@wix/evalforge-types 0.67.0 → 0.69.0

package/build/index.mjs

@@ -1161,7 +1161,19 @@ var EvalRunSchema = TenantEntitySchema.extend({
   /** Tags used to select scenarios for this run (for traceability) */
   tags: z28.array(z28.string()).optional(),
   /** How many times each scenario is executed within this eval run. Default: 1. Max: 20. */
-  runsPerScenario: z28.number().int().min(1).max(20).optional()
+  runsPerScenario: z28.number().int().min(1).max(20).optional(),
+  /** Snapshot of agent configuration captured at run creation time */
+  agentSnapshot: z28.object({
+    name: z28.string().optional(),
+    agentType: AgentTypeSchema.optional(),
+    runCommand: AgentRunCommandSchema.optional(),
+    systemPrompt: z28.string().nullable().optional(),
+    modelConfig: ModelConfigSchema.optional()
+  }).optional(),
+  /** UUID linking all runs in a comparison group */
+  comparisonGroupId: z28.string().optional(),
+  /** Human-readable label for this variant (e.g., "MCP: Wix Stores") */
+  comparisonLabel: z28.string().optional()
 });
 var CreateEvalRunInputSchema = EvalRunSchema.omit({
   id: true,
@@ -1173,7 +1185,8 @@ var CreateEvalRunInputSchema = EvalRunSchema.omit({
   aggregateMetrics: true,
   startedAt: true,
   completedAt: true,
-  scenarioIds: true
+  scenarioIds: true,
+  agentSnapshot: true
 }).extend({
   /** Optional on input — backend resolves from tags when not provided */
   scenarioIds: z28.array(z28.string()).optional()
@@ -1285,6 +1298,8 @@ var EvalRunResultSchema = z30.object({
   skillVersion: z30.string().optional(),
   scenarioId: z30.string(),
   scenarioName: z30.string(),
+  /** Snapshot of the trigger prompt used during the run (prevents stale display after edits) */
+  triggerPrompt: z30.string().optional(),
   modelConfig: ModelConfigSchema.optional(),
   assertionResults: z30.array(AssertionResultSchema),
   metrics: EvalMetricsSchema.optional(),