agent-inspect 6.31.17 → 6.31.18
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.
- package/CHANGELOG.md +6 -0
- package/README.md +1 -1
- package/docs/COMPARE.md +1 -1
- package/package.json +2 -2
- package/packages/cli/dist/{chunk-USZJEEME.mjs → chunk-AHARYDFO.mjs} +58 -24
- package/packages/cli/dist/chunk-AHARYDFO.mjs.map +1 -0
- package/packages/cli/dist/index.cjs +112 -25
- package/packages/cli/dist/index.cjs.map +1 -1
- package/packages/cli/dist/index.mjs +58 -6
- package/packages/cli/dist/index.mjs.map +1 -1
- package/packages/cli/dist/{src-EY4EY456.mjs → src-YNZZ2KYR.mjs} +3 -3
- package/packages/cli/dist/{src-EY4EY456.mjs.map → src-YNZZ2KYR.mjs.map} +1 -1
- package/packages/core/dist/advanced.mjs +2 -2
- package/packages/core/dist/checks.cjs +56 -22
- package/packages/core/dist/checks.cjs.map +1 -1
- package/packages/core/dist/checks.mjs +1 -1
- package/packages/core/dist/{chunk-GTA6SNDK.mjs → chunk-FRBBZUEP.mjs} +58 -24
- package/packages/core/dist/chunk-FRBBZUEP.mjs.map +1 -0
- package/packages/core/dist/exporters.cjs +54 -2
- package/packages/core/dist/exporters.cjs.map +1 -1
- package/packages/core/dist/exporters.mjs +54 -2
- package/packages/core/dist/exporters.mjs.map +1 -1
- package/packages/cli/dist/chunk-USZJEEME.mjs.map +0 -1
- package/packages/core/dist/chunk-GTA6SNDK.mjs.map +0 -1
|
@@ -1,4 +1,4 @@
|
|
|
1
|
-
export { buildTraceFacts, createBaselineRegressionRule, createDecisionRule, createGuardrailRule, createLlmUsageRule, createMaxStepDurationRule, createObservedOutcomeRule, createRequireCompletedRule, createRetrievalRule, createRunDepthRule, createRunDurationRule, createRunEventCountRule, createRunStatusRule, createSafetyOversizedAttributeRule, createSafetyRawContentRule, createSafetyRedactionRule, createSafetySecretPatternRule, createStallDetectionRule, createStepOrderingRule, createStructureCycleRule, createStructureIncompleteRule, createStructureOrphanRule, createStructureParallelWidthRule, createStructureRelationshipRule, createToolFailureRule, createToolOrderingRule, createToolUsageRule, defineTraceContract, deriveFailureFacts, deriveRelationshipFacts, evaluateToolArgumentValue, evaluateTraceContract, evaluateTraceContractRead, explainTraceContract, extractToolArgumentPayload, lintTraceContract, resolveJsonPointer, resolveTraceContractScope, runTraceChecks, summarizeSemanticParity, workflowMetadataForRun } from './chunk-
|
|
1
|
+
export { buildTraceFacts, createBaselineRegressionRule, createDecisionRule, createGuardrailRule, createLlmUsageRule, createMaxStepDurationRule, createObservedOutcomeRule, createRequireCompletedRule, createRetrievalRule, createRunDepthRule, createRunDurationRule, createRunEventCountRule, createRunStatusRule, createSafetyOversizedAttributeRule, createSafetyRawContentRule, createSafetyRedactionRule, createSafetySecretPatternRule, createStallDetectionRule, createStepOrderingRule, createStructureCycleRule, createStructureIncompleteRule, createStructureOrphanRule, createStructureParallelWidthRule, createStructureRelationshipRule, createToolFailureRule, createToolOrderingRule, createToolUsageRule, defineTraceContract, deriveFailureFacts, deriveRelationshipFacts, evaluateToolArgumentValue, evaluateTraceContract, evaluateTraceContractRead, explainTraceContract, extractToolArgumentPayload, lintTraceContract, resolveJsonPointer, resolveTraceContractScope, runTraceChecks, summarizeSemanticParity, workflowMetadataForRun } from './chunk-FRBBZUEP.mjs';
|
|
2
2
|
import './chunk-77UZIMMD.mjs';
|
|
3
3
|
import './chunk-EOXUEQB6.mjs';
|
|
4
4
|
import './chunk-VLVCLAPB.mjs';
|
|
@@ -1238,24 +1238,26 @@ function attemptIdOf(event) {
|
|
|
1238
1238
|
return typeof value === "string" && value.trim() !== "" ? value : void 0;
|
|
1239
1239
|
}
|
|
1240
1240
|
function countOperationAttempts(members) {
|
|
1241
|
-
const
|
|
1242
|
-
|
|
1243
|
-
)
|
|
1244
|
-
|
|
1245
|
-
|
|
1241
|
+
const identified = /* @__PURE__ */ new Set();
|
|
1242
|
+
let unidentified = 0;
|
|
1243
|
+
for (const event of members) {
|
|
1244
|
+
if (event.kind !== "TOOL" && event.kind !== "LLM") continue;
|
|
1245
|
+
const id = attemptIdOf(event);
|
|
1246
|
+
if (id !== void 0) {
|
|
1247
|
+
identified.add(id);
|
|
1248
|
+
continue;
|
|
1249
|
+
}
|
|
1250
|
+
unidentified += 1;
|
|
1251
|
+
}
|
|
1252
|
+
if (identified.size > 0) {
|
|
1253
|
+
return identified.size + unidentified;
|
|
1246
1254
|
}
|
|
1247
1255
|
const numbers = members.map((event) => attemptNumberOf(event)).filter((value) => value !== void 0).sort((a, b) => a - b);
|
|
1248
1256
|
if (numbers.length > 0) {
|
|
1249
1257
|
const unique = [...new Set(numbers)];
|
|
1250
|
-
const min = unique[0];
|
|
1251
|
-
const max = unique[unique.length - 1];
|
|
1252
|
-
const contiguous = unique.length === max - min + 1 && unique.every((value, index) => value === min + index);
|
|
1253
|
-
if (contiguous) {
|
|
1254
|
-
return unique.length;
|
|
1255
|
-
}
|
|
1256
1258
|
return unique.length;
|
|
1257
1259
|
}
|
|
1258
|
-
return
|
|
1260
|
+
return unidentified;
|
|
1259
1261
|
}
|
|
1260
1262
|
function classifyRetryAttempt(event, priorOrdered, allMembers) {
|
|
1261
1263
|
const workflow = workflowFor(event);
|
|
@@ -1530,8 +1532,30 @@ function eventEvidence2(event) {
|
|
|
1530
1532
|
function eventTime2(event) {
|
|
1531
1533
|
return event.startedAt ?? event.timestamp ?? "";
|
|
1532
1534
|
}
|
|
1535
|
+
function parseComparableMs(value) {
|
|
1536
|
+
if (!value) return void 0;
|
|
1537
|
+
const parsed = Date.parse(value);
|
|
1538
|
+
return Number.isFinite(parsed) ? parsed : void 0;
|
|
1539
|
+
}
|
|
1540
|
+
function consumerStartMs(event) {
|
|
1541
|
+
return parseComparableMs(event.startedAt) ?? parseComparableMs(event.timestamp);
|
|
1542
|
+
}
|
|
1543
|
+
function resultAvailableMs(event) {
|
|
1544
|
+
const ended = parseComparableMs(event.endedAt);
|
|
1545
|
+
if (ended !== void 0) return ended;
|
|
1546
|
+
const started = parseComparableMs(event.startedAt) ?? parseComparableMs(event.timestamp);
|
|
1547
|
+
if (started !== void 0 && event.durationMs !== void 0 && Number.isFinite(event.durationMs)) {
|
|
1548
|
+
return started + event.durationMs;
|
|
1549
|
+
}
|
|
1550
|
+
return parseComparableMs(event.timestamp);
|
|
1551
|
+
}
|
|
1533
1552
|
function sortByTime2(events) {
|
|
1534
|
-
return [...events].sort((a, b) =>
|
|
1553
|
+
return [...events].sort((a, b) => {
|
|
1554
|
+
const aMs = consumerStartMs(a) ?? 0;
|
|
1555
|
+
const bMs = consumerStartMs(b) ?? 0;
|
|
1556
|
+
if (aMs !== bMs) return aMs - bMs;
|
|
1557
|
+
return eventTime2(a).localeCompare(eventTime2(b));
|
|
1558
|
+
});
|
|
1535
1559
|
}
|
|
1536
1560
|
function errorCodeOf(event) {
|
|
1537
1561
|
if (typeof event.error?.code === "string" && event.error.code.trim() !== "") {
|
|
@@ -1599,13 +1623,20 @@ function sameArguments(left, right) {
|
|
|
1599
1623
|
}
|
|
1600
1624
|
return { ok: true };
|
|
1601
1625
|
}
|
|
1626
|
+
var UNEVALUABLE_WRITE_ERROR_CODES = /* @__PURE__ */ new Set([
|
|
1627
|
+
"TIMEOUT",
|
|
1628
|
+
"ETIMEDOUT",
|
|
1629
|
+
"UNKNOWN_COMPLETION",
|
|
1630
|
+
"DEADLINE_EXCEEDED"
|
|
1631
|
+
]);
|
|
1602
1632
|
function isUnevaluableWriteCompletion(event) {
|
|
1603
1633
|
if (event.status === "running" || event.status === "unknown") return true;
|
|
1604
1634
|
const attrs2 = event.attributes ?? {};
|
|
1605
1635
|
if (attrs2.timeout === true) return true;
|
|
1636
|
+
if (attrs2.unknownCompletion === true) return true;
|
|
1606
1637
|
if (attrs2.completionState === "unknown" || attrs2.completionState === "timeout") return true;
|
|
1607
1638
|
const code = errorCodeOf(event)?.toUpperCase();
|
|
1608
|
-
if (code
|
|
1639
|
+
if (code && UNEVALUABLE_WRITE_ERROR_CODES.has(code)) {
|
|
1609
1640
|
return true;
|
|
1610
1641
|
}
|
|
1611
1642
|
return false;
|
|
@@ -1653,14 +1684,17 @@ function collectExplicitReferences(event) {
|
|
|
1653
1684
|
return refs;
|
|
1654
1685
|
}
|
|
1655
1686
|
function llmReferencesTool(llm, toolEvent, requireExplicit) {
|
|
1656
|
-
if (
|
|
1657
|
-
|
|
1658
|
-
|
|
1659
|
-
|
|
1660
|
-
|
|
1661
|
-
|
|
1662
|
-
|
|
1663
|
-
|
|
1687
|
+
if (llm.runId !== toolEvent.runId) return false;
|
|
1688
|
+
if (requireExplicit) {
|
|
1689
|
+
const refs = collectExplicitReferences(llm);
|
|
1690
|
+
const toolCallId = workflowFor2(toolEvent).toolCallId;
|
|
1691
|
+
const hasRef = refs.has(toolEvent.eventId) || toolCallId !== void 0 && refs.has(toolCallId);
|
|
1692
|
+
if (!hasRef) return false;
|
|
1693
|
+
}
|
|
1694
|
+
const resultMs = resultAvailableMs(toolEvent);
|
|
1695
|
+
const consumerMs = consumerStartMs(llm);
|
|
1696
|
+
if (resultMs === void 0 || consumerMs === void 0) return false;
|
|
1697
|
+
return resultMs <= consumerMs;
|
|
1664
1698
|
}
|
|
1665
1699
|
function attemptIdOf2(event) {
|
|
1666
1700
|
const value = workflowFor2(event).attemptId;
|
|
@@ -5254,5 +5288,5 @@ function runTraceChecks(input, options = {}) {
|
|
|
5254
5288
|
}
|
|
5255
5289
|
|
|
5256
5290
|
export { buildTraceFacts, createBaselineRegressionRule, createDecisionRule, createGuardrailRule, createLlmUsageRule, createMaxStepDurationRule, createObservedOutcomeRule, createRequireCompletedRule, createRetrievalRule, createRunDepthRule, createRunDurationRule, createRunEventCountRule, createRunStatusRule, createSafetyOversizedAttributeRule, createSafetyRawContentRule, createSafetyRedactionRule, createSafetySecretPatternRule, createStallDetectionRule, createStepOrderingRule, createStructureCycleRule, createStructureIncompleteRule, createStructureOrphanRule, createStructureParallelWidthRule, createStructureRelationshipRule, createToolFailureRule, createToolOrderingRule, createToolUsageRule, defineTraceContract, deriveFailureFacts, deriveRelationshipFacts, evaluateToolArgumentValue, evaluateTraceContract, evaluateTraceContractRead, explainTraceContract, extractSessionWorkflowMetadata, extractToolArgumentPayload, lintTraceContract, resolveJsonPointer, resolveTraceContractScope, runTraceChecks, sessionKeyForRun, summarizeSemanticParity, workflowMetadataForRun };
|
|
5257
|
-
//# sourceMappingURL=chunk-
|
|
5258
|
-
//# sourceMappingURL=chunk-
|
|
5291
|
+
//# sourceMappingURL=chunk-FRBBZUEP.mjs.map
|
|
5292
|
+
//# sourceMappingURL=chunk-FRBBZUEP.mjs.map
|