@artemiskit/core 0.5.1 → 0.5.2

package/src/runner/runner.ts

@@ -2,6 +2,7 @@
  * Scenario runner - main entry point for running test scenarios
  */
 
+import { nanoid } from 'nanoid';
 import { createRunManifest } from '../artifacts/manifest';
 import type { CaseResult, ManifestRedactionInfo } from '../artifacts/types';
 import { createExecutionProvenance, createWorkloadIdentity } from '../provenance';
@@ -22,6 +23,8 @@ export async function runScenario(options: RunOptions): Promise<RunResult> {
     concurrency = 1,
     timeout,
     retries,
+    repetition = { index: 1, total: 1 },
+    costProvenance,
     redaction,
     toolExecutor,
     onCaseComplete,
@@ -42,6 +45,7 @@ export async function runScenario(options: RunOptions): Promise<RunResult> {
   onProgress?.(`Running ${cases.length} test cases...`);
 
   const startTime = new Date();
+  const runId = nanoid(12);
   const results: CaseResult[] = [];
 
   if (concurrency === 1) {
@@ -54,6 +58,8 @@ export async function runScenario(options: RunOptions): Promise<RunResult> {
         requestedModel: resolvedConfig?.model,
         timeout: testCase.timeout || timeout,
         retries: testCase.retries ?? retries,
+        runId,
+        repetition,
         redaction,
         toolExecutor,
       });
@@ -74,6 +80,8 @@ export async function runScenario(options: RunOptions): Promise<RunResult> {
             requestedModel: resolvedConfig?.model,
             timeout: testCase.timeout || timeout,
             retries: testCase.retries ?? retries,
+            runId,
+            repetition,
             redaction,
             toolExecutor,
           });
@@ -132,9 +140,21 @@ export async function runScenario(options: RunOptions): Promise<RunResult> {
       seed: scenario.seed,
       cases: results,
     }),
+    attemptEvidence: {
+      schema_version: '1',
+      repetition,
+      retry_policy: {
+        default_max_retries: retries ?? 0,
+        backoff: 'exponential',
+        initial_delay_ms: 1000,
+      },
+      ...(timeout ? { timeout: { default_ms: timeout } } : {}),
+    },
+    costProvenance,
     cases: results,
     startTime,
     endTime,
+    runId,
     redaction: redactionInfo,
   });
 

package/src/runner/types.ts

@@ -3,7 +3,7 @@
  */
 
 import type { ModelClient } from '../adapters/types';
-import type { CaseResult, ResolvedConfig, RunManifest } from '../artifacts/types';
+import type { CaseResult, CostProvenance, ResolvedConfig, RunManifest } from '../artifacts/types';
 import type { RedactionConfig } from '../redaction/types';
 import type { Scenario } from '../scenario/schema';
 import type { ToolExecutor } from '../tools';
@@ -28,6 +28,10 @@ export interface RunOptions {
   timeout?: number;
   /** Number of retries per case */
   retries?: number;
+  /** One-based coordinate for an independently planned repetition. */
+  repetition?: { index: number; total: number };
+  /** Attested or operator-supplied monetary evidence; omitted means unavailable. */
+  costProvenance?: CostProvenance;
   /** Redaction configuration (CLI overrides scenario) */
   redaction?: RedactionConfig;
   /** SDK-only executor for explicitly supplied real tools. */
@@ -60,6 +64,8 @@ export interface ExecutorContext {
   requestedModel?: string;
   timeout?: number;
   retries?: number;
+  runId?: string;
+  repetition?: { index: number; total: number };
   /** Redaction configuration for this execution */
   redaction?: RedactionConfig;
   toolExecutor?: ToolExecutor;

package/src/scenario/schema.ts

@@ -225,7 +225,7 @@ export const TestCaseSchema = z.object({
   tags: z.array(z.string()).optional().default([]),
   metadata: z.record(z.unknown()).optional().default({}),
   timeout: z.number().optional(),
-  retries: z.number().optional().default(0),
+  retries: z.number().int().min(0).max(99).optional().default(0),
   provider: ProviderSchema.optional(),
   model: z.string().optional(),
   variables: VariablesSchema.optional(),