).optional()
+});
+var BatchSummarySchema = z27.object({
+  total: z27.number(),
+  valid: z27.number(),
+  invalid: z27.number(),
+  created: z27.number()
+});
+var BatchImportResponseSchema = z27.object({
   summary: BatchSummarySchema,
-  results: z26.array(BatchResultItemSchema)
+  results: z27.array(BatchResultItemSchema)
 });
 
 // src/suite/test-suite.ts
-import { z as z27 } from "zod";
+import { z as z28 } from "zod";
 var TestSuiteSchema = TenantEntitySchema.extend({
   /** IDs of test scenarios in this suite */
-  scenarioIds: z27.array(z27.string())
+  scenarioIds: z28.array(z28.string())
 });
 var CreateTestSuiteInputSchema = TestSuiteSchema.omit({
   id: true,
@@ -1406,21 +1517,21 @@ var CreateTestSuiteInputSchema = TestSuiteSchema.omit({
 var UpdateTestSuiteInputSchema = CreateTestSuiteInputSchema.partial();
 
 // src/evaluation/metrics.ts
-import { z as z28 } from "zod";
-var TokenUsageSchema = z28.object({
-  prompt: z28.number(),
-  completion: z28.number(),
-  total: z28.number()
-});
-var EvalMetricsSchema = z28.object({
-  totalAssertions: z28.number(),
-  passed: z28.number(),
-  failed: z28.number(),
-  skipped: z28.number(),
-  errors: z28.number(),
-  passRate: z28.number(),
-  avgDuration: z28.number(),
-  totalDuration: z28.number()
+import { z as z29 } from "zod";
+var TokenUsageSchema = z29.object({
+  prompt: z29.number(),
+  completion: z29.number(),
+  total: z29.number()
+});
+var EvalMetricsSchema = z29.object({
+  totalAssertions: z29.number(),
+  passed: z29.number(),
+  failed: z29.number(),
+  skipped: z29.number(),
+  errors: z29.number(),
+  passRate: z29.number(),
+  avgDuration: z29.number(),
+  totalDuration: z29.number()
 });
 var EvalStatus = /* @__PURE__ */ ((EvalStatus2) => {
   EvalStatus2["PENDING"] = "pending";
@@ -1430,7 +1541,7 @@ var EvalStatus = /* @__PURE__ */ ((EvalStatus2) => {
   EvalStatus2["CANCELLED"] = "cancelled";
   return EvalStatus2;
 })(EvalStatus || {});
-var EvalStatusSchema = z28.enum(EvalStatus);
+var EvalStatusSchema = z29.enum(EvalStatus);
 var LLMStepType = /* @__PURE__ */ ((LLMStepType2) => {
   LLMStepType2["COMPLETION"] = "completion";
   LLMStepType2["TOOL_USE"] = "tool_use";
@@ -1438,54 +1549,54 @@ var LLMStepType = /* @__PURE__ */ ((LLMStepType2) => {
   LLMStepType2["THINKING"] = "thinking";
   return LLMStepType2;
 })(LLMStepType || {});
-var LLMTraceStepSchema = z28.object({
-  id: z28.string(),
-  stepNumber: z28.number(),
-  type: z28.enum(LLMStepType),
-  model: z28.string(),
-  provider: z28.string(),
-  startedAt: z28.string(),
-  durationMs: z28.number(),
+var LLMTraceStepSchema = z29.object({
+  id: z29.string(),
+  stepNumber: z29.number(),
+  type: z29.enum(LLMStepType),
+  model: z29.string(),
+  provider: z29.string(),
+  startedAt: z29.string(),
+  durationMs: z29.number(),
   tokenUsage: TokenUsageSchema,
-  costUsd: z28.number(),
-  toolName: z28.string().optional(),
-  toolArguments: z28.string().optional(),
-  inputPreview: z28.string().optional(),
-  outputPreview: z28.string().optional(),
-  success: z28.boolean(),
-  error: z28.string().optional(),
-  turnIndex: z28.number().optional()
-});
-var LLMBreakdownStatsSchema = z28.object({
-  count: z28.number(),
-  durationMs: z28.number(),
-  tokens: z28.number(),
-  costUsd: z28.number()
-});
-var LLMTraceSummarySchema = z28.object({
-  totalSteps: z28.number(),
-  totalTurns: z28.number().optional(),
-  totalDurationMs: z28.number(),
+  costUsd: z29.number(),
+  toolName: z29.string().optional(),
+  toolArguments: z29.string().optional(),
+  inputPreview: z29.string().optional(),
+  outputPreview: z29.string().optional(),
+  success: z29.boolean(),
+  error: z29.string().optional(),
+  turnIndex: z29.number().optional()
+});
+var LLMBreakdownStatsSchema = z29.object({
+  count: z29.number(),
+  durationMs: z29.number(),
+  tokens: z29.number(),
+  costUsd: z29.number()
+});
+var LLMTraceSummarySchema = z29.object({
+  totalSteps: z29.number(),
+  totalTurns: z29.number().optional(),
+  totalDurationMs: z29.number(),
   totalTokens: TokenUsageSchema,
-  totalCostUsd: z28.number(),
-  stepTypeBreakdown: z28.record(z28.string(), LLMBreakdownStatsSchema).optional(),
-  modelBreakdown: z28.record(z28.string(), LLMBreakdownStatsSchema),
-  modelsUsed: z28.array(z28.string())
-});
-var LLMTraceSchema = z28.object({
-  id: z28.string(),
-  steps: z28.array(LLMTraceStepSchema),
+  totalCostUsd: z29.number(),
+  stepTypeBreakdown: z29.record(z29.string(), LLMBreakdownStatsSchema).optional(),
+  modelBreakdown: z29.record(z29.string(), LLMBreakdownStatsSchema),
+  modelsUsed: z29.array(z29.string())
+});
+var LLMTraceSchema = z29.object({
+  id: z29.string(),
+  steps: z29.array(LLMTraceStepSchema),
   summary: LLMTraceSummarySchema
 });
 
 // src/evaluation/eval-result.ts
-import { z as z32 } from "zod";
+import { z as z33 } from "zod";
 
 // src/evaluation/eval-run.ts
-import { z as z30 } from "zod";
+import { z as z31 } from "zod";
 
 // src/evaluation/live-trace.ts
-import { z as z29 } from "zod";
+import { z as z30 } from "zod";
 var LiveTraceEventType = /* @__PURE__ */ ((LiveTraceEventType2) => {
   LiveTraceEventType2["THINKING"] = "thinking";
   LiveTraceEventType2["TOOL_USE"] = "tool_use";
@@ -1499,37 +1610,37 @@ var LiveTraceEventType = /* @__PURE__ */ ((LiveTraceEventType2) => {
   LiveTraceEventType2["USER"] = "user";
   return LiveTraceEventType2;
 })(LiveTraceEventType || {});
-var LiveTraceEventSchema = z29.object({
+var LiveTraceEventSchema = z30.object({
   /** The evaluation run ID */
-  evalRunId: z29.string(),
+  evalRunId: z30.string(),
   /** The scenario ID being executed */
-  scenarioId: z29.string(),
+  scenarioId: z30.string(),
   /** The scenario name for display */
-  scenarioName: z29.string(),
+  scenarioName: z30.string(),
   /** The target ID (skill, agent, etc.) */
-  targetId: z29.string(),
+  targetId: z30.string(),
   /** The target name for display */
-  targetName: z29.string(),
+  targetName: z30.string(),
   /** Step number in the current scenario execution */
-  stepNumber: z29.number(),
+  stepNumber: z30.number(),
   /** Type of trace event */
-  type: z29.enum(LiveTraceEventType),
+  type: z30.enum(LiveTraceEventType),
   /** Tool name if this is a tool_use event */
-  toolName: z29.string().optional(),
+  toolName: z30.string().optional(),
   /** Tool arguments preview (truncated JSON) */
-  toolArgs: z29.string().optional(),
+  toolArgs: z30.string().optional(),
   /** Output preview (truncated text) */
-  outputPreview: z29.string().optional(),
+  outputPreview: z30.string().optional(),
   /** File path for file operations */
-  filePath: z29.string().optional(),
+  filePath: z30.string().optional(),
   /** Elapsed time in milliseconds for progress events */
-  elapsedMs: z29.number().optional(),
+  elapsedMs: z30.number().optional(),
   /** Thinking/reasoning text from Claude */
-  thinking: z29.string().optional(),
+  thinking: z30.string().optional(),
   /** Timestamp when this event occurred */
-  timestamp: z29.string(),
+  timestamp: z30.string(),
   /** Whether this is the final event for this scenario */
-  isComplete: z29.boolean()
+  isComplete: z30.boolean()
 });
 var TRACE_EVENT_PREFIX = "TRACE_EVENT:";
 function parseTraceEventLine(line) {
@@ -1558,40 +1669,40 @@ var TriggerType = /* @__PURE__ */ ((TriggerType2) => {
   TriggerType2["SCHEDULED"] = "SCHEDULED";
   return TriggerType2;
 })(TriggerType || {});
-var TriggerMetadataSchema = z30.object({
-  version: z30.string().optional(),
-  resourceUpdated: z30.array(z30.string()).optional(),
-  scheduleId: z30.string().optional()
+var TriggerMetadataSchema = z31.object({
+  version: z31.string().optional(),
+  resourceUpdated: z31.array(z31.string()).optional(),
+  scheduleId: z31.string().optional()
 });
-var TriggerSchema = z30.object({
-  id: z30.string(),
+var TriggerSchema = z31.object({
+  id: z31.string(),
   metadata: TriggerMetadataSchema.optional(),
-  type: z30.nativeEnum(TriggerType)
+  type: z31.nativeEnum(TriggerType)
 });
-var DiffLineTypeSchema = z30.enum(["added", "removed", "unchanged"]);
-var DiffLineSchema = z30.object({
+var DiffLineTypeSchema = z31.enum(["added", "removed", "unchanged"]);
+var DiffLineSchema = z31.object({
   type: DiffLineTypeSchema,
-  content: z30.string(),
-  lineNumber: z30.number()
-});
-var DiffContentSchema = z30.object({
-  path: z30.string(),
-  expected: z30.string(),
-  actual: z30.string(),
-  diffLines: z30.array(DiffLineSchema),
-  renamedFrom: z30.string().optional(),
+  content: z31.string(),
+  lineNumber: z31.number()
+});
+var DiffContentSchema = z31.object({
+  path: z31.string(),
+  expected: z31.string(),
+  actual: z31.string(),
+  diffLines: z31.array(DiffLineSchema),
+  renamedFrom: z31.string().optional(),
   /** Whether this file is an infrastructure/config file (e.g. .claude/settings.json, .mcp.json) */
-  isInfrastructure: z30.boolean().optional()
+  isInfrastructure: z31.boolean().optional()
 });
-var CommandExecutionSchema = z30.object({
-  command: z30.string(),
-  exitCode: z30.number(),
-  output: z30.string().optional(),
-  duration: z30.number()
+var CommandExecutionSchema = z31.object({
+  command: z31.string(),
+  exitCode: z31.number(),
+  output: z31.string().optional(),
+  duration: z31.number()
 });
-var FileModificationSchema = z30.object({
-  path: z30.string(),
-  action: z30.enum(["created", "modified", "deleted"])
+var FileModificationSchema = z31.object({
+  path: z31.string(),
+  action: z31.enum(["created", "modified", "deleted"])
 });
 var TemplateFileStatus = /* @__PURE__ */ ((TemplateFileStatus2) => {
   TemplateFileStatus2["NEW"] = "new";
@@ -1599,58 +1710,58 @@ var TemplateFileStatus = /* @__PURE__ */ ((TemplateFileStatus2) => {
   TemplateFileStatus2["UNCHANGED"] = "unchanged";
   return TemplateFileStatus2;
 })(TemplateFileStatus || {});
-var TemplateFileSchema = z30.object({
+var TemplateFileSchema = z31.object({
   /** Relative path within the template */
-  path: z30.string(),
+  path: z31.string(),
   /** Full file content after execution */
-  content: z30.string(),
+  content: z31.string(),
   /** File status (new, modified, unchanged) */
-  status: z30.enum(["new", "modified", "unchanged"]),
+  status: z31.enum(["new", "modified", "unchanged"]),
   /** Whether this file is an infrastructure/config file (e.g. .claude/settings.json, .mcp.json) */
-  isInfrastructure: z30.boolean().optional()
+  isInfrastructure: z31.boolean().optional()
 });
-var ApiCallSchema = z30.object({
-  endpoint: z30.string(),
-  tokensUsed: z30.number(),
-  duration: z30.number()
+var ApiCallSchema = z31.object({
+  endpoint: z31.string(),
+  tokensUsed: z31.number(),
+  duration: z31.number()
 });
-var ExecutionTraceSchema = z30.object({
-  commands: z30.array(CommandExecutionSchema),
-  filesModified: z30.array(FileModificationSchema),
-  apiCalls: z30.array(ApiCallSchema),
-  totalDuration: z30.number()
+var ExecutionTraceSchema = z31.object({
+  commands: z31.array(CommandExecutionSchema),
+  filesModified: z31.array(FileModificationSchema),
+  apiCalls: z31.array(ApiCallSchema),
+  totalDuration: z31.number()
 });
-var RunAnalysisFindingSchema = z30.object({
-  category: z30.enum([
+var RunAnalysisFindingSchema = z31.object({
+  category: z31.enum([
     "failure_pattern",
     "cost_waste",
     "flakiness",
     "inefficiency",
     "positive"
   ]),
-  severity: z30.enum(["high", "medium", "low"]),
-  description: z30.string(),
-  affectedScenarios: z30.array(z30.string()),
-  recommendation: z30.string().optional()
+  severity: z31.enum(["high", "medium", "low"]),
+  description: z31.string(),
+  affectedScenarios: z31.array(z31.string()),
+  recommendation: z31.string().optional()
 });
-var RunAnalysisSchema = z30.object({
-  generatedAt: z30.string(),
-  summary: z30.string(),
-  findings: z30.array(RunAnalysisFindingSchema)
+var RunAnalysisSchema = z31.object({
+  generatedAt: z31.string(),
+  summary: z31.string(),
+  findings: z31.array(RunAnalysisFindingSchema)
 });
 var EvalRunSchema = TenantEntitySchema.extend({
   /** Agent ID for this run */
-  agentId: z30.string().optional(),
+  agentId: z31.string().optional(),
   /** Preset ID that originated this run (optional) */
-  presetId: z30.string().optional(),
+  presetId: z31.string().optional(),
   /** Scenario IDs to run (always present — resolved server-side from tags when needed) */
-  scenarioIds: z30.array(z30.string()),
+  scenarioIds: z31.array(z31.string()),
   /** Current status */
   status: EvalStatusSchema,
   /** Progress percentage (0-100) */
-  progress: z30.number(),
+  progress: z31.number(),
   /** Results for each scenario/target combination (lazy to break eval-result ↔ eval-run cycle) */
-  results: z30.array(z30.lazy(() => EvalRunResultSchema)),
+  results: z31.array(z31.lazy(() => EvalRunResultSchema)),
   /** Aggregated metrics across all results */
   aggregateMetrics: EvalMetricsSchema,
   /** Aggregated LLM trace summary */
@@ -1658,49 +1769,49 @@ var EvalRunSchema = TenantEntitySchema.extend({
   /** What triggered this run */
   trigger: TriggerSchema.optional(),
   /** When the run started (set when evaluation is triggered) */
-  startedAt: z30.string().optional(),
+  startedAt: z31.string().optional(),
   /** When the run completed */
-  completedAt: z30.string().optional(),
+  completedAt: z31.string().optional(),
   /** Live trace events captured during execution (for playback on results page) */
-  liveTraceEvents: z30.array(LiveTraceEventSchema).optional(),
+  liveTraceEvents: z31.array(LiveTraceEventSchema).optional(),
   /** Remote job ID for tracking execution in Dev Machines */
-  jobId: z30.string().optional(),
+  jobId: z31.string().optional(),
   /** Remote job status from the Dev Machine API (PENDING, RUNNING, COMPLETED, FAILED, CANCELLED) */
-  jobStatus: z30.string().optional(),
+  jobStatus: z31.string().optional(),
   /** Remote job error message if the job failed */
-  jobError: z30.string().optional(),
+  jobError: z31.string().optional(),
   /** Timestamp of the last job status check */
-  jobStatusCheckedAt: z30.string().optional(),
+  jobStatusCheckedAt: z31.string().optional(),
   /** Unified capability IDs */
-  capabilityIds: z30.array(z30.string()).optional(),
+  capabilityIds: z31.array(z31.string()).optional(),
   /** Map of capabilityId to capabilityVersionId for version pinning */
-  capabilityVersions: z30.record(z30.string(), z30.string()).optional(),
+  capabilityVersions: z31.record(z31.string(), z31.string()).optional(),
   /** Tags used to select scenarios for this run (for traceability) */
-  tags: z30.array(z30.string()).optional(),
+  tags: z31.array(z31.string()).optional(),
   /** How many times each scenario is executed within this eval run. Default: 1. Max: 20. */
-  runsPerScenario: z30.number().int().min(1).max(20).optional(),
+  runsPerScenario: z31.number().int().min(1).max(20).optional(),
   /** Variable values to substitute in scenario trigger prompts at runtime */
-  variables: z30.record(z30.string(), z30.string()).optional(),
+  variables: z31.record(z31.string(), z31.string()).optional(),
   /** Snapshot of agent configuration captured at run creation time */
-  agentSnapshot: z30.object({
-    name: z30.string().optional(),
+  agentSnapshot: z31.object({
+    name: z31.string().optional(),
     agentType: AgentTypeSchema.optional(),
     runCommand: AgentRunCommandSchema.optional(),
-    systemPrompt: z30.string().nullable().optional(),
+    systemPrompt: z31.string().nullable().optional(),
     /** @deprecated retained for backward compat with stored snapshots */
     modelConfig: ModelConfigSchema.optional(),
-    config: z30.record(z30.string(), z30.unknown()).optional()
+    config: z31.record(z31.string(), z31.unknown()).optional()
   }).optional(),
   /** UUID linking all runs in a comparison group */
-  comparisonGroupId: z30.string().optional(),
+  comparisonGroupId: z31.string().optional(),
   /** Human-readable label for this variant (e.g., "MCP: Wix Stores") */
-  comparisonLabel: z30.string().optional(),
+  comparisonLabel: z31.string().optional(),
   /** LLM-generated analysis of the completed run */
   runAnalysis: RunAnalysisSchema.optional(),
   /** IDs of folders this run belongs to (read-only, managed via AddRunToFolder / RemoveRunFromFolder) */
-  folderIds: z30.array(z30.string()).optional(),
+  folderIds: z31.array(z31.string()).optional(),
   /** ID of the schedule that triggered this run, if any (read-only) */
-  scheduleId: z30.string().optional()
+  scheduleId: z31.string().optional()
 });
 var CreateEvalRunInputSchema = EvalRunSchema.omit({
   id: true,
@@ -1716,60 +1827,60 @@ var CreateEvalRunInputSchema = EvalRunSchema.omit({
   agentSnapshot: true
 }).extend({
   /** Optional on input — backend resolves from tags when not provided */
-  scenarioIds: z30.array(z30.string()).optional()
+  scenarioIds: z31.array(z31.string()).optional()
 }).refine(
   (data) => data.scenarioIds && data.scenarioIds.length > 0 || data.tags && data.tags.length > 0,
   { message: "Either scenarioIds or tags must be provided" }
 );
-var EvaluationProgressSchema = z30.object({
-  runId: z30.string(),
-  targetId: z30.string(),
-  totalScenarios: z30.number(),
-  completedScenarios: z30.number(),
-  scenarioProgress: z30.array(
-    z30.object({
-      scenarioId: z30.string(),
-      currentStep: z30.string(),
-      error: z30.string().optional()
+var EvaluationProgressSchema = z31.object({
+  runId: z31.string(),
+  targetId: z31.string(),
+  totalScenarios: z31.number(),
+  completedScenarios: z31.number(),
+  scenarioProgress: z31.array(
+    z31.object({
+      scenarioId: z31.string(),
+      currentStep: z31.string(),
+      error: z31.string().optional()
     })
   ),
-  createdAt: z30.number()
-});
-var EvaluationLogSchema = z30.object({
-  runId: z30.string(),
-  scenarioId: z30.string(),
-  log: z30.object({
-    level: z30.enum(["info", "error", "debug"]),
-    message: z30.string().optional(),
-    args: z30.array(z30.any()).optional(),
-    error: z30.string().optional()
+  createdAt: z31.number()
+});
+var EvaluationLogSchema = z31.object({
+  runId: z31.string(),
+  scenarioId: z31.string(),
+  log: z31.object({
+    level: z31.enum(["info", "error", "debug"]),
+    message: z31.string().optional(),
+    args: z31.array(z31.any()).optional(),
+    error: z31.string().optional()
   })
 });
 var LLM_TIMEOUT = 12e4;
 
 // src/evaluation/conversation.ts
-import { z as z31 } from "zod";
-var TextBlockSchema = z31.object({
-  type: z31.literal("text"),
-  text: z31.string()
-});
-var ThinkingBlockSchema = z31.object({
-  type: z31.literal("thinking"),
-  thinking: z31.string()
-});
-var ToolUseBlockSchema = z31.object({
-  type: z31.literal("tool_use"),
-  toolName: z31.string(),
-  toolId: z31.string(),
-  input: z31.unknown()
-});
-var ToolResultBlockSchema = z31.object({
-  type: z31.literal("tool_result"),
-  toolUseId: z31.string(),
-  content: z31.string(),
-  isError: z31.boolean().optional()
-});
-var ConversationBlockSchema = z31.discriminatedUnion("type", [
+import { z as z32 } from "zod";
+var TextBlockSchema = z32.object({
+  type: z32.literal("text"),
+  text: z32.string()
+});
+var ThinkingBlockSchema = z32.object({
+  type: z32.literal("thinking"),
+  thinking: z32.string()
+});
+var ToolUseBlockSchema = z32.object({
+  type: z32.literal("tool_use"),
+  toolName: z32.string(),
+  toolId: z32.string(),
+  input: z32.unknown()
+});
+var ToolResultBlockSchema = z32.object({
+  type: z32.literal("tool_result"),
+  toolUseId: z32.string(),
+  content: z32.string(),
+  isError: z32.boolean().optional()
+});
+var ConversationBlockSchema = z32.discriminatedUnion("type", [
   TextBlockSchema,
   ThinkingBlockSchema,
   ToolUseBlockSchema,
@@ -1780,22 +1891,22 @@ var ConversationMessageRoles = [
   "user",
   "system"
 ];
-var ConversationMessageSchema = z31.object({
-  role: z31.enum(ConversationMessageRoles),
-  content: z31.array(ConversationBlockSchema),
-  timestamp: z31.string()
+var ConversationMessageSchema = z32.object({
+  role: z32.enum(ConversationMessageRoles),
+  content: z32.array(ConversationBlockSchema),
+  timestamp: z32.string()
 });
-var ScenarioConversationSchema = z31.object({
-  id: z31.string(),
-  projectId: z31.string(),
-  evalRunId: z31.string(),
-  resultId: z31.string(),
-  messages: z31.array(ConversationMessageSchema),
-  createdAt: z31.string()
+var ScenarioConversationSchema = z32.object({
+  id: z32.string(),
+  projectId: z32.string(),
+  evalRunId: z32.string(),
+  resultId: z32.string(),
+  messages: z32.array(ConversationMessageSchema),
+  createdAt: z32.string()
 });
-var ConversationResponseSchema = z31.object({
-  messages: z31.array(ConversationMessageSchema),
-  isPartial: z31.boolean()
+var ConversationResponseSchema = z32.object({
+  messages: z32.array(ConversationMessageSchema),
+  isPartial: z32.boolean()
 });
 
 // src/evaluation/eval-result.ts
@@ -1806,98 +1917,98 @@ var AssertionResultStatus = /* @__PURE__ */ ((AssertionResultStatus2) => {
   AssertionResultStatus2["ERROR"] = "error";
   return AssertionResultStatus2;
 })(AssertionResultStatus || {});
-var AssertionResultSchema = z32.object({
-  id: z32.string(),
-  assertionId: z32.string(),
-  assertionType: z32.string(),
-  assertionName: z32.string(),
-  status: z32.enum(AssertionResultStatus),
-  message: z32.string().optional(),
-  expected: z32.string().optional(),
-  actual: z32.string().optional(),
-  duration: z32.number().optional(),
-  details: z32.record(z32.string(), z32.unknown()).optional(),
-  llmTraceSteps: z32.array(LLMTraceStepSchema).optional()
-});
-var EvalRunResultSchema = z32.object({
-  id: z32.string(),
-  targetId: z32.string(),
-  targetName: z32.string().optional(),
+var AssertionResultSchema = z33.object({
+  id: z33.string(),
+  assertionId: z33.string(),
+  assertionType: z33.string(),
+  assertionName: z33.string(),
+  status: z33.enum(AssertionResultStatus),
+  message: z33.string().optional(),
+  expected: z33.string().optional(),
+  actual: z33.string().optional(),
+  duration: z33.number().optional(),
+  details: z33.record(z33.string(), z33.unknown()).optional(),
+  llmTraceSteps: z33.array(LLMTraceStepSchema).optional()
+});
+var EvalRunResultSchema = z33.object({
+  id: z33.string(),
+  targetId: z33.string(),
+  targetName: z33.string().optional(),
   /** SkillVersion ID used for this evaluation (for version tracking) */
-  skillVersionId: z32.string().optional(),
+  skillVersionId: z33.string().optional(),
   /** SkillVersion semver string (e.g., "1.0.0", "1.2.3") for display */
-  skillVersion: z32.string().optional(),
-  scenarioId: z32.string(),
-  scenarioName: z32.string(),
+  skillVersion: z33.string().optional(),
+  scenarioId: z33.string(),
+  scenarioName: z33.string(),
   /** Snapshot of the trigger prompt used during the run (prevents stale display after edits) */
-  triggerPrompt: z32.string().optional(),
+  triggerPrompt: z33.string().optional(),
   modelConfig: ModelConfigSchema.optional(),
-  assertionResults: z32.array(AssertionResultSchema),
+  assertionResults: z33.array(AssertionResultSchema),
   metrics: EvalMetricsSchema.optional(),
-  passed: z32.number(),
-  failed: z32.number(),
-  passRate: z32.number(),
-  duration: z32.number(),
-  outputText: z32.string().optional(),
-  files: z32.array(ExpectedFileSchema).optional(),
-  fileDiffs: z32.array(DiffContentSchema).optional(),
+  passed: z33.number(),
+  failed: z33.number(),
+  passRate: z33.number(),
+  duration: z33.number(),
+  outputText: z33.string().optional(),
+  files: z33.array(ExpectedFileSchema).optional(),
+  fileDiffs: z33.array(DiffContentSchema).optional(),
   /** Full template files after execution with status indicators */
-  templateFiles: z32.array(TemplateFileSchema).optional(),
-  startedAt: z32.string().optional(),
-  completedAt: z32.string().optional(),
+  templateFiles: z33.array(TemplateFileSchema).optional(),
+  startedAt: z33.string().optional(),
+  completedAt: z33.string().optional(),
   llmTrace: LLMTraceSchema.optional(),
   /** Full conversation messages (only present in transit; stripped before DB storage) */
-  conversation: z32.array(ConversationMessageSchema).optional(),
+  conversation: z33.array(ConversationMessageSchema).optional(),
   /** 0-based iteration index when a scenario is run multiple times within a single eval run */
-  iterationIndex: z32.number().int().min(0).optional()
-});
-var PromptResultSchema = z32.object({
-  text: z32.string(),
-  files: z32.array(z32.unknown()).optional(),
-  finishReason: z32.string().optional(),
-  reasoning: z32.string().optional(),
-  reasoningDetails: z32.unknown().optional(),
-  toolCalls: z32.array(z32.unknown()).optional(),
-  toolResults: z32.array(z32.unknown()).optional(),
-  warnings: z32.array(z32.unknown()).optional(),
-  sources: z32.array(z32.unknown()).optional(),
-  steps: z32.array(z32.unknown()),
-  generationTimeMs: z32.number(),
-  prompt: z32.string(),
-  systemPrompt: z32.string(),
-  usage: z32.object({
-    totalTokens: z32.number().optional(),
-    totalMicrocentsSpent: z32.number().optional()
+  iterationIndex: z33.number().int().min(0).optional()
+});
+var PromptResultSchema = z33.object({
+  text: z33.string(),
+  files: z33.array(z33.unknown()).optional(),
+  finishReason: z33.string().optional(),
+  reasoning: z33.string().optional(),
+  reasoningDetails: z33.unknown().optional(),
+  toolCalls: z33.array(z33.unknown()).optional(),
+  toolResults: z33.array(z33.unknown()).optional(),
+  warnings: z33.array(z33.unknown()).optional(),
+  sources: z33.array(z33.unknown()).optional(),
+  steps: z33.array(z33.unknown()),
+  generationTimeMs: z33.number(),
+  prompt: z33.string(),
+  systemPrompt: z33.string(),
+  usage: z33.object({
+    totalTokens: z33.number().optional(),
+    totalMicrocentsSpent: z33.number().optional()
   })
 });
-var EvaluationResultSchema = z32.object({
-  id: z32.string(),
-  runId: z32.string(),
-  timestamp: z32.number(),
+var EvaluationResultSchema = z33.object({
+  id: z33.string(),
+  runId: z33.string(),
+  timestamp: z33.number(),
   promptResult: PromptResultSchema,
-  testResults: z32.array(z32.unknown()),
-  tags: z32.array(z32.string()).optional(),
-  feedback: z32.string().optional(),
-  score: z32.number(),
-  suiteId: z32.string().optional()
-});
-var LeanEvaluationResultSchema = z32.object({
-  id: z32.string(),
-  runId: z32.string(),
-  timestamp: z32.number(),
-  tags: z32.array(z32.string()).optional(),
-  scenarioId: z32.string(),
-  scenarioVersion: z32.number().optional(),
-  targetId: z32.string(),
-  targetVersion: z32.number().optional(),
-  suiteId: z32.string().optional(),
-  score: z32.number(),
-  time: z32.number().optional(),
-  microcentsSpent: z32.number().optional()
+  testResults: z33.array(z33.unknown()),
+  tags: z33.array(z33.string()).optional(),
+  feedback: z33.string().optional(),
+  score: z33.number(),
+  suiteId: z33.string().optional()
+});
+var LeanEvaluationResultSchema = z33.object({
+  id: z33.string(),
+  runId: z33.string(),
+  timestamp: z33.number(),
+  tags: z33.array(z33.string()).optional(),
+  scenarioId: z33.string(),
+  scenarioVersion: z33.number().optional(),
+  targetId: z33.string(),
+  targetVersion: z33.number().optional(),
+  suiteId: z33.string().optional(),
+  score: z33.number(),
+  time: z33.number().optional(),
+  microcentsSpent: z33.number().optional()
 });
 
 // src/evaluation/eval-run-folder.ts
-import { z as z33 } from "zod";
+import { z as z34 } from "zod";
 var EvalRunFolderSchema = TenantEntitySchema.extend({});
 var CreateEvalRunFolderInputSchema = EvalRunFolderSchema.omit({
   id: true,
@@ -1911,26 +2022,26 @@ var UpdateEvalRunFolderInputSchema = EvalRunFolderSchema.omit({
   updatedAt: true,
   deleted: true
 }).partial();
-var EvalRunFolderMembershipSchema = z33.object({
-  folderId: z33.string(),
-  evalRunId: z33.string(),
-  projectId: z33.string(),
-  createdAt: z33.string()
+var EvalRunFolderMembershipSchema = z34.object({
+  folderId: z34.string(),
+  evalRunId: z34.string(),
+  projectId: z34.string(),
+  createdAt: z34.string()
 });
 
 // src/project/project.ts
-import { z as z34 } from "zod";
+import { z as z35 } from "zod";
 var ProjectSchema = BaseEntitySchema.extend({
-  appId: z34.string().optional().describe("The ID of the app in Dev Center"),
-  scenarioTags: z34.array(z34.string()).optional().describe("Project-level tag vocabulary for scenarios"),
+  appId: z35.string().optional().describe("The ID of the app in Dev Center"),
+  scenarioTags: z35.array(z35.string()).optional().describe("Project-level tag vocabulary for scenarios"),
   /** Per-project Wix auth token (write-only — never returned in GET responses). null = clear. */
-  wixAuthToken: z34.string().nullable().optional().describe("Wix auth token for CLI/MCP authentication (encrypted at rest)"),
+  wixAuthToken: z35.string().nullable().optional().describe("Wix auth token for CLI/MCP authentication (encrypted at rest)"),
   /** Per-project Base44 auth file content (write-only — never returned in GET responses). null = clear. */
-  base44AuthFile: z34.string().nullable().optional().describe("Base64-encoded Base44 auth file content (encrypted at rest)"),
+  base44AuthFile: z35.string().nullable().optional().describe("Base64-encoded Base44 auth file content (encrypted at rest)"),
   /** Resolved at runtime from the encrypted Wix auth token */
-  wixAuthEmail: z34.string().optional().describe("Email associated with the Wix auth token (resolved at runtime)"),
+  wixAuthEmail: z35.string().optional().describe("Email associated with the Wix auth token (resolved at runtime)"),
   /** Resolved at runtime from the encrypted Base44 auth file */
-  base44AuthEmail: z34.string().optional().describe("Email from the Base44 auth file (resolved at runtime)")
+  base44AuthEmail: z35.string().optional().describe("Email from the Base44 auth file (resolved at runtime)")
 });
 var CreateProjectInputSchema = ProjectSchema.omit({
   id: true,
@@ -1940,34 +2051,34 @@ var CreateProjectInputSchema = ProjectSchema.omit({
   wixAuthEmail: true,
   base44AuthEmail: true
 }).extend({
-  appId: z34.string().describe(
+  appId: z35.string().describe(
     "Required: The ID of the app in Dev Center for credential scoping"
   )
 });
 var UpdateProjectInputSchema = CreateProjectInputSchema.partial();
 
 // src/template/template.ts
-import { z as z35 } from "zod";
-var SourceFileSchema = z35.object({
-  path: z35.string().min(1),
-  content: z35.string()
+import { z as z36 } from "zod";
+var SourceFileSchema = z36.object({
+  path: z36.string().min(1),
+  content: z36.string()
 });
-var ExtraFileSchema = z35.object({
-  path: z35.string().min(1),
-  content: z35.string().optional(),
+var ExtraFileSchema = z36.object({
+  path: z36.string().min(1),
+  content: z36.string().optional(),
   gitSource: GitHubSourceSchema.optional()
 }).refine((ef) => ef.content !== void 0 || ef.gitSource !== void 0, {
   message: "ExtraFile must have either content or gitSource"
 });
-var TemplateFileEntrySchema = z35.object({
-  path: z35.string().min(1),
-  content: z35.string(),
-  extra: z35.boolean()
+var TemplateFileEntrySchema = z36.object({
+  path: z36.string().min(1),
+  content: z36.string(),
+  extra: z36.boolean()
 });
 var TemplateSchema = TenantEntitySchema.extend({
   source: GitHubSourceSchema.optional(),
-  sourceFiles: z35.array(SourceFileSchema).optional(),
-  extraFiles: z35.array(ExtraFileSchema).optional()
+  sourceFiles: z36.array(SourceFileSchema).optional(),
+  extraFiles: z36.array(ExtraFileSchema).optional()
 });
 var singleSourceKind = (t) => !(t.source && t.sourceFiles?.length);
 var singleSourceKindError = {
@@ -1987,66 +2098,66 @@ var UpdateTemplateInputSchema = TemplateSchema.omit({
 }).partial().refine(singleSourceKind, singleSourceKindError);
 
 // src/agent/agent-config.ts
-import { z as z36 } from "zod";
-var BaseAgentConfigSchema = z36.object({
+import { z as z37 } from "zod";
+var BaseAgentConfigSchema = z37.object({
   /** Model ID (Claude or OpenAI). */
   model: AnyModelSchema.optional(),
   /** Sampling temperature (0–1). */
-  temperature: z36.number().min(0).max(1).optional(),
+  temperature: z37.number().min(0).max(1).optional(),
   /** Max output tokens per turn. */
-  maxTokens: z36.number().int().min(1).optional(),
+  maxTokens: z37.number().int().min(1).optional(),
   /** Number of agentic turns. 0 = unlimited. */
-  maxTurns: z36.number().int().min(0).optional(),
+  maxTurns: z37.number().int().min(0).optional(),
   /** Execution timeout in milliseconds. Overrides the default maxTurns-based calculation. */
-  maxDurationMs: z36.number().int().min(0).optional()
+  maxDurationMs: z37.number().int().min(0).optional()
 });
-var EffortLevelSchema = z36.enum(["low", "medium", "high", "max"]);
+var EffortLevelSchema = z37.enum(["low", "medium", "high", "max"]);
 var ClaudeCodeConfigSchema = BaseAgentConfigSchema.extend({
   /** Extended thinking token budget. */
-  maxThinkingTokens: z36.number().int().min(0).optional(),
+  maxThinkingTokens: z37.number().int().min(0).optional(),
   /** Override the default allowedTools list passed to the SDK. */
-  allowedTools: z36.array(z36.string()).optional(),
+  allowedTools: z37.array(z37.string()).optional(),
   /** Tools to remove from the model's context entirely. */
-  disallowedTools: z36.array(z36.string()).optional(),
+  disallowedTools: z37.array(z37.string()).optional(),
   /** Controls thinking depth: low, medium, high, max. */
   effort: EffortLevelSchema.optional(),
   /** Maximum USD spend per run. Stops execution when reached. */
-  maxBudgetUsd: z36.number().min(0).optional()
+  maxBudgetUsd: z37.number().min(0).optional()
 });
-var PermissionValueSchema = z36.enum(["allow", "deny"]);
-var OpenCodePermissionSchema = z36.record(
-  z36.string(),
-  z36.union([PermissionValueSchema, z36.record(z36.string(), PermissionValueSchema)])
+var PermissionValueSchema = z37.enum(["allow", "deny"]);
+var OpenCodePermissionSchema = z37.record(
+  z37.string(),
+  z37.union([PermissionValueSchema, z37.record(z37.string(), PermissionValueSchema)])
 );
-var ThinkingVariantSchema = z36.enum(["high", "low", "none"]);
+var ThinkingVariantSchema = z37.enum(["high", "low", "none"]);
 var OpenCodeConfigSchema = BaseAgentConfigSchema.extend({
   /** Permission overrides (defaults: allow-all). */
   permission: OpenCodePermissionSchema.optional(),
   /** Maps to `--variant` CLI flag. 'none' omits --thinking entirely. Default: 'high'. */
   thinkingVariant: ThinkingVariantSchema.optional(),
   /** Nucleus sampling (0–1). Alternative to temperature. */
-  topP: z36.number().min(0).max(1).optional()
+  topP: z37.number().min(0).max(1).optional()
 }).omit({ maxTokens: true });
-var ReasoningEffortSchema = z36.enum(["low", "medium", "high"]);
+var ReasoningEffortSchema = z37.enum(["low", "medium", "high"]);
 var SimpleAgentConfigSchema = BaseAgentConfigSchema.extend({
   /** Anthropic thinking budget in tokens. Default: 10 000. */
-  thinkingBudgetTokens: z36.number().int().min(0).optional(),
+  thinkingBudgetTokens: z37.number().int().min(0).optional(),
   /** Nucleus sampling (0–1). Alternative to temperature. */
-  topP: z36.number().min(0).max(1).optional(),
+  topP: z37.number().min(0).max(1).optional(),
   /** Integer seed for deterministic/reproducible results (if model supports it). */
-  seed: z36.number().int().optional(),
+  seed: z37.number().int().optional(),
   /** Stop sequences — model stops when generating any of these strings. */
-  stopSequences: z36.array(z36.string()).optional(),
+  stopSequences: z37.array(z37.string()).optional(),
   /** OpenAI reasoning effort level. Default: 'high'. */
   reasoningEffort: ReasoningEffortSchema.optional(),
   /** Frequency penalty (−2 to 2). Reduces repetition of same tokens. */
-  frequencyPenalty: z36.number().min(-2).max(2).optional(),
+  frequencyPenalty: z37.number().min(-2).max(2).optional(),
   /** Presence penalty (−2 to 2). Encourages topic diversity. */
-  presencePenalty: z36.number().min(-2).max(2).optional()
+  presencePenalty: z37.number().min(-2).max(2).optional()
 });
 
 // src/schedule/eval-schedule.ts
-import { z as z37 } from "zod";
+import { z as z38 } from "zod";
 var FrequencyType = /* @__PURE__ */ ((FrequencyType2) => {
   FrequencyType2["DAILY"] = "daily";
   FrequencyType2["WEEKDAY"] = "weekday";
@@ -2056,31 +2167,31 @@ var FrequencyType = /* @__PURE__ */ ((FrequencyType2) => {
 })(FrequencyType || {});
 var EvalScheduleSchema = TenantEntitySchema.extend({
   /** Whether the schedule is active */
-  enabled: z37.boolean(),
+  enabled: z38.boolean(),
   /** Test suite to run */
-  suiteId: z37.string(),
+  suiteId: z38.string(),
   /** Preset that provides agent + entities for this schedule */
-  presetId: z37.string(),
+  presetId: z38.string(),
   /** How often to run */
-  frequencyType: z37.nativeEnum(FrequencyType),
+  frequencyType: z38.nativeEnum(FrequencyType),
   /** Time of day in 24h format (HH:MM), hours 00-23, minutes 00-59 */
-  timeOfDay: z37.string().regex(/^([01]\d|2[0-3]):[0-5]\d$/),
+  timeOfDay: z38.string().regex(/^([01]\d|2[0-3]):[0-5]\d$/),
   /** Day of week (0=Sun, 6=Sat) for weekly schedules */
-  dayOfWeek: z37.number().min(0).max(6).optional(),
+  dayOfWeek: z38.number().min(0).max(6).optional(),
   /** Day of month (1-31) for monthly schedules */
-  dayOfMonth: z37.number().min(1).max(31).optional(),
+  dayOfMonth: z38.number().min(1).max(31).optional(),
   /** IANA timezone (e.g., 'America/New_York') */
-  timezone: z37.string(),
+  timezone: z38.string(),
   /** ID of the last eval run created by this schedule */
-  lastRunId: z37.string().optional(),
+  lastRunId: z38.string().optional(),
   /** Denormalized status of the last run */
-  lastRunStatus: z37.string().optional(),
+  lastRunStatus: z38.string().optional(),
   /** ISO timestamp of the last run */
-  lastRunAt: z37.string().optional(),
+  lastRunAt: z38.string().optional(),
   /** Next scheduled run time in UTC (pre-computed for efficient querying, set by backend) */
-  nextRunAt: z37.string().optional(),
+  nextRunAt: z38.string().optional(),
   /** Per-scenario variable values forwarded to runs triggered by this schedule (scenarioId → varName → value) */
-  variables: z37.record(z37.string(), z37.record(z37.string(), z37.string())).optional()
+  variables: z38.record(z38.string(), z38.record(z38.string(), z38.string())).optional()
 });
 function isValidTimezone(tz) {
   try {
@@ -2093,14 +2204,14 @@ function isValidTimezone(tz) {
 function validateScheduleFields(data, ctx, options) {
   if (data.frequencyType === "weekly" /* WEEKLY */ && data.dayOfWeek == null) {
     ctx.addIssue({
-      code: z37.ZodIssueCode.custom,
+      code: z38.ZodIssueCode.custom,
       message: "dayOfWeek is required for weekly schedules",
       path: ["dayOfWeek"]
     });
   }
   if (data.frequencyType === "monthly" /* MONTHLY */ && data.dayOfMonth == null) {
     ctx.addIssue({
-      code: z37.ZodIssueCode.custom,
+      code: z38.ZodIssueCode.custom,
       message: "dayOfMonth is required for monthly schedules",
       path: ["dayOfMonth"]
     });
@@ -2108,7 +2219,7 @@ function validateScheduleFields(data, ctx, options) {
   const shouldValidateTz = options.partial ? data.timezone !== void 0 : true;
   if (shouldValidateTz && !isValidTimezone(data.timezone)) {
     ctx.addIssue({
-      code: z37.ZodIssueCode.custom,
+      code: z38.ZodIssueCode.custom,
       message: "Invalid IANA timezone",
       path: ["timezone"]
     });
@@ -2271,6 +2382,7 @@ export {
   PresetSchema,
   ProjectSchema,
   PromptResultSchema,
+  ProvisionedSiteSchema,
   RUN_COMMAND_LABELS,
   ReasoningEffortSchema,
   RuleSchema,
@@ -2278,13 +2390,20 @@ export {
   RunAnalysisFindingSchema,
   RunAnalysisSchema,
   SEMVER_REGEX,
+  SITE_SETUP_EXCLUSIVE_VARIABLES,
   SKILL_FOLDER_NAME_REGEX,
   SYSTEM_ASSERTIONS,
   SYSTEM_ASSERTION_IDS,
   ScenarioAssertionLinkSchema,
   ScenarioConversationSchema,
   SimpleAgentConfigSchema,
+  SiteBootstrapHttpMethodSchema,
+  SiteBootstrapResultSchema,
+  SiteBootstrapSchema,
+  SiteBootstrapStepResultSchema,
+  SiteBootstrapStepSchema,
   SiteConfigTestSchema,
+  SiteSetupConfigSchema,
   SkillFileSchema,
   SkillMetadataSchema,
   SkillSchema,
@@ -2338,6 +2457,7 @@ export {
   UpdateTestScenarioInputSchema,
   UpdateTestSuiteInputSchema,
   VitestTestSchema,
+  WixSiteSummarySchema,
   capabilityToMcp,
   capabilityToRule,
   capabilityToSkill,
@@ -2345,6 +2465,7 @@ export {
   capabilityToSubAgent,
   capabilityVersionToSkillVersion,
   classifyAssertionRef,
+  extractVariableNamesFromPrompt,
   formatTraceEventLine,
   getSystemAssertion,
   getSystemAssertions,
@@ -2357,7 +2478,10 @@ export {
   normalizeModelId,
   parseBuildCommandToArgv,
   parseTraceEventLine,
+  promptUsesSiteSetupExclusiveVariables,
+  resolveWixOriginTemplateId,
   validateAssertionConfig,
-  validateBuildPassedParamsInAssertionLinks
+  validateBuildPassedParamsInAssertionLinks,
+  validateSiteSetupExclusivity
 };
 //# sourceMappingURL=index.mjs.map