@agentv/core 2.12.0 → 2.13.0

package/dist/index.d.cts

@@ -1060,6 +1060,12 @@ interface ExecutionError {
     readonly message: string;
     readonly stage: FailureStage;
 }
+/**
+ * Tolerance for execution errors in an eval run.
+ * - `true`: halt on first execution error
+ * - `false`: never halt on errors (default)
+ */
+type FailOnError = boolean;
 /**
  * Evaluator scorecard for a single eval case run.
  */
@@ -1194,6 +1200,7 @@ type ExecutionDefaults = {
     readonly otel_file?: string;
 };
 type AgentVConfig$1 = {
+    readonly required_version?: string;
     readonly guideline_patterns?: readonly string[];
     readonly eval_patterns?: readonly string[];
     readonly execution?: ExecutionDefaults;
@@ -1238,6 +1245,12 @@ interface CacheConfig {
  * Returns undefined when no cache config is specified.
  */
 declare function extractCacheConfig(suite: JsonObject): CacheConfig | undefined;
+/**
+ * Extract `execution.fail_on_error` from parsed eval suite.
+ * Accepts `true` or `false`.
+ * Returns undefined when not specified.
+ */
+declare function extractFailOnError(suite: JsonObject): FailOnError | undefined;
 
 /**
  * Formatting mode for segment content.
@@ -1297,6 +1310,8 @@ type EvalSuiteResult = {
     readonly metadata?: EvalMetadata;
     /** Suite-level total cost budget in USD */
     readonly totalBudgetUsd?: number;
+    /** Execution error tolerance: true or false */
+    readonly failOnError?: FailOnError;
 };
 /**
  * Load tests and suite metadata from a single parse.
@@ -2440,6 +2455,8 @@ interface RunEvaluationOptions {
     readonly streamCallbacks?: ProviderStreamCallbacks;
     /** Suite-level total cost budget in USD (stops dispatching when exceeded) */
     readonly totalBudgetUsd?: number;
+    /** Execution error tolerance: true halts on first error */
+    readonly failOnError?: FailOnError;
 }
 declare function runEvaluation(options: RunEvaluationOptions): Promise<readonly EvaluationResult[]>;
 declare function runEvalCase(options: RunEvalCaseOptions): Promise<EvaluationResult>;
@@ -3156,4 +3173,4 @@ type AgentKernel = {
 };
 declare function createAgentKernel(): AgentKernel;
 
-export { AgentJudgeEvaluator, type AgentJudgeEvaluatorConfig, type AgentJudgeEvaluatorOptions, type AgentKernel, type AgentVConfig$1 as AgentVConfig, type AgentVConfig as AgentVTsConfig, type AnthropicResolvedConfig, type ArgsMatchMode, type AssertionResult, type AssistantTestMessage, type AzureResolvedConfig, type CacheConfig, type ChildEvaluatorResult, type ClaudeResolvedConfig, type CliResolvedConfig, CodeEvaluator, type CodeEvaluatorConfig, type CodeEvaluatorOptions, type CompositeAggregatorConfig, CompositeEvaluator, type CompositeEvaluatorConfig, type CompositeEvaluatorOptions, type ConfidenceIntervalAggregation, type ContainsAllEvaluatorConfig, type ContainsAnyEvaluatorConfig, type ContainsEvaluatorConfig, type CopilotCliResolvedConfig, type CopilotSdkResolvedConfig, CostEvaluator, type CostEvaluatorConfig, type CostEvaluatorOptions, DEFAULT_EVALUATOR_TEMPLATE, DEFAULT_EVAL_PATTERNS, DEFAULT_EXPLORATION_TOOLS, DeterministicAssertionEvaluator, type EndsWithEvaluatorConfig, type EnsureSubagentsOptions, type EnsureSubagentsResult, type EnvLookup, type EqualsEvaluatorConfig, type EvalAssertionInput, type EvalCase, type EvalConfig, type EvalMetadata, type EvalRunResult, type EvalSuiteResult, type EvalSummary, type EvalTest, type EvalTestInput, type EvaluationCache, type EvaluationContext, type EvaluationResult, type EvaluationScore, type EvaluationVerdict, type Evaluator, type EvaluatorConfig, type EvaluatorDispatchContext, type EvaluatorFactory, type EvaluatorFactoryFn, type EvaluatorKind, EvaluatorRegistry, type EvaluatorResult, type ExecutionDefaults, type ExecutionError, type ExecutionMetrics, ExecutionMetricsEvaluator, type ExecutionMetricsEvaluatorConfig, type ExecutionMetricsEvaluatorOptions, type ExecutionStatus, type FailureStage, FieldAccuracyEvaluator, type FieldAccuracyEvaluatorConfig, type FieldAccuracyEvaluatorOptions, type FieldAggregationType, type FieldConfig, type FieldMatchType, type GeminiResolvedConfig, type GenerateRubricsOptions, type IcontainsAllEvaluatorConfig, type IcontainsAnyEvaluatorConfig, type IcontainsEvaluatorConfig, type IsJsonEvaluatorConfig, type JsonObject, type JsonPrimitive, type JsonValue, LatencyEvaluator, type LatencyEvaluatorConfig, type LatencyEvaluatorOptions, LlmJudgeEvaluator, type LlmJudgeEvaluatorConfig, type LlmJudgeEvaluatorOptions, type LlmJudgePromptAssembly, type MeanAggregation, type Message, type MockResolvedConfig, OTEL_BACKEND_PRESETS, type OtelBackendPreset, type OtelExportOptions, OtelStreamingObserver, OtelTraceExporter, OtlpJsonFileExporter, type OutputMessage, type PassAtKAggregation, type PiAgentSdkResolvedConfig, type PiCodingAgentResolvedConfig, type ProgressEvent, type PromptInputs, type PromptScriptConfig, type Provider, type ProviderFactoryFn, type ProviderKind, ProviderRegistry, type ProviderRequest, type ProviderResponse, type ProviderStreamCallbacks, type ProviderTokenUsage, type RegexEvaluatorConfig, type RepoCheckout, type RepoClone, type RepoConfig, RepoManager, type RepoSource, type ResetConfig, type ResolvedTarget, type ResolvedWorkspaceTemplate, ResponseCache, type RubricItem, type RubricsEvaluatorConfig, type RunEvalCaseOptions, type RunEvaluationOptions, type ScoreRange, type ScriptExecutionContext, SimpleTraceFileExporter, type StartsWithEvaluatorConfig, type SystemTestMessage, TEST_MESSAGE_ROLES, type TargetAccessConfig, type TargetDefinition, TemplateNotDirectoryError, TemplateNotFoundError, type TestMessage, type TestMessageContent, type TestMessageRole, type TokenUsage, TokenUsageEvaluator, type TokenUsageEvaluatorConfig, type TokenUsageEvaluatorOptions, type ToolCall, type ToolTestMessage, ToolTrajectoryEvaluator, type ToolTrajectoryEvaluatorConfig, type ToolTrajectoryEvaluatorOptions, type ToolTrajectoryExpectedItem, type TraceComputeResult, type TraceSummary, type TrialAggregation, type TrialResult, type TrialStrategy, type TrialsConfig, type UserTestMessage, type VSCodeResolvedConfig, type WorkspaceConfig, WorkspaceCreationError, type WorkspaceScriptConfig, assembleLlmJudgePrompt, avgToolDurationMs, buildDirectoryChain, buildOutputSchema, buildPromptInputs, buildRubricOutputSchema, buildScoreRangeOutputSchema, buildSearchRoots, calculateRubricScore, captureFileChanges, clampScore, cleanupEvalWorkspaces, cleanupWorkspace, computeTraceSummary, consumeClaudeLogEntries, consumeCodexLogEntries, consumeCopilotCliLogEntries, consumeCopilotSdkLogEntries, consumePiLogEntries, createAgentKernel, createBuiltinProviderRegistry, createBuiltinRegistry, createProvider, createTempWorkspace, deepEqual, defineConfig, detectFormat, discoverAssertions, discoverProviders, ensureVSCodeSubagents, evaluate, executeScript, executeWorkspaceScript, explorationRatio, extractCacheConfig, extractJsonBlob, extractTargetFromSuite, extractTargetsFromSuite, extractTargetsFromTestCase, extractTrialsConfig, fileExists, findGitRoot, freeformEvaluationSchema, generateRubrics, getHitCount, getWorkspacePath, initializeBaseline, isEvaluatorKind, isGuidelineFile, isJsonObject, isJsonValue, isNonEmptyString, isTestMessage, isTestMessageRole, listTargetNames, loadConfig, loadEvalCaseById, loadEvalCases, loadEvalSuite, loadTestById, loadTestSuite, loadTests, loadTsConfig, mergeExecutionMetrics, negateScore, normalizeLineEndings, parseJsonFromText, parseJsonSafe, readJsonFile, readTargetDefinitions, readTestSuiteMetadata, readTextFile, resolveAndCreateProvider, resolveFileReference, resolveTargetDefinition, resolveWorkspaceTemplate, rubricEvaluationSchema, runContainsAllAssertion, runContainsAnyAssertion, runContainsAssertion, runEndsWithAssertion, runEqualsAssertion, runEvalCase, runEvaluation, runIcontainsAllAssertion, runIcontainsAnyAssertion, runIcontainsAssertion, runIsJsonAssertion, runRegexAssertion, runStartsWithAssertion, scoreToVerdict, shouldEnableCache, shouldSkipCacheForTemperature, subscribeToClaudeLogEntries, subscribeToCodexLogEntries, subscribeToCopilotCliLogEntries, subscribeToCopilotSdkLogEntries, subscribeToPiLogEntries, substituteVariables, toCamelCaseDeep, toSnakeCaseDeep, tokensPerTool, trimBaselineResult };
+export { AgentJudgeEvaluator, type AgentJudgeEvaluatorConfig, type AgentJudgeEvaluatorOptions, type AgentKernel, type AgentVConfig$1 as AgentVConfig, type AgentVConfig as AgentVTsConfig, type AnthropicResolvedConfig, type ArgsMatchMode, type AssertionResult, type AssistantTestMessage, type AzureResolvedConfig, type CacheConfig, type ChildEvaluatorResult, type ClaudeResolvedConfig, type CliResolvedConfig, CodeEvaluator, type CodeEvaluatorConfig, type CodeEvaluatorOptions, type CompositeAggregatorConfig, CompositeEvaluator, type CompositeEvaluatorConfig, type CompositeEvaluatorOptions, type ConfidenceIntervalAggregation, type ContainsAllEvaluatorConfig, type ContainsAnyEvaluatorConfig, type ContainsEvaluatorConfig, type CopilotCliResolvedConfig, type CopilotSdkResolvedConfig, CostEvaluator, type CostEvaluatorConfig, type CostEvaluatorOptions, DEFAULT_EVALUATOR_TEMPLATE, DEFAULT_EVAL_PATTERNS, DEFAULT_EXPLORATION_TOOLS, DeterministicAssertionEvaluator, type EndsWithEvaluatorConfig, type EnsureSubagentsOptions, type EnsureSubagentsResult, type EnvLookup, type EqualsEvaluatorConfig, type EvalAssertionInput, type EvalCase, type EvalConfig, type EvalMetadata, type EvalRunResult, type EvalSuiteResult, type EvalSummary, type EvalTest, type EvalTestInput, type EvaluationCache, type EvaluationContext, type EvaluationResult, type EvaluationScore, type EvaluationVerdict, type Evaluator, type EvaluatorConfig, type EvaluatorDispatchContext, type EvaluatorFactory, type EvaluatorFactoryFn, type EvaluatorKind, EvaluatorRegistry, type EvaluatorResult, type ExecutionDefaults, type ExecutionError, type ExecutionMetrics, ExecutionMetricsEvaluator, type ExecutionMetricsEvaluatorConfig, type ExecutionMetricsEvaluatorOptions, type ExecutionStatus, type FailOnError, type FailureStage, FieldAccuracyEvaluator, type FieldAccuracyEvaluatorConfig, type FieldAccuracyEvaluatorOptions, type FieldAggregationType, type FieldConfig, type FieldMatchType, type GeminiResolvedConfig, type GenerateRubricsOptions, type IcontainsAllEvaluatorConfig, type IcontainsAnyEvaluatorConfig, type IcontainsEvaluatorConfig, type IsJsonEvaluatorConfig, type JsonObject, type JsonPrimitive, type JsonValue, LatencyEvaluator, type LatencyEvaluatorConfig, type LatencyEvaluatorOptions, LlmJudgeEvaluator, type LlmJudgeEvaluatorConfig, type LlmJudgeEvaluatorOptions, type LlmJudgePromptAssembly, type MeanAggregation, type Message, type MockResolvedConfig, OTEL_BACKEND_PRESETS, type OtelBackendPreset, type OtelExportOptions, OtelStreamingObserver, OtelTraceExporter, OtlpJsonFileExporter, type OutputMessage, type PassAtKAggregation, type PiAgentSdkResolvedConfig, type PiCodingAgentResolvedConfig, type ProgressEvent, type PromptInputs, type PromptScriptConfig, type Provider, type ProviderFactoryFn, type ProviderKind, ProviderRegistry, type ProviderRequest, type ProviderResponse, type ProviderStreamCallbacks, type ProviderTokenUsage, type RegexEvaluatorConfig, type RepoCheckout, type RepoClone, type RepoConfig, RepoManager, type RepoSource, type ResetConfig, type ResolvedTarget, type ResolvedWorkspaceTemplate, ResponseCache, type RubricItem, type RubricsEvaluatorConfig, type RunEvalCaseOptions, type RunEvaluationOptions, type ScoreRange, type ScriptExecutionContext, SimpleTraceFileExporter, type StartsWithEvaluatorConfig, type SystemTestMessage, TEST_MESSAGE_ROLES, type TargetAccessConfig, type TargetDefinition, TemplateNotDirectoryError, TemplateNotFoundError, type TestMessage, type TestMessageContent, type TestMessageRole, type TokenUsage, TokenUsageEvaluator, type TokenUsageEvaluatorConfig, type TokenUsageEvaluatorOptions, type ToolCall, type ToolTestMessage, ToolTrajectoryEvaluator, type ToolTrajectoryEvaluatorConfig, type ToolTrajectoryEvaluatorOptions, type ToolTrajectoryExpectedItem, type TraceComputeResult, type TraceSummary, type TrialAggregation, type TrialResult, type TrialStrategy, type TrialsConfig, type UserTestMessage, type VSCodeResolvedConfig, type WorkspaceConfig, WorkspaceCreationError, type WorkspaceScriptConfig, assembleLlmJudgePrompt, avgToolDurationMs, buildDirectoryChain, buildOutputSchema, buildPromptInputs, buildRubricOutputSchema, buildScoreRangeOutputSchema, buildSearchRoots, calculateRubricScore, captureFileChanges, clampScore, cleanupEvalWorkspaces, cleanupWorkspace, computeTraceSummary, consumeClaudeLogEntries, consumeCodexLogEntries, consumeCopilotCliLogEntries, consumeCopilotSdkLogEntries, consumePiLogEntries, createAgentKernel, createBuiltinProviderRegistry, createBuiltinRegistry, createProvider, createTempWorkspace, deepEqual, defineConfig, detectFormat, discoverAssertions, discoverProviders, ensureVSCodeSubagents, evaluate, executeScript, executeWorkspaceScript, explorationRatio, extractCacheConfig, extractFailOnError, extractJsonBlob, extractTargetFromSuite, extractTargetsFromSuite, extractTargetsFromTestCase, extractTrialsConfig, fileExists, findGitRoot, freeformEvaluationSchema, generateRubrics, getHitCount, getWorkspacePath, initializeBaseline, isEvaluatorKind, isGuidelineFile, isJsonObject, isJsonValue, isNonEmptyString, isTestMessage, isTestMessageRole, listTargetNames, loadConfig, loadEvalCaseById, loadEvalCases, loadEvalSuite, loadTestById, loadTestSuite, loadTests, loadTsConfig, mergeExecutionMetrics, negateScore, normalizeLineEndings, parseJsonFromText, parseJsonSafe, readJsonFile, readTargetDefinitions, readTestSuiteMetadata, readTextFile, resolveAndCreateProvider, resolveFileReference, resolveTargetDefinition, resolveWorkspaceTemplate, rubricEvaluationSchema, runContainsAllAssertion, runContainsAnyAssertion, runContainsAssertion, runEndsWithAssertion, runEqualsAssertion, runEvalCase, runEvaluation, runIcontainsAllAssertion, runIcontainsAnyAssertion, runIcontainsAssertion, runIsJsonAssertion, runRegexAssertion, runStartsWithAssertion, scoreToVerdict, shouldEnableCache, shouldSkipCacheForTemperature, subscribeToClaudeLogEntries, subscribeToCodexLogEntries, subscribeToCopilotCliLogEntries, subscribeToCopilotSdkLogEntries, subscribeToPiLogEntries, substituteVariables, toCamelCaseDeep, toSnakeCaseDeep, tokensPerTool, trimBaselineResult };

package/dist/index.d.ts

@@ -1060,6 +1060,12 @@ interface ExecutionError {
     readonly message: string;
     readonly stage: FailureStage;
 }
+/**
+ * Tolerance for execution errors in an eval run.
+ * - `true`: halt on first execution error
+ * - `false`: never halt on errors (default)
+ */
+type FailOnError = boolean;
 /**
  * Evaluator scorecard for a single eval case run.
  */
@@ -1194,6 +1200,7 @@ type ExecutionDefaults = {
     readonly otel_file?: string;
 };
 type AgentVConfig$1 = {
+    readonly required_version?: string;
     readonly guideline_patterns?: readonly string[];
     readonly eval_patterns?: readonly string[];
     readonly execution?: ExecutionDefaults;
@@ -1238,6 +1245,12 @@ interface CacheConfig {
  * Returns undefined when no cache config is specified.
  */
 declare function extractCacheConfig(suite: JsonObject): CacheConfig | undefined;
+/**
+ * Extract `execution.fail_on_error` from parsed eval suite.
+ * Accepts `true` or `false`.
+ * Returns undefined when not specified.
+ */
+declare function extractFailOnError(suite: JsonObject): FailOnError | undefined;
 
 /**
  * Formatting mode for segment content.
@@ -1297,6 +1310,8 @@ type EvalSuiteResult = {
     readonly metadata?: EvalMetadata;
     /** Suite-level total cost budget in USD */
     readonly totalBudgetUsd?: number;
+    /** Execution error tolerance: true or false */
+    readonly failOnError?: FailOnError;
 };
 /**
  * Load tests and suite metadata from a single parse.
@@ -2440,6 +2455,8 @@ interface RunEvaluationOptions {
     readonly streamCallbacks?: ProviderStreamCallbacks;
     /** Suite-level total cost budget in USD (stops dispatching when exceeded) */
     readonly totalBudgetUsd?: number;
+    /** Execution error tolerance: true halts on first error */
+    readonly failOnError?: FailOnError;
 }
 declare function runEvaluation(options: RunEvaluationOptions): Promise<readonly EvaluationResult[]>;
 declare function runEvalCase(options: RunEvalCaseOptions): Promise<EvaluationResult>;
@@ -3156,4 +3173,4 @@ type AgentKernel = {
 };
 declare function createAgentKernel(): AgentKernel;
 
-export { AgentJudgeEvaluator, type AgentJudgeEvaluatorConfig, type AgentJudgeEvaluatorOptions, type AgentKernel, type AgentVConfig$1 as AgentVConfig, type AgentVConfig as AgentVTsConfig, type AnthropicResolvedConfig, type ArgsMatchMode, type AssertionResult, type AssistantTestMessage, type AzureResolvedConfig, type CacheConfig, type ChildEvaluatorResult, type ClaudeResolvedConfig, type CliResolvedConfig, CodeEvaluator, type CodeEvaluatorConfig, type CodeEvaluatorOptions, type CompositeAggregatorConfig, CompositeEvaluator, type CompositeEvaluatorConfig, type CompositeEvaluatorOptions, type ConfidenceIntervalAggregation, type ContainsAllEvaluatorConfig, type ContainsAnyEvaluatorConfig, type ContainsEvaluatorConfig, type CopilotCliResolvedConfig, type CopilotSdkResolvedConfig, CostEvaluator, type CostEvaluatorConfig, type CostEvaluatorOptions, DEFAULT_EVALUATOR_TEMPLATE, DEFAULT_EVAL_PATTERNS, DEFAULT_EXPLORATION_TOOLS, DeterministicAssertionEvaluator, type EndsWithEvaluatorConfig, type EnsureSubagentsOptions, type EnsureSubagentsResult, type EnvLookup, type EqualsEvaluatorConfig, type EvalAssertionInput, type EvalCase, type EvalConfig, type EvalMetadata, type EvalRunResult, type EvalSuiteResult, type EvalSummary, type EvalTest, type EvalTestInput, type EvaluationCache, type EvaluationContext, type EvaluationResult, type EvaluationScore, type EvaluationVerdict, type Evaluator, type EvaluatorConfig, type EvaluatorDispatchContext, type EvaluatorFactory, type EvaluatorFactoryFn, type EvaluatorKind, EvaluatorRegistry, type EvaluatorResult, type ExecutionDefaults, type ExecutionError, type ExecutionMetrics, ExecutionMetricsEvaluator, type ExecutionMetricsEvaluatorConfig, type ExecutionMetricsEvaluatorOptions, type ExecutionStatus, type FailureStage, FieldAccuracyEvaluator, type FieldAccuracyEvaluatorConfig, type FieldAccuracyEvaluatorOptions, type FieldAggregationType, type FieldConfig, type FieldMatchType, type GeminiResolvedConfig, type GenerateRubricsOptions, type IcontainsAllEvaluatorConfig, type IcontainsAnyEvaluatorConfig, type IcontainsEvaluatorConfig, type IsJsonEvaluatorConfig, type JsonObject, type JsonPrimitive, type JsonValue, LatencyEvaluator, type LatencyEvaluatorConfig, type LatencyEvaluatorOptions, LlmJudgeEvaluator, type LlmJudgeEvaluatorConfig, type LlmJudgeEvaluatorOptions, type LlmJudgePromptAssembly, type MeanAggregation, type Message, type MockResolvedConfig, OTEL_BACKEND_PRESETS, type OtelBackendPreset, type OtelExportOptions, OtelStreamingObserver, OtelTraceExporter, OtlpJsonFileExporter, type OutputMessage, type PassAtKAggregation, type PiAgentSdkResolvedConfig, type PiCodingAgentResolvedConfig, type ProgressEvent, type PromptInputs, type PromptScriptConfig, type Provider, type ProviderFactoryFn, type ProviderKind, ProviderRegistry, type ProviderRequest, type ProviderResponse, type ProviderStreamCallbacks, type ProviderTokenUsage, type RegexEvaluatorConfig, type RepoCheckout, type RepoClone, type RepoConfig, RepoManager, type RepoSource, type ResetConfig, type ResolvedTarget, type ResolvedWorkspaceTemplate, ResponseCache, type RubricItem, type RubricsEvaluatorConfig, type RunEvalCaseOptions, type RunEvaluationOptions, type ScoreRange, type ScriptExecutionContext, SimpleTraceFileExporter, type StartsWithEvaluatorConfig, type SystemTestMessage, TEST_MESSAGE_ROLES, type TargetAccessConfig, type TargetDefinition, TemplateNotDirectoryError, TemplateNotFoundError, type TestMessage, type TestMessageContent, type TestMessageRole, type TokenUsage, TokenUsageEvaluator, type TokenUsageEvaluatorConfig, type TokenUsageEvaluatorOptions, type ToolCall, type ToolTestMessage, ToolTrajectoryEvaluator, type ToolTrajectoryEvaluatorConfig, type ToolTrajectoryEvaluatorOptions, type ToolTrajectoryExpectedItem, type TraceComputeResult, type TraceSummary, type TrialAggregation, type TrialResult, type TrialStrategy, type TrialsConfig, type UserTestMessage, type VSCodeResolvedConfig, type WorkspaceConfig, WorkspaceCreationError, type WorkspaceScriptConfig, assembleLlmJudgePrompt, avgToolDurationMs, buildDirectoryChain, buildOutputSchema, buildPromptInputs, buildRubricOutputSchema, buildScoreRangeOutputSchema, buildSearchRoots, calculateRubricScore, captureFileChanges, clampScore, cleanupEvalWorkspaces, cleanupWorkspace, computeTraceSummary, consumeClaudeLogEntries, consumeCodexLogEntries, consumeCopilotCliLogEntries, consumeCopilotSdkLogEntries, consumePiLogEntries, createAgentKernel, createBuiltinProviderRegistry, createBuiltinRegistry, createProvider, createTempWorkspace, deepEqual, defineConfig, detectFormat, discoverAssertions, discoverProviders, ensureVSCodeSubagents, evaluate, executeScript, executeWorkspaceScript, explorationRatio, extractCacheConfig, extractJsonBlob, extractTargetFromSuite, extractTargetsFromSuite, extractTargetsFromTestCase, extractTrialsConfig, fileExists, findGitRoot, freeformEvaluationSchema, generateRubrics, getHitCount, getWorkspacePath, initializeBaseline, isEvaluatorKind, isGuidelineFile, isJsonObject, isJsonValue, isNonEmptyString, isTestMessage, isTestMessageRole, listTargetNames, loadConfig, loadEvalCaseById, loadEvalCases, loadEvalSuite, loadTestById, loadTestSuite, loadTests, loadTsConfig, mergeExecutionMetrics, negateScore, normalizeLineEndings, parseJsonFromText, parseJsonSafe, readJsonFile, readTargetDefinitions, readTestSuiteMetadata, readTextFile, resolveAndCreateProvider, resolveFileReference, resolveTargetDefinition, resolveWorkspaceTemplate, rubricEvaluationSchema, runContainsAllAssertion, runContainsAnyAssertion, runContainsAssertion, runEndsWithAssertion, runEqualsAssertion, runEvalCase, runEvaluation, runIcontainsAllAssertion, runIcontainsAnyAssertion, runIcontainsAssertion, runIsJsonAssertion, runRegexAssertion, runStartsWithAssertion, scoreToVerdict, shouldEnableCache, shouldSkipCacheForTemperature, subscribeToClaudeLogEntries, subscribeToCodexLogEntries, subscribeToCopilotCliLogEntries, subscribeToCopilotSdkLogEntries, subscribeToPiLogEntries, substituteVariables, toCamelCaseDeep, toSnakeCaseDeep, tokensPerTool, trimBaselineResult };
+export { AgentJudgeEvaluator, type AgentJudgeEvaluatorConfig, type AgentJudgeEvaluatorOptions, type AgentKernel, type AgentVConfig$1 as AgentVConfig, type AgentVConfig as AgentVTsConfig, type AnthropicResolvedConfig, type ArgsMatchMode, type AssertionResult, type AssistantTestMessage, type AzureResolvedConfig, type CacheConfig, type ChildEvaluatorResult, type ClaudeResolvedConfig, type CliResolvedConfig, CodeEvaluator, type CodeEvaluatorConfig, type CodeEvaluatorOptions, type CompositeAggregatorConfig, CompositeEvaluator, type CompositeEvaluatorConfig, type CompositeEvaluatorOptions, type ConfidenceIntervalAggregation, type ContainsAllEvaluatorConfig, type ContainsAnyEvaluatorConfig, type ContainsEvaluatorConfig, type CopilotCliResolvedConfig, type CopilotSdkResolvedConfig, CostEvaluator, type CostEvaluatorConfig, type CostEvaluatorOptions, DEFAULT_EVALUATOR_TEMPLATE, DEFAULT_EVAL_PATTERNS, DEFAULT_EXPLORATION_TOOLS, DeterministicAssertionEvaluator, type EndsWithEvaluatorConfig, type EnsureSubagentsOptions, type EnsureSubagentsResult, type EnvLookup, type EqualsEvaluatorConfig, type EvalAssertionInput, type EvalCase, type EvalConfig, type EvalMetadata, type EvalRunResult, type EvalSuiteResult, type EvalSummary, type EvalTest, type EvalTestInput, type EvaluationCache, type EvaluationContext, type EvaluationResult, type EvaluationScore, type EvaluationVerdict, type Evaluator, type EvaluatorConfig, type EvaluatorDispatchContext, type EvaluatorFactory, type EvaluatorFactoryFn, type EvaluatorKind, EvaluatorRegistry, type EvaluatorResult, type ExecutionDefaults, type ExecutionError, type ExecutionMetrics, ExecutionMetricsEvaluator, type ExecutionMetricsEvaluatorConfig, type ExecutionMetricsEvaluatorOptions, type ExecutionStatus, type FailOnError, type FailureStage, FieldAccuracyEvaluator, type FieldAccuracyEvaluatorConfig, type FieldAccuracyEvaluatorOptions, type FieldAggregationType, type FieldConfig, type FieldMatchType, type GeminiResolvedConfig, type GenerateRubricsOptions, type IcontainsAllEvaluatorConfig, type IcontainsAnyEvaluatorConfig, type IcontainsEvaluatorConfig, type IsJsonEvaluatorConfig, type JsonObject, type JsonPrimitive, type JsonValue, LatencyEvaluator, type LatencyEvaluatorConfig, type LatencyEvaluatorOptions, LlmJudgeEvaluator, type LlmJudgeEvaluatorConfig, type LlmJudgeEvaluatorOptions, type LlmJudgePromptAssembly, type MeanAggregation, type Message, type MockResolvedConfig, OTEL_BACKEND_PRESETS, type OtelBackendPreset, type OtelExportOptions, OtelStreamingObserver, OtelTraceExporter, OtlpJsonFileExporter, type OutputMessage, type PassAtKAggregation, type PiAgentSdkResolvedConfig, type PiCodingAgentResolvedConfig, type ProgressEvent, type PromptInputs, type PromptScriptConfig, type Provider, type ProviderFactoryFn, type ProviderKind, ProviderRegistry, type ProviderRequest, type ProviderResponse, type ProviderStreamCallbacks, type ProviderTokenUsage, type RegexEvaluatorConfig, type RepoCheckout, type RepoClone, type RepoConfig, RepoManager, type RepoSource, type ResetConfig, type ResolvedTarget, type ResolvedWorkspaceTemplate, ResponseCache, type RubricItem, type RubricsEvaluatorConfig, type RunEvalCaseOptions, type RunEvaluationOptions, type ScoreRange, type ScriptExecutionContext, SimpleTraceFileExporter, type StartsWithEvaluatorConfig, type SystemTestMessage, TEST_MESSAGE_ROLES, type TargetAccessConfig, type TargetDefinition, TemplateNotDirectoryError, TemplateNotFoundError, type TestMessage, type TestMessageContent, type TestMessageRole, type TokenUsage, TokenUsageEvaluator, type TokenUsageEvaluatorConfig, type TokenUsageEvaluatorOptions, type ToolCall, type ToolTestMessage, ToolTrajectoryEvaluator, type ToolTrajectoryEvaluatorConfig, type ToolTrajectoryEvaluatorOptions, type ToolTrajectoryExpectedItem, type TraceComputeResult, type TraceSummary, type TrialAggregation, type TrialResult, type TrialStrategy, type TrialsConfig, type UserTestMessage, type VSCodeResolvedConfig, type WorkspaceConfig, WorkspaceCreationError, type WorkspaceScriptConfig, assembleLlmJudgePrompt, avgToolDurationMs, buildDirectoryChain, buildOutputSchema, buildPromptInputs, buildRubricOutputSchema, buildScoreRangeOutputSchema, buildSearchRoots, calculateRubricScore, captureFileChanges, clampScore, cleanupEvalWorkspaces, cleanupWorkspace, computeTraceSummary, consumeClaudeLogEntries, consumeCodexLogEntries, consumeCopilotCliLogEntries, consumeCopilotSdkLogEntries, consumePiLogEntries, createAgentKernel, createBuiltinProviderRegistry, createBuiltinRegistry, createProvider, createTempWorkspace, deepEqual, defineConfig, detectFormat, discoverAssertions, discoverProviders, ensureVSCodeSubagents, evaluate, executeScript, executeWorkspaceScript, explorationRatio, extractCacheConfig, extractFailOnError, extractJsonBlob, extractTargetFromSuite, extractTargetsFromSuite, extractTargetsFromTestCase, extractTrialsConfig, fileExists, findGitRoot, freeformEvaluationSchema, generateRubrics, getHitCount, getWorkspacePath, initializeBaseline, isEvaluatorKind, isGuidelineFile, isJsonObject, isJsonValue, isNonEmptyString, isTestMessage, isTestMessageRole, listTargetNames, loadConfig, loadEvalCaseById, loadEvalCases, loadEvalSuite, loadTestById, loadTestSuite, loadTests, loadTsConfig, mergeExecutionMetrics, negateScore, normalizeLineEndings, parseJsonFromText, parseJsonSafe, readJsonFile, readTargetDefinitions, readTestSuiteMetadata, readTextFile, resolveAndCreateProvider, resolveFileReference, resolveTargetDefinition, resolveWorkspaceTemplate, rubricEvaluationSchema, runContainsAllAssertion, runContainsAnyAssertion, runContainsAssertion, runEndsWithAssertion, runEqualsAssertion, runEvalCase, runEvaluation, runIcontainsAllAssertion, runIcontainsAnyAssertion, runIcontainsAssertion, runIsJsonAssertion, runRegexAssertion, runStartsWithAssertion, scoreToVerdict, shouldEnableCache, shouldSkipCacheForTemperature, subscribeToClaudeLogEntries, subscribeToCodexLogEntries, subscribeToCopilotCliLogEntries, subscribeToCopilotSdkLogEntries, subscribeToPiLogEntries, substituteVariables, toCamelCaseDeep, toSnakeCaseDeep, tokensPerTool, trimBaselineResult };

package/dist/index.js

@@ -17,7 +17,7 @@ import {
   readTextFile,
   resolveFileReference,
   resolveTargetDefinition
-} from "./chunk-7HPKTRFZ.js";
+} from "./chunk-JHER2LQ5.js";
 import {
   OtlpJsonFileExporter
 } from "./chunk-HFSYZHGF.js";
@@ -396,6 +396,11 @@ async function loadConfig(evalFilePath, repoRoot) {
         continue;
       }
       const config = parsed;
+      const requiredVersion = parsed.required_version;
+      if (requiredVersion !== void 0 && typeof requiredVersion !== "string") {
+        logWarning(`Invalid required_version in ${configPath}, expected string`);
+        continue;
+      }
       const guidelinePatterns = config.guideline_patterns;
       if (guidelinePatterns !== void 0 && !Array.isArray(guidelinePatterns)) {
         logWarning(`Invalid guideline_patterns in ${configPath}, expected array`);
@@ -419,6 +424,7 @@ async function loadConfig(evalFilePath, repoRoot) {
         configPath
       );
       return {
+        required_version: requiredVersion,
         guideline_patterns: guidelinePatterns,
         eval_patterns: evalPatterns,
         execution: executionDefaults
@@ -562,6 +568,22 @@ function extractTotalBudgetUsd(suite) {
   );
   return void 0;
 }
+function extractFailOnError(suite) {
+  const execution = suite.execution;
+  if (!execution || typeof execution !== "object" || Array.isArray(execution)) {
+    return void 0;
+  }
+  const executionObj = execution;
+  const raw = executionObj.fail_on_error ?? executionObj.failOnError;
+  if (raw === void 0 || raw === null) {
+    return void 0;
+  }
+  if (typeof raw === "boolean") {
+    return raw;
+  }
+  logWarning(`Invalid execution.fail_on_error: ${raw}. Must be true or false. Ignoring.`);
+  return void 0;
+}
 function parseExecutionDefaults(raw, configPath) {
   if (!raw || typeof raw !== "object" || Array.isArray(raw)) {
     return void 0;
@@ -2757,13 +2779,15 @@ async function loadTestSuite(evalFilePath, repoRoot, options) {
   }
   const { tests, parsed } = await loadTestsFromYaml(evalFilePath, repoRoot, options);
   const metadata = parseMetadata(parsed);
+  const failOnError = extractFailOnError(parsed);
   return {
     tests,
     trials: extractTrialsConfig(parsed),
     targets: extractTargetsFromSuite(parsed),
     cacheConfig: extractCacheConfig(parsed),
     totalBudgetUsd: extractTotalBudgetUsd(parsed),
-    ...metadata !== void 0 && { metadata }
+    ...metadata !== void 0 && { metadata },
+    ...failOnError !== void 0 && { failOnError }
   };
 }
 var loadEvalSuite = loadTestSuite;
@@ -12921,7 +12945,8 @@ async function runEvaluation(options) {
     cleanupWorkspaces,
     trials,
     streamCallbacks,
-    totalBudgetUsd
+    totalBudgetUsd,
+    failOnError
   } = options;
   let useCache = options.useCache;
   if (trials && trials.count > 1 && useCache) {
@@ -13117,6 +13142,7 @@ async function runEvaluation(options) {
   let beforeAllOutputAttached = false;
   let cumulativeBudgetCost = 0;
   let budgetExhausted = false;
+  let failOnErrorTriggered = false;
   const promises = filteredEvalCases.map(
     (evalCase) => limit(async () => {
       const workerId = nextWorkerId++;
@@ -13155,6 +13181,37 @@ async function runEvaluation(options) {
         }
         return budgetResult;
       }
+      if (failOnError === true && failOnErrorTriggered) {
+        const errorMsg = "Halted: execution error encountered with fail_on_error enabled";
+        const haltResult = {
+          timestamp: (now ?? (() => /* @__PURE__ */ new Date()))().toISOString(),
+          testId: evalCase.id,
+          dataset: evalCase.dataset,
+          score: 0,
+          hits: [],
+          misses: [],
+          answer: "",
+          target: target.name,
+          error: errorMsg,
+          executionStatus: "execution_error",
+          failureStage: "setup",
+          failureReasonCode: "error_threshold_exceeded",
+          executionError: { message: errorMsg, stage: "setup" }
+        };
+        if (onProgress) {
+          await onProgress({
+            workerId,
+            testId: evalCase.id,
+            status: "failed",
+            completedAt: Date.now(),
+            error: haltResult.error
+          });
+        }
+        if (onResult) {
+          await onResult(haltResult);
+        }
+        return haltResult;
+      }
       if (onProgress) {
         await onProgress({
           workerId,
@@ -13207,6 +13264,9 @@ async function runEvaluation(options) {
             }
           }
         }
+        if (failOnError === true && result.executionStatus === "execution_error") {
+          failOnErrorTriggered = true;
+        }
         if (beforeAllOutput && !beforeAllOutputAttached) {
           result = { ...result, beforeAllOutput };
           beforeAllOutputAttached = true;
@@ -15268,6 +15328,7 @@ export {
   executeWorkspaceScript,
   explorationRatio,
   extractCacheConfig,
+  extractFailOnError,
   extractJsonBlob,
   extractTargetFromSuite,
   extractTargetsFromSuite,