agent-inspect 4.2.0 → 4.4.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (97) hide show
  1. package/CHANGELOG.md +12 -0
  2. package/docs/API.md +3 -0
  3. package/docs/CLI.md +36 -0
  4. package/docs/SCHEMA.md +27 -1
  5. package/package.json +1 -1
  6. package/packages/cli/dist/{chunk-5VSPJEZ7.mjs → chunk-ORHDQUT6.mjs} +385 -10
  7. package/packages/cli/dist/chunk-ORHDQUT6.mjs.map +1 -0
  8. package/packages/cli/dist/index.cjs +2548 -1567
  9. package/packages/cli/dist/index.cjs.map +1 -1
  10. package/packages/cli/dist/index.mjs +2028 -1495
  11. package/packages/cli/dist/index.mjs.map +1 -1
  12. package/packages/cli/dist/{src-VQT7QLGV.mjs → src-FJOHPC4X.mjs} +3 -3
  13. package/packages/cli/dist/{src-VQT7QLGV.mjs.map → src-FJOHPC4X.mjs.map} +1 -1
  14. package/packages/core/dist/advanced.cjs +397 -12
  15. package/packages/core/dist/advanced.cjs.map +1 -1
  16. package/packages/core/dist/advanced.d.cts +105 -9
  17. package/packages/core/dist/advanced.d.ts +105 -9
  18. package/packages/core/dist/advanced.mjs +219 -17
  19. package/packages/core/dist/advanced.mjs.map +1 -1
  20. package/packages/core/dist/checks.cjs +117 -18
  21. package/packages/core/dist/checks.cjs.map +1 -1
  22. package/packages/core/dist/checks.d.cts +11 -3
  23. package/packages/core/dist/checks.d.ts +11 -3
  24. package/packages/core/dist/checks.mjs +42 -1
  25. package/packages/core/dist/checks.mjs.map +1 -1
  26. package/packages/core/dist/{chunk-2ZXEMWGK.mjs → chunk-3USJVBLA.mjs} +3 -3
  27. package/packages/core/dist/{chunk-2ZXEMWGK.mjs.map → chunk-3USJVBLA.mjs.map} +1 -1
  28. package/packages/core/dist/{chunk-HY7H3CQM.mjs → chunk-CAWPF22J.mjs} +4 -3
  29. package/packages/core/dist/chunk-CAWPF22J.mjs.map +1 -0
  30. package/packages/core/dist/{chunk-XUNJMAPD.mjs → chunk-KNYL56KZ.mjs} +38 -3
  31. package/packages/core/dist/chunk-KNYL56KZ.mjs.map +1 -0
  32. package/packages/core/dist/{chunk-IZBJAZGF.mjs → chunk-P366CCL2.mjs} +34 -3
  33. package/packages/core/dist/chunk-P366CCL2.mjs.map +1 -0
  34. package/packages/core/dist/{chunk-4JVGFNQH.mjs → chunk-ULCGIRTC.mjs} +95 -8
  35. package/packages/core/dist/chunk-ULCGIRTC.mjs.map +1 -0
  36. package/packages/core/dist/chunk-VFO76UH3.mjs +228 -0
  37. package/packages/core/dist/chunk-VFO76UH3.mjs.map +1 -0
  38. package/packages/core/dist/{chunk-NQZPZ2FN.mjs → chunk-WRE4CNPF.mjs} +6 -3
  39. package/packages/core/dist/chunk-WRE4CNPF.mjs.map +1 -0
  40. package/packages/core/dist/{chunk-A4V2H3YU.mjs → chunk-ZWA5RWMV.mjs} +40 -4
  41. package/packages/core/dist/chunk-ZWA5RWMV.mjs.map +1 -0
  42. package/packages/core/dist/{context-DjY-jvQk.d.cts → context-BbfkMYjz.d.cts} +5 -3
  43. package/packages/core/dist/{context-BDZmi53V.d.ts → context-Cp8NRbBX.d.ts} +5 -3
  44. package/packages/core/dist/diff.cjs.map +1 -1
  45. package/packages/core/dist/diff.d.cts +3 -3
  46. package/packages/core/dist/diff.d.ts +3 -3
  47. package/packages/core/dist/diff.mjs +1 -1
  48. package/packages/core/dist/exporters.cjs +137 -1
  49. package/packages/core/dist/exporters.cjs.map +1 -1
  50. package/packages/core/dist/exporters.d.cts +6 -3
  51. package/packages/core/dist/exporters.d.ts +6 -3
  52. package/packages/core/dist/exporters.mjs +71 -6
  53. package/packages/core/dist/exporters.mjs.map +1 -1
  54. package/packages/core/dist/index.cjs +181 -6
  55. package/packages/core/dist/index.cjs.map +1 -1
  56. package/packages/core/dist/index.d.cts +14 -7
  57. package/packages/core/dist/index.d.ts +14 -7
  58. package/packages/core/dist/index.mjs +51 -7
  59. package/packages/core/dist/index.mjs.map +1 -1
  60. package/packages/core/dist/{inspect-event-CYAV7Hxm.d.cts → inspect-event-IqwOHbGo.d.cts} +1 -1
  61. package/packages/core/dist/{inspect-event-CYAV7Hxm.d.ts → inspect-event-IqwOHbGo.d.ts} +1 -1
  62. package/packages/core/dist/{log-config-BNQ9UTmP.d.cts → log-config-CW8w5YQ9.d.cts} +1 -1
  63. package/packages/core/dist/{log-config-D3Yp1SZN.d.ts → log-config-SRA4xTp8.d.ts} +1 -1
  64. package/packages/core/dist/logs.d.cts +3 -3
  65. package/packages/core/dist/logs.d.ts +3 -3
  66. package/packages/core/dist/{persisted-inspect-event-BLvb0jSX.d.ts → persisted-inspect-event-DruwqARZ.d.ts} +1 -1
  67. package/packages/core/dist/{persisted-inspect-event-D5SjBz9b.d.cts → persisted-inspect-event-Gd-W8vB7.d.cts} +1 -1
  68. package/packages/core/dist/persisted.cjs +67 -1
  69. package/packages/core/dist/persisted.cjs.map +1 -1
  70. package/packages/core/dist/persisted.d.cts +5 -5
  71. package/packages/core/dist/persisted.d.ts +5 -5
  72. package/packages/core/dist/persisted.mjs +3 -3
  73. package/packages/core/dist/readers.cjs +72 -3
  74. package/packages/core/dist/readers.cjs.map +1 -1
  75. package/packages/core/dist/readers.d.cts +2 -2
  76. package/packages/core/dist/readers.d.ts +2 -2
  77. package/packages/core/dist/readers.mjs +5 -5
  78. package/packages/core/dist/readers.mjs.map +1 -1
  79. package/packages/core/dist/{types-CQun9LW6.d.ts → types-B1X87dUW.d.cts} +16 -2
  80. package/packages/core/dist/{types-C-llnPH0.d.cts → types-BVkjSd0q.d.ts} +16 -2
  81. package/packages/core/dist/types-CvsKGweN.d.cts +12 -0
  82. package/packages/core/dist/types-CvsKGweN.d.ts +12 -0
  83. package/packages/core/dist/writers.cjs +2 -1
  84. package/packages/core/dist/writers.cjs.map +1 -1
  85. package/packages/core/dist/writers.d.cts +2 -2
  86. package/packages/core/dist/writers.d.ts +2 -2
  87. package/packages/core/dist/writers.mjs +3 -4
  88. package/packages/core/dist/writers.mjs.map +1 -1
  89. package/packages/cli/dist/chunk-5VSPJEZ7.mjs.map +0 -1
  90. package/packages/core/dist/chunk-4JVGFNQH.mjs.map +0 -1
  91. package/packages/core/dist/chunk-A4V2H3YU.mjs.map +0 -1
  92. package/packages/core/dist/chunk-EDTQHZPM.mjs +0 -88
  93. package/packages/core/dist/chunk-EDTQHZPM.mjs.map +0 -1
  94. package/packages/core/dist/chunk-HY7H3CQM.mjs.map +0 -1
  95. package/packages/core/dist/chunk-IZBJAZGF.mjs.map +0 -1
  96. package/packages/core/dist/chunk-NQZPZ2FN.mjs.map +0 -1
  97. package/packages/core/dist/chunk-XUNJMAPD.mjs.map +0 -1
@@ -1,6 +1,7 @@
1
1
  import { TraceReadResult, TraceReadWarning } from './readers.cjs';
2
- import { b as InspectRunTree, c as InspectNode, A as AttributionConfidence } from './inspect-event-CYAV7Hxm.cjs';
3
- import { P as PersistedInspectEvent } from './persisted-inspect-event-D5SjBz9b.cjs';
2
+ import { b as InspectRunTree, c as InspectNode, A as AttributionConfidence } from './inspect-event-IqwOHbGo.cjs';
3
+ import { P as PersistedInspectEvent } from './persisted-inspect-event-Gd-W8vB7.cjs';
4
+ import { a as ObservedOutcomeStatus } from './types-CvsKGweN.cjs';
4
5
 
5
6
  /**
6
7
  * Experimental trace-check finding severity.
@@ -596,6 +597,13 @@ declare function createSafetyOversizedAttributeRule(options: SafetyOversizedAttr
596
597
  * evolve during the v1.x experimental period.
597
598
  */
598
599
  declare function createBaselineRegressionRule(options: BaselineRegressionRuleOptions): TraceCheckRule;
600
+ /**
601
+ * Experimental rule for observed real-world outcomes (v4.4+).
602
+ */
603
+ interface ObservedOutcomeRuleOptions {
604
+ failOn?: readonly ObservedOutcomeStatus[];
605
+ }
606
+ declare function createObservedOutcomeRule(options?: ObservedOutcomeRuleOptions): TraceCheckRule;
599
607
  /**
600
608
  * Run experimental deterministic trace checks against normalized reader output.
601
609
  *
@@ -607,4 +615,4 @@ declare function createBaselineRegressionRule(options: BaselineRegressionRuleOpt
607
615
  */
608
616
  declare function runTraceChecks(input: TraceCheckInput, options?: RunTraceChecksOptions): TraceCheckResult;
609
617
 
610
- export { type BaselineRegressionRuleOptions, type DecisionRuleOptions, type GuardrailRuleOptions, type LlmUsageRuleOptions, type MaxStepDurationRuleOptions, type RetrievalRuleOptions, type RunDepthRuleOptions, type RunDurationRuleOptions, type RunEventCountRuleOptions, type RunStatusRuleOptions, type RunTraceChecksOptions, type SafetyOversizedAttributeRuleOptions, type SafetyRawContentRuleOptions, type SafetyRedactionRuleOptions, type SafetySecretPattern, type SafetySecretPatternRuleOptions, type StallDetectionRuleOptions, type StructureIncompleteRuleOptions, type StructureOrphanRuleOptions, type StructureParallelWidthRuleOptions, type StructureRelationshipRuleOptions, type ToolFailureRuleOptions, type ToolOrderingRuleOptions, type ToolUsageRuleOptions, type TraceCheckContext, type TraceCheckDiagnostic, type TraceCheckDiagnosticCode, type TraceCheckEvidence, type TraceCheckFacts, type TraceCheckFinding, type TraceCheckFindingStatus, type TraceCheckInput, type TraceCheckResult, type TraceCheckRule, type TraceCheckRuleCategory, type TraceCheckSeverity, type TraceCheckStatus, type TraceCheckSummary, type TraceSignalRuleOptions, createBaselineRegressionRule, createDecisionRule, createGuardrailRule, createLlmUsageRule, createMaxStepDurationRule, createRequireCompletedRule, createRetrievalRule, createRunDepthRule, createRunDurationRule, createRunEventCountRule, createRunStatusRule, createSafetyOversizedAttributeRule, createSafetyRawContentRule, createSafetyRedactionRule, createSafetySecretPatternRule, createStallDetectionRule, createStructureCycleRule, createStructureIncompleteRule, createStructureOrphanRule, createStructureParallelWidthRule, createStructureRelationshipRule, createToolFailureRule, createToolOrderingRule, createToolUsageRule, runTraceChecks };
618
+ export { type BaselineRegressionRuleOptions, type DecisionRuleOptions, type GuardrailRuleOptions, type LlmUsageRuleOptions, type MaxStepDurationRuleOptions, type ObservedOutcomeRuleOptions, type RetrievalRuleOptions, type RunDepthRuleOptions, type RunDurationRuleOptions, type RunEventCountRuleOptions, type RunStatusRuleOptions, type RunTraceChecksOptions, type SafetyOversizedAttributeRuleOptions, type SafetyRawContentRuleOptions, type SafetyRedactionRuleOptions, type SafetySecretPattern, type SafetySecretPatternRuleOptions, type StallDetectionRuleOptions, type StructureIncompleteRuleOptions, type StructureOrphanRuleOptions, type StructureParallelWidthRuleOptions, type StructureRelationshipRuleOptions, type ToolFailureRuleOptions, type ToolOrderingRuleOptions, type ToolUsageRuleOptions, type TraceCheckContext, type TraceCheckDiagnostic, type TraceCheckDiagnosticCode, type TraceCheckEvidence, type TraceCheckFacts, type TraceCheckFinding, type TraceCheckFindingStatus, type TraceCheckInput, type TraceCheckResult, type TraceCheckRule, type TraceCheckRuleCategory, type TraceCheckSeverity, type TraceCheckStatus, type TraceCheckSummary, type TraceSignalRuleOptions, createBaselineRegressionRule, createDecisionRule, createGuardrailRule, createLlmUsageRule, createMaxStepDurationRule, createObservedOutcomeRule, createRequireCompletedRule, createRetrievalRule, createRunDepthRule, createRunDurationRule, createRunEventCountRule, createRunStatusRule, createSafetyOversizedAttributeRule, createSafetyRawContentRule, createSafetyRedactionRule, createSafetySecretPatternRule, createStallDetectionRule, createStructureCycleRule, createStructureIncompleteRule, createStructureOrphanRule, createStructureParallelWidthRule, createStructureRelationshipRule, createToolFailureRule, createToolOrderingRule, createToolUsageRule, runTraceChecks };
@@ -1,6 +1,7 @@
1
1
  import { TraceReadResult, TraceReadWarning } from './readers.js';
2
- import { b as InspectRunTree, c as InspectNode, A as AttributionConfidence } from './inspect-event-CYAV7Hxm.js';
3
- import { P as PersistedInspectEvent } from './persisted-inspect-event-BLvb0jSX.js';
2
+ import { b as InspectRunTree, c as InspectNode, A as AttributionConfidence } from './inspect-event-IqwOHbGo.js';
3
+ import { P as PersistedInspectEvent } from './persisted-inspect-event-DruwqARZ.js';
4
+ import { a as ObservedOutcomeStatus } from './types-CvsKGweN.js';
4
5
 
5
6
  /**
6
7
  * Experimental trace-check finding severity.
@@ -596,6 +597,13 @@ declare function createSafetyOversizedAttributeRule(options: SafetyOversizedAttr
596
597
  * evolve during the v1.x experimental period.
597
598
  */
598
599
  declare function createBaselineRegressionRule(options: BaselineRegressionRuleOptions): TraceCheckRule;
600
+ /**
601
+ * Experimental rule for observed real-world outcomes (v4.4+).
602
+ */
603
+ interface ObservedOutcomeRuleOptions {
604
+ failOn?: readonly ObservedOutcomeStatus[];
605
+ }
606
+ declare function createObservedOutcomeRule(options?: ObservedOutcomeRuleOptions): TraceCheckRule;
599
607
  /**
600
608
  * Run experimental deterministic trace checks against normalized reader output.
601
609
  *
@@ -607,4 +615,4 @@ declare function createBaselineRegressionRule(options: BaselineRegressionRuleOpt
607
615
  */
608
616
  declare function runTraceChecks(input: TraceCheckInput, options?: RunTraceChecksOptions): TraceCheckResult;
609
617
 
610
- export { type BaselineRegressionRuleOptions, type DecisionRuleOptions, type GuardrailRuleOptions, type LlmUsageRuleOptions, type MaxStepDurationRuleOptions, type RetrievalRuleOptions, type RunDepthRuleOptions, type RunDurationRuleOptions, type RunEventCountRuleOptions, type RunStatusRuleOptions, type RunTraceChecksOptions, type SafetyOversizedAttributeRuleOptions, type SafetyRawContentRuleOptions, type SafetyRedactionRuleOptions, type SafetySecretPattern, type SafetySecretPatternRuleOptions, type StallDetectionRuleOptions, type StructureIncompleteRuleOptions, type StructureOrphanRuleOptions, type StructureParallelWidthRuleOptions, type StructureRelationshipRuleOptions, type ToolFailureRuleOptions, type ToolOrderingRuleOptions, type ToolUsageRuleOptions, type TraceCheckContext, type TraceCheckDiagnostic, type TraceCheckDiagnosticCode, type TraceCheckEvidence, type TraceCheckFacts, type TraceCheckFinding, type TraceCheckFindingStatus, type TraceCheckInput, type TraceCheckResult, type TraceCheckRule, type TraceCheckRuleCategory, type TraceCheckSeverity, type TraceCheckStatus, type TraceCheckSummary, type TraceSignalRuleOptions, createBaselineRegressionRule, createDecisionRule, createGuardrailRule, createLlmUsageRule, createMaxStepDurationRule, createRequireCompletedRule, createRetrievalRule, createRunDepthRule, createRunDurationRule, createRunEventCountRule, createRunStatusRule, createSafetyOversizedAttributeRule, createSafetyRawContentRule, createSafetyRedactionRule, createSafetySecretPatternRule, createStallDetectionRule, createStructureCycleRule, createStructureIncompleteRule, createStructureOrphanRule, createStructureParallelWidthRule, createStructureRelationshipRule, createToolFailureRule, createToolOrderingRule, createToolUsageRule, runTraceChecks };
618
+ export { type BaselineRegressionRuleOptions, type DecisionRuleOptions, type GuardrailRuleOptions, type LlmUsageRuleOptions, type MaxStepDurationRuleOptions, type ObservedOutcomeRuleOptions, type RetrievalRuleOptions, type RunDepthRuleOptions, type RunDurationRuleOptions, type RunEventCountRuleOptions, type RunStatusRuleOptions, type RunTraceChecksOptions, type SafetyOversizedAttributeRuleOptions, type SafetyRawContentRuleOptions, type SafetyRedactionRuleOptions, type SafetySecretPattern, type SafetySecretPatternRuleOptions, type StallDetectionRuleOptions, type StructureIncompleteRuleOptions, type StructureOrphanRuleOptions, type StructureParallelWidthRuleOptions, type StructureRelationshipRuleOptions, type ToolFailureRuleOptions, type ToolOrderingRuleOptions, type ToolUsageRuleOptions, type TraceCheckContext, type TraceCheckDiagnostic, type TraceCheckDiagnosticCode, type TraceCheckEvidence, type TraceCheckFacts, type TraceCheckFinding, type TraceCheckFindingStatus, type TraceCheckInput, type TraceCheckResult, type TraceCheckRule, type TraceCheckRuleCategory, type TraceCheckSeverity, type TraceCheckStatus, type TraceCheckSummary, type TraceSignalRuleOptions, createBaselineRegressionRule, createDecisionRule, createGuardrailRule, createLlmUsageRule, createMaxStepDurationRule, createObservedOutcomeRule, createRequireCompletedRule, createRetrievalRule, createRunDepthRule, createRunDurationRule, createRunEventCountRule, createRunStatusRule, createSafetyOversizedAttributeRule, createSafetyRawContentRule, createSafetyRedactionRule, createSafetySecretPatternRule, createStallDetectionRule, createStructureCycleRule, createStructureIncompleteRule, createStructureOrphanRule, createStructureParallelWidthRule, createStructureRelationshipRule, createToolFailureRule, createToolOrderingRule, createToolUsageRule, runTraceChecks };
@@ -1,3 +1,10 @@
1
+ import { extractOutcomesFromPersistedEvents, outcomesMatchingStatus } from './chunk-VFO76UH3.mjs';
2
+ import './chunk-ULCGIRTC.mjs';
3
+ import './chunk-VU6O5QAH.mjs';
4
+ import './chunk-WRE4CNPF.mjs';
5
+ import './chunk-P366CCL2.mjs';
6
+ import './chunk-7TGZLWEE.mjs';
7
+
1
8
  // packages/core/src/checks/index.ts
2
9
  var SEVERITY_RANK = {
3
10
  error: 0,
@@ -1564,6 +1571,40 @@ function createBaselineRegressionRule(options) {
1564
1571
  }
1565
1572
  };
1566
1573
  }
1574
+ function createObservedOutcomeRule(options = {}) {
1575
+ const failOn = options.failOn ?? ["failed"];
1576
+ return {
1577
+ id: "outcome.status",
1578
+ category: "run",
1579
+ defaultSeverity: "error",
1580
+ evaluate(context) {
1581
+ const outcomes = extractOutcomesFromPersistedEvents(context.events);
1582
+ const matching = outcomesMatchingStatus(outcomes, failOn);
1583
+ if (matching.length === 0) return [];
1584
+ return [
1585
+ failFinding(
1586
+ "outcome.status",
1587
+ `Observed outcome count ${matching.length} matched [${failOn.join(", ")}].`,
1588
+ matching.map((outcome) => ({
1589
+ runId: outcome.runId,
1590
+ eventId: outcome.outcomeId,
1591
+ ...outcome.parentId !== void 0 ? { parentId: outcome.parentId } : {},
1592
+ kind: "OUTCOME",
1593
+ name: outcome.name,
1594
+ status: outcome.status,
1595
+ path: `outcome.${outcome.name}`
1596
+ })),
1597
+ { failOn },
1598
+ matching.map((outcome) => ({
1599
+ name: outcome.name,
1600
+ status: outcome.status,
1601
+ expectation: outcome.expectation
1602
+ }))
1603
+ )
1604
+ ];
1605
+ }
1606
+ };
1607
+ }
1567
1608
  function runTraceChecks(input, options = {}) {
1568
1609
  const selected = resolveSelectedRun(input, options.runId);
1569
1610
  if (selected.diagnostics.length > 0) {
@@ -1609,6 +1650,6 @@ function runTraceChecks(input, options = {}) {
1609
1650
  };
1610
1651
  }
1611
1652
 
1612
- export { createBaselineRegressionRule, createDecisionRule, createGuardrailRule, createLlmUsageRule, createMaxStepDurationRule, createRequireCompletedRule, createRetrievalRule, createRunDepthRule, createRunDurationRule, createRunEventCountRule, createRunStatusRule, createSafetyOversizedAttributeRule, createSafetyRawContentRule, createSafetyRedactionRule, createSafetySecretPatternRule, createStallDetectionRule, createStructureCycleRule, createStructureIncompleteRule, createStructureOrphanRule, createStructureParallelWidthRule, createStructureRelationshipRule, createToolFailureRule, createToolOrderingRule, createToolUsageRule, runTraceChecks };
1653
+ export { createBaselineRegressionRule, createDecisionRule, createGuardrailRule, createLlmUsageRule, createMaxStepDurationRule, createObservedOutcomeRule, createRequireCompletedRule, createRetrievalRule, createRunDepthRule, createRunDurationRule, createRunEventCountRule, createRunStatusRule, createSafetyOversizedAttributeRule, createSafetyRawContentRule, createSafetyRedactionRule, createSafetySecretPatternRule, createStallDetectionRule, createStructureCycleRule, createStructureIncompleteRule, createStructureOrphanRule, createStructureParallelWidthRule, createStructureRelationshipRule, createToolFailureRule, createToolOrderingRule, createToolUsageRule, runTraceChecks };
1613
1654
  //# sourceMappingURL=checks.mjs.map
1614
1655
  //# sourceMappingURL=checks.mjs.map