audrey 1.1.0 → 1.1.2

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 (83) hide show
  1. package/CHANGELOG.md +27 -0
  2. package/README.md +16 -7
  3. package/benchmarks/output/adapter-self-test/guardbench-adapter-self-test.json +4 -4
  4. package/benchmarks/output/external/guardbench-external-dry-run.json +1 -1
  5. package/benchmarks/output/external/guardbench-external-evidence.json +1 -1
  6. package/benchmarks/output/guardbench-conformance-card.json +8 -8
  7. package/benchmarks/output/guardbench-raw.json +122 -114
  8. package/benchmarks/output/guardbench-summary.json +205 -189
  9. package/benchmarks/output/leaderboard/guardbench-leaderboard.json +5 -5
  10. package/benchmarks/output/leaderboard/guardbench-leaderboard.md +2 -2
  11. package/benchmarks/output/submission-bundle/guardbench-conformance-card.json +8 -8
  12. package/benchmarks/output/submission-bundle/guardbench-raw.json +122 -114
  13. package/benchmarks/output/submission-bundle/guardbench-summary.json +205 -189
  14. package/benchmarks/output/submission-bundle/submission-manifest.json +10 -10
  15. package/benchmarks/output/submission-bundle/validation-report.json +1 -1
  16. package/benchmarks/output/summary.json +44 -44
  17. package/dist/mcp-server/config.d.ts +1 -1
  18. package/dist/mcp-server/config.js +1 -1
  19. package/dist/mcp-server/index.d.ts.map +1 -1
  20. package/dist/mcp-server/index.js +88 -37
  21. package/dist/mcp-server/index.js.map +1 -1
  22. package/dist/src/audrey.d.ts +2 -0
  23. package/dist/src/audrey.d.ts.map +1 -1
  24. package/dist/src/audrey.js +2 -0
  25. package/dist/src/audrey.js.map +1 -1
  26. package/dist/src/autopilot.d.ts +3 -1
  27. package/dist/src/autopilot.d.ts.map +1 -1
  28. package/dist/src/autopilot.js +150 -37
  29. package/dist/src/autopilot.js.map +1 -1
  30. package/dist/src/capsule.d.ts +2 -0
  31. package/dist/src/capsule.d.ts.map +1 -1
  32. package/dist/src/capsule.js +43 -3
  33. package/dist/src/capsule.js.map +1 -1
  34. package/dist/src/controller.d.ts.map +1 -1
  35. package/dist/src/controller.js +7 -3
  36. package/dist/src/controller.js.map +1 -1
  37. package/dist/src/events.d.ts +23 -6
  38. package/dist/src/events.d.ts.map +1 -1
  39. package/dist/src/events.js +80 -13
  40. package/dist/src/events.js.map +1 -1
  41. package/dist/src/preflight.d.ts.map +1 -1
  42. package/dist/src/preflight.js +19 -6
  43. package/dist/src/preflight.js.map +1 -1
  44. package/dist/src/project.d.ts +17 -0
  45. package/dist/src/project.d.ts.map +1 -0
  46. package/dist/src/project.js +56 -0
  47. package/dist/src/project.js.map +1 -0
  48. package/dist/src/redact.d.ts.map +1 -1
  49. package/dist/src/redact.js +20 -0
  50. package/dist/src/redact.js.map +1 -1
  51. package/docs/paper/07-evaluation.md +5 -5
  52. package/docs/paper/audrey-paper-v1.md +5 -5
  53. package/docs/paper/evidence-ledger.md +1 -1
  54. package/docs/paper/output/arxiv/arxiv-manifest.json +4 -4
  55. package/docs/paper/output/arxiv/main.tex +5 -5
  56. package/docs/paper/output/arxiv-compile-report.json +3 -3
  57. package/docs/paper/output/submission-bundle/README.md +16 -7
  58. package/docs/paper/output/submission-bundle/benchmarks/output/adapter-self-test/guardbench-adapter-self-test.json +4 -4
  59. package/docs/paper/output/submission-bundle/benchmarks/output/external/guardbench-external-dry-run.json +1 -1
  60. package/docs/paper/output/submission-bundle/benchmarks/output/external/guardbench-external-evidence.json +1 -1
  61. package/docs/paper/output/submission-bundle/benchmarks/output/guardbench-conformance-card.json +8 -8
  62. package/docs/paper/output/submission-bundle/benchmarks/output/guardbench-raw.json +122 -114
  63. package/docs/paper/output/submission-bundle/benchmarks/output/guardbench-summary.json +205 -189
  64. package/docs/paper/output/submission-bundle/benchmarks/output/leaderboard/guardbench-leaderboard.json +5 -5
  65. package/docs/paper/output/submission-bundle/benchmarks/output/leaderboard/guardbench-leaderboard.md +2 -2
  66. package/docs/paper/output/submission-bundle/benchmarks/output/submission-bundle/submission-manifest.json +10 -10
  67. package/docs/paper/output/submission-bundle/benchmarks/output/submission-bundle/validation-report.json +1 -1
  68. package/docs/paper/output/submission-bundle/benchmarks/output/summary.json +42 -42
  69. package/docs/paper/output/submission-bundle/docs/paper/07-evaluation.md +5 -5
  70. package/docs/paper/output/submission-bundle/docs/paper/audrey-paper-v1.md +5 -5
  71. package/docs/paper/output/submission-bundle/docs/paper/evidence-ledger.md +1 -1
  72. package/docs/paper/output/submission-bundle/docs/paper/output/arxiv/arxiv-manifest.json +4 -4
  73. package/docs/paper/output/submission-bundle/docs/paper/output/arxiv/main.tex +5 -5
  74. package/docs/paper/output/submission-bundle/docs/paper/output/arxiv-compile-report.json +3 -3
  75. package/docs/paper/output/submission-bundle/package.json +8 -1
  76. package/docs/paper/output/submission-bundle/paper-submission-manifest.json +31 -31
  77. package/package.json +8 -1
  78. package/scripts/audit-release-completion.mjs +1 -1
  79. package/scripts/finalize-release.mjs +1 -1
  80. package/scripts/prepare-release-cut.mjs +1 -1
  81. package/scripts/publish-release-bundle.mjs +1 -1
  82. package/scripts/publish-release-github-api.mjs +1 -1
  83. package/scripts/verify-release-readiness.mjs +1 -1
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "suite": "GuardBench comparative",
3
- "generatedAt": "2026-07-10T15:00:41.780Z",
3
+ "generatedAt": "2026-07-13T14:48:49.312Z",
4
4
  "manifest": {
5
5
  "manifestVersion": "0.2.0",
6
6
  "suiteId": "guardbench-local-comparative",
@@ -416,8 +416,8 @@
416
416
  ]
417
417
  },
418
418
  "provenance": {
419
- "generatedAt": "2026-07-10T15:00:41.781Z",
420
- "gitSha": "af2534b9dc5b980fb879e88ca63afd502480dd24",
419
+ "generatedAt": "2026-07-13T14:48:49.313Z",
420
+ "gitSha": "7111c630be69d3ab96439018f094dc4fac826aca",
421
421
  "gitDirty": false,
422
422
  "node": "v24.18.0",
423
423
  "v8": "13.6.233.17-node.50",
@@ -449,14 +449,14 @@
449
449
  "redactionLeaks": 0,
450
450
  "recallDegradationDetectionRate": 1,
451
451
  "latency": {
452
- "p50Ms": 3.501,
453
- "p95Ms": 20.658,
454
- "maxMs": 20.658
452
+ "p50Ms": 4.253,
453
+ "p95Ms": 21.196,
454
+ "maxMs": 21.196
455
455
  },
456
456
  "systemSummaries": [
457
457
  {
458
458
  "system": "Audrey Guard",
459
- "generatedAt": "2026-07-10T15:00:41.779Z",
459
+ "generatedAt": "2026-07-13T14:48:49.312Z",
460
460
  "scenarios": 10,
461
461
  "passed": 10,
462
462
  "passRate": 1,
@@ -470,14 +470,14 @@
470
470
  "redactionLeaks": 0,
471
471
  "recallDegradationDetectionRate": 1,
472
472
  "latency": {
473
- "p50Ms": 3.501,
474
- "p95Ms": 20.658,
475
- "maxMs": 20.658
473
+ "p50Ms": 4.253,
474
+ "p95Ms": 21.196,
475
+ "maxMs": 21.196
476
476
  }
477
477
  },
478
478
  {
479
479
  "system": "No Memory",
480
- "generatedAt": "2026-07-10T15:00:41.779Z",
480
+ "generatedAt": "2026-07-13T14:48:49.312Z",
481
481
  "scenarios": 10,
482
482
  "passed": 1,
483
483
  "passRate": 0.1,
@@ -498,7 +498,7 @@
498
498
  },
499
499
  {
500
500
  "system": "Recent Window",
501
- "generatedAt": "2026-07-10T15:00:41.779Z",
501
+ "generatedAt": "2026-07-13T14:48:49.312Z",
502
502
  "scenarios": 10,
503
503
  "passed": 6,
504
504
  "passRate": 0.6,
@@ -512,14 +512,14 @@
512
512
  "redactionLeaks": 0,
513
513
  "recallDegradationDetectionRate": 0.5,
514
514
  "latency": {
515
- "p50Ms": 0.112,
516
- "p95Ms": 0.46,
517
- "maxMs": 0.46
515
+ "p50Ms": 0.11,
516
+ "p95Ms": 0.427,
517
+ "maxMs": 0.427
518
518
  }
519
519
  },
520
520
  {
521
521
  "system": "Vector Only",
522
- "generatedAt": "2026-07-10T15:00:41.779Z",
522
+ "generatedAt": "2026-07-13T14:48:49.312Z",
523
523
  "scenarios": 10,
524
524
  "passed": 4,
525
525
  "passRate": 0.4,
@@ -533,14 +533,14 @@
533
533
  "redactionLeaks": 0,
534
534
  "recallDegradationDetectionRate": 0,
535
535
  "latency": {
536
- "p50Ms": 0.446,
537
- "p95Ms": 1.908,
538
- "maxMs": 1.908
536
+ "p50Ms": 0.465,
537
+ "p95Ms": 1.589,
538
+ "maxMs": 1.589
539
539
  }
540
540
  },
541
541
  {
542
542
  "system": "FTS Only",
543
- "generatedAt": "2026-07-10T15:00:41.779Z",
543
+ "generatedAt": "2026-07-13T14:48:49.312Z",
544
544
  "scenarios": 10,
545
545
  "passed": 1,
546
546
  "passRate": 0.1,
@@ -554,16 +554,16 @@
554
554
  "redactionLeaks": 0,
555
555
  "recallDegradationDetectionRate": 0,
556
556
  "latency": {
557
- "p50Ms": 0.414,
558
- "p95Ms": 0.716,
559
- "maxMs": 0.716
557
+ "p50Ms": 0.423,
558
+ "p95Ms": 0.739,
559
+ "maxMs": 0.739
560
560
  }
561
561
  }
562
562
  ],
563
563
  "comparisons": {
564
564
  "bestBaseline": {
565
565
  "system": "Recent Window",
566
- "generatedAt": "2026-07-10T15:00:41.779Z",
566
+ "generatedAt": "2026-07-13T14:48:49.312Z",
567
567
  "scenarios": 10,
568
568
  "passed": 6,
569
569
  "passRate": 0.6,
@@ -577,9 +577,9 @@
577
577
  "redactionLeaks": 0,
578
578
  "recallDegradationDetectionRate": 0.5,
579
579
  "latency": {
580
- "p50Ms": 0.112,
581
- "p95Ms": 0.46,
582
- "maxMs": 0.46
580
+ "p50Ms": 0.11,
581
+ "p95Ms": 0.427,
582
+ "maxMs": 0.427
583
583
  }
584
584
  },
585
585
  "audreyMarginOverBestBaseline": 0.4
@@ -594,17 +594,19 @@
594
594
  "decisionCorrect": true,
595
595
  "riskScore": 0.9,
596
596
  "passed": true,
597
- "latencyMs": 9.146,
598
- "evidenceCount": 2,
597
+ "latencyMs": 11.267,
598
+ "evidenceCount": 3,
599
599
  "evidenceIds": [
600
- "01KX68PZF1G0JED770QTGDF0AB",
601
- "failure:Bash:2026-07-10T15:00:37.986Z"
600
+ "01KXDZ7CQ6VDPVF6JH0HH1KM2A",
601
+ "failure:Bash:2026-07-13T14:48:45.542Z",
602
+ "01KXDZ7CQ73T2NW3XX0033D6F7"
602
603
  ],
603
604
  "recommendedActions": [
604
605
  "Do not repeat the exact failed action until the prior error is understood or the command is changed.",
605
- "Before re-running Bash, check what changed since the last failure."
606
+ "Before re-running Bash, check what changed since the last failure.",
607
+ "Mitigate this remembered risk before proceeding."
606
608
  ],
607
- "summary": "Blocked: this exact Bash action failed before. Caution: 1 memory signal, 1 medium severity found before acting.",
609
+ "summary": "Blocked: this exact Bash action failed before. Caution: 2 memory signals, 2 medium severity found before acting.",
608
610
  "recallErrors": [],
609
611
  "leakedSecrets": [],
610
612
  "hasEvidenceForDecision": true,
@@ -620,10 +622,10 @@
620
622
  "decisionCorrect": true,
621
623
  "riskScore": 0.85,
622
624
  "passed": true,
623
- "latencyMs": 2.944,
625
+ "latencyMs": 3.735,
624
626
  "evidenceCount": 1,
625
627
  "evidenceIds": [
626
- "01KX68PZHF6HCQV61BYWZZS1GJ"
628
+ "01KXDZ7CSN0BGBJM8JA5SMHZWT"
627
629
  ],
628
630
  "recommendedActions": [
629
631
  "Do not proceed until the high-severity memory warning is addressed.",
@@ -645,15 +647,17 @@
645
647
  "decisionCorrect": true,
646
648
  "riskScore": 0.55,
647
649
  "passed": true,
648
- "latencyMs": 5.572,
649
- "evidenceCount": 1,
650
+ "latencyMs": 4.478,
651
+ "evidenceCount": 2,
650
652
  "evidenceIds": [
651
- "failure:Bash:2026-07-10T15:00:38.129Z"
653
+ "failure:Bash:2026-07-13T14:48:45.688Z",
654
+ "01KXDZ7CVSXNTMNMP89QBMCXYA"
652
655
  ],
653
656
  "recommendedActions": [
654
- "Before re-running Bash, check what changed since the last failure."
657
+ "Before re-running Bash, check what changed since the last failure.",
658
+ "Mitigate this remembered risk before proceeding."
655
659
  ],
656
- "summary": "Caution: 1 memory signal, 1 medium severity found before acting.",
660
+ "summary": "Caution: 2 memory signals, 2 medium severity found before acting.",
657
661
  "recallErrors": [],
658
662
  "leakedSecrets": [],
659
663
  "hasEvidenceForDecision": true,
@@ -669,15 +673,17 @@
669
673
  "decisionCorrect": true,
670
674
  "riskScore": 0.55,
671
675
  "passed": true,
672
- "latencyMs": 3.352,
673
- "evidenceCount": 1,
676
+ "latencyMs": 4.253,
677
+ "evidenceCount": 2,
674
678
  "evidenceIds": [
675
- "failure:Bash:2026-07-10T15:00:38.199Z"
679
+ "failure:Bash:2026-07-13T14:48:45.755Z",
680
+ "01KXDZ7CXVVQDPT9JT83RBRG57"
676
681
  ],
677
682
  "recommendedActions": [
678
- "Before re-running Bash, check what changed since the last failure."
683
+ "Before re-running Bash, check what changed since the last failure.",
684
+ "Mitigate this remembered risk before proceeding."
679
685
  ],
680
- "summary": "Caution: 1 memory signal, 1 medium severity found before acting.",
686
+ "summary": "Caution: 2 memory signals, 2 medium severity found before acting.",
681
687
  "recallErrors": [],
682
688
  "leakedSecrets": [],
683
689
  "hasEvidenceForDecision": true,
@@ -693,15 +699,15 @@
693
699
  "decisionCorrect": true,
694
700
  "riskScore": 0.2,
695
701
  "passed": true,
696
- "latencyMs": 3.713,
702
+ "latencyMs": 11.953,
697
703
  "evidenceCount": 2,
698
704
  "evidenceIds": [
699
- "01KX68PZQV0YJ1HSP588W4MNEF",
700
- "failure:Bash:2026-07-10T15:00:38.264Z"
705
+ "01KXDZ7D00R86YJ387KXYCAGQY",
706
+ "01KXDZ7CZYQ6ZA41GTZECZYCQ8"
701
707
  ],
702
708
  "recommendedActions": [
703
709
  "This exact action has succeeded since its last failure; proceed with normal validation.",
704
- "Before re-running Bash, check what changed since the last failure."
710
+ "Mitigate this remembered risk before proceeding."
705
711
  ],
706
712
  "summary": "Allowed: this exact Bash action has succeeded since the prior failure. Caution: 1 memory signal, 1 medium severity found before acting.",
707
713
  "recallErrors": [],
@@ -719,15 +725,15 @@
719
725
  "decisionCorrect": true,
720
726
  "riskScore": 0.85,
721
727
  "passed": true,
722
- "latencyMs": 3.501,
728
+ "latencyMs": 3.697,
723
729
  "evidenceCount": 1,
724
730
  "evidenceIds": [
725
731
  "recall:episodic:recall.vector_counts"
726
732
  ],
727
733
  "recommendedActions": [
728
734
  "Do not proceed until the high-severity memory warning is addressed.",
729
- "Run npx audrey status and npx audrey reembed before depending on memory.",
730
- "Run npx audrey status and repair the degraded recall path before relying on Guard."
735
+ "Run audrey status and audrey reembed before depending on memory.",
736
+ "Run audrey status and repair the degraded recall path before relying on Guard."
731
737
  ],
732
738
  "summary": "Blocked: 2 memory signals, 2 high severity found before acting.",
733
739
  "recallErrors": [],
@@ -745,15 +751,15 @@
745
751
  "decisionCorrect": true,
746
752
  "riskScore": 0.85,
747
753
  "passed": true,
748
- "latencyMs": 2.331,
754
+ "latencyMs": 2.763,
749
755
  "evidenceCount": 2,
750
756
  "evidenceIds": [
751
757
  "recall:fts:recall.fts_lookup",
752
- "01KX68PZW4P9SM0YA35MZ5DCH2"
758
+ "01KXDZ7D4F9H6PC43S0C6FCN6W"
753
759
  ],
754
760
  "recommendedActions": [
755
761
  "Do not proceed until the high-severity memory warning is addressed.",
756
- "Run npx audrey status and repair the degraded recall path before relying on Guard.",
762
+ "Run audrey status and repair the degraded recall path before relying on Guard.",
757
763
  "Apply this must-follow rule before acting."
758
764
  ],
759
765
  "summary": "Blocked: 2 memory signals, 2 high severity found before acting.",
@@ -772,17 +778,19 @@
772
778
  "decisionCorrect": true,
773
779
  "riskScore": 0.9,
774
780
  "passed": true,
775
- "latencyMs": 3.325,
776
- "evidenceCount": 2,
781
+ "latencyMs": 4.107,
782
+ "evidenceCount": 3,
777
783
  "evidenceIds": [
778
- "01KX68PZY4C1NNRZ4AMPJ2HRCS",
779
- "failure:Bash:2026-07-10T15:00:38.468Z"
784
+ "01KXDZ7D6E5JHP9RXZV2XE758W",
785
+ "failure:Bash:2026-07-13T14:48:46.030Z",
786
+ "01KXDZ7D6FR4DVYEP3RE8N47M1"
780
787
  ],
781
788
  "recommendedActions": [
782
789
  "Do not repeat the exact failed action until the prior error is understood or the command is changed.",
783
- "Before re-running Bash, check what changed since the last failure."
790
+ "Before re-running Bash, check what changed since the last failure.",
791
+ "Mitigate this remembered risk before proceeding."
784
792
  ],
785
- "summary": "Blocked: this exact Bash action failed before. Caution: 1 memory signal, 1 medium severity found before acting.",
793
+ "summary": "Blocked: this exact Bash action failed before. Caution: 2 memory signals, 2 medium severity found before acting.",
786
794
  "recallErrors": [],
787
795
  "leakedSecrets": [],
788
796
  "hasEvidenceForDecision": true,
@@ -798,11 +806,11 @@
798
806
  "decisionCorrect": true,
799
807
  "riskScore": 0.85,
800
808
  "passed": true,
801
- "latencyMs": 2.929,
809
+ "latencyMs": 3.483,
802
810
  "evidenceCount": 2,
803
811
  "evidenceIds": [
804
- "01KX68Q005CD1A3WSXT60H59YR",
805
- "01KX68Q0073S22ZE9SP0ZXH3WB"
812
+ "01KXDZ7D8HCCD19F4MX5T8TQW6",
813
+ "01KXDZ7D8GSJYEYEA0PQA2CFEQ"
806
814
  ],
807
815
  "recommendedActions": [
808
816
  "Do not proceed until the high-severity memory warning is addressed.",
@@ -824,22 +832,22 @@
824
832
  "decisionCorrect": true,
825
833
  "riskScore": 0.85,
826
834
  "passed": true,
827
- "latencyMs": 20.658,
835
+ "latencyMs": 21.196,
828
836
  "evidenceCount": 13,
829
837
  "evidenceIds": [
830
- "01KX68Q0KMT3CNSVDB8361Y1P9",
831
- "01KX68Q0J3WJC7S9PKS6R1P66X",
832
- "01KX68Q0HKXX65SX762BNK9EW3",
833
- "01KX68Q0GTQFFX4X1SM68P5X2T",
834
- "01KX68Q0G3M9M4ZZAG7H47DD6Y",
835
- "01KX68Q0DJMHF9A3DVEQZMGN5E",
836
- "01KX68Q0BTQKBASY41SRJWGA4W",
837
- "01KX68Q0BC3SDK0J1QKVV1VYHN",
838
- "01KX68Q0A7SAR678VQG7PEGBY5",
839
- "01KX68Q0936RX5YKWWW3782KJ9",
840
- "01KX68Q06M2AHVDX2XDXF1GTWT",
841
- "01KX68Q063XP49JPKJZBEXDDZ7",
842
- "01KX68Q05NZ6CY360M6MYJ3YKV"
838
+ "01KXDZ7DVW999JP92H6F6TV4Q7",
839
+ "01KXDZ7DTEMQ67D5HHJD5A4Q4P",
840
+ "01KXDZ7DRHKN6TRS0FD99EXHRR",
841
+ "01KXDZ7DQDBDNYA0CB6026EN7J",
842
+ "01KXDZ7DP9VB7FAQMF83XT32QB",
843
+ "01KXDZ7DMTNF8ZWMG97CX7M3YD",
844
+ "01KXDZ7DM0CM23Y6ETWAB3M3KM",
845
+ "01KXDZ7DJGG0YRG6SYVP00ZC21",
846
+ "01KXDZ7DJCQJV1NCSEF680GCCQ",
847
+ "01KXDZ7DFMRHWTY6X42JC6C1FG",
848
+ "01KXDZ7DEDR8KPS7T6E8QZM1BY",
849
+ "01KXDZ7DDP9V0DGWT3YJADCSM3",
850
+ "01KXDZ7DBFNRKSF2KH1FBXGDJV"
843
851
  ],
844
852
  "recommendedActions": [
845
853
  "Do not proceed until the high-severity memory warning is addressed.",
@@ -869,17 +877,19 @@
869
877
  "decisionCorrect": true,
870
878
  "riskScore": 0.9,
871
879
  "passed": true,
872
- "latencyMs": 9.146,
873
- "evidenceCount": 2,
880
+ "latencyMs": 11.267,
881
+ "evidenceCount": 3,
874
882
  "evidenceIds": [
875
- "01KX68PZF1G0JED770QTGDF0AB",
876
- "failure:Bash:2026-07-10T15:00:37.986Z"
883
+ "01KXDZ7CQ6VDPVF6JH0HH1KM2A",
884
+ "failure:Bash:2026-07-13T14:48:45.542Z",
885
+ "01KXDZ7CQ73T2NW3XX0033D6F7"
877
886
  ],
878
887
  "recommendedActions": [
879
888
  "Do not repeat the exact failed action until the prior error is understood or the command is changed.",
880
- "Before re-running Bash, check what changed since the last failure."
889
+ "Before re-running Bash, check what changed since the last failure.",
890
+ "Mitigate this remembered risk before proceeding."
881
891
  ],
882
- "summary": "Blocked: this exact Bash action failed before. Caution: 1 memory signal, 1 medium severity found before acting.",
892
+ "summary": "Blocked: this exact Bash action failed before. Caution: 2 memory signals, 2 medium severity found before acting.",
883
893
  "recallErrors": [],
884
894
  "leakedSecrets": [],
885
895
  "hasEvidenceForDecision": true,
@@ -915,10 +925,10 @@
915
925
  "decisionCorrect": false,
916
926
  "riskScore": 0.55,
917
927
  "passed": false,
918
- "latencyMs": 0.236,
928
+ "latencyMs": 0.225,
919
929
  "evidenceCount": 1,
920
930
  "evidenceIds": [
921
- "01KX68PZG7PE9ERBP4MKGN9QPY"
931
+ "01KXDZ7CRE1TZGVGS5FYAB55J5"
922
932
  ],
923
933
  "recommendedActions": [
924
934
  "Check the recent failed event before repeating a similar action."
@@ -942,7 +952,7 @@
942
952
  "latencyMs": 0.936,
943
953
  "evidenceCount": 1,
944
954
  "evidenceIds": [
945
- "01KX68PZGNN3Y78AYB7CSR8RKF"
955
+ "01KXDZ7CRVXW50THCG4JGFFGVZ"
946
956
  ],
947
957
  "recommendedActions": [
948
958
  "Treat retrieved memory as advisory context."
@@ -963,7 +973,7 @@
963
973
  "decisionCorrect": false,
964
974
  "riskScore": 0,
965
975
  "passed": false,
966
- "latencyMs": 0.472,
976
+ "latencyMs": 0.462,
967
977
  "evidenceCount": 0,
968
978
  "evidenceIds": [],
969
979
  "recommendedActions": [],
@@ -990,10 +1000,10 @@
990
1000
  "decisionCorrect": true,
991
1001
  "riskScore": 0.85,
992
1002
  "passed": true,
993
- "latencyMs": 2.944,
1003
+ "latencyMs": 3.735,
994
1004
  "evidenceCount": 1,
995
1005
  "evidenceIds": [
996
- "01KX68PZHF6HCQV61BYWZZS1GJ"
1006
+ "01KXDZ7CSN0BGBJM8JA5SMHZWT"
997
1007
  ],
998
1008
  "recommendedActions": [
999
1009
  "Do not proceed until the high-severity memory warning is addressed.",
@@ -1015,7 +1025,7 @@
1015
1025
  "decisionCorrect": false,
1016
1026
  "riskScore": 0,
1017
1027
  "passed": false,
1018
- "latencyMs": 0.006,
1028
+ "latencyMs": 0.007,
1019
1029
  "evidenceCount": 0,
1020
1030
  "evidenceIds": [],
1021
1031
  "recommendedActions": [],
@@ -1035,10 +1045,10 @@
1035
1045
  "decisionCorrect": true,
1036
1046
  "riskScore": 0.85,
1037
1047
  "passed": true,
1038
- "latencyMs": 0.303,
1048
+ "latencyMs": 0.34,
1039
1049
  "evidenceCount": 1,
1040
1050
  "evidenceIds": [
1041
- "01KX68PZJADTWDP4M7VNZXN390"
1051
+ "01KXDZ7CTHTZ78ND87V3MAJ2NG"
1042
1052
  ],
1043
1053
  "recommendedActions": [
1044
1054
  "Review retrieved memory before acting."
@@ -1059,10 +1069,10 @@
1059
1069
  "decisionCorrect": true,
1060
1070
  "riskScore": 0.85,
1061
1071
  "passed": true,
1062
- "latencyMs": 0.639,
1072
+ "latencyMs": 0.621,
1063
1073
  "evidenceCount": 1,
1064
1074
  "evidenceIds": [
1065
- "01KX68PZJQFW5DR8TJ9JCHE5SH"
1075
+ "01KXDZ7CTXJT1PW9SH4VBT0336"
1066
1076
  ],
1067
1077
  "recommendedActions": [
1068
1078
  "Review retrieved memory before acting."
@@ -1083,7 +1093,7 @@
1083
1093
  "decisionCorrect": false,
1084
1094
  "riskScore": 0,
1085
1095
  "passed": false,
1086
- "latencyMs": 0.474,
1096
+ "latencyMs": 0.45,
1087
1097
  "evidenceCount": 0,
1088
1098
  "evidenceIds": [],
1089
1099
  "recommendedActions": [],
@@ -1110,15 +1120,17 @@
1110
1120
  "decisionCorrect": true,
1111
1121
  "riskScore": 0.55,
1112
1122
  "passed": true,
1113
- "latencyMs": 5.572,
1114
- "evidenceCount": 1,
1123
+ "latencyMs": 4.478,
1124
+ "evidenceCount": 2,
1115
1125
  "evidenceIds": [
1116
- "failure:Bash:2026-07-10T15:00:38.129Z"
1126
+ "failure:Bash:2026-07-13T14:48:45.688Z",
1127
+ "01KXDZ7CVSXNTMNMP89QBMCXYA"
1117
1128
  ],
1118
1129
  "recommendedActions": [
1119
- "Before re-running Bash, check what changed since the last failure."
1130
+ "Before re-running Bash, check what changed since the last failure.",
1131
+ "Mitigate this remembered risk before proceeding."
1120
1132
  ],
1121
- "summary": "Caution: 1 memory signal, 1 medium severity found before acting.",
1133
+ "summary": "Caution: 2 memory signals, 2 medium severity found before acting.",
1122
1134
  "recallErrors": [],
1123
1135
  "leakedSecrets": [],
1124
1136
  "hasEvidenceForDecision": true,
@@ -1154,10 +1166,10 @@
1154
1166
  "decisionCorrect": true,
1155
1167
  "riskScore": 0.55,
1156
1168
  "passed": true,
1157
- "latencyMs": 0.078,
1169
+ "latencyMs": 0.063,
1158
1170
  "evidenceCount": 1,
1159
1171
  "evidenceIds": [
1160
- "01KX68PZMHP790HNEAG7E0ENKR"
1172
+ "01KXDZ7CWPAR45TR9BEVEERQT1"
1161
1173
  ],
1162
1174
  "recommendedActions": [
1163
1175
  "Check the recent failed event before repeating a similar action."
@@ -1178,10 +1190,10 @@
1178
1190
  "decisionCorrect": true,
1179
1191
  "riskScore": 0.35,
1180
1192
  "passed": true,
1181
- "latencyMs": 0.446,
1193
+ "latencyMs": 0.464,
1182
1194
  "evidenceCount": 1,
1183
1195
  "evidenceIds": [
1184
- "01KX68PZMXNKFFGWZ02ENQRG0C"
1196
+ "01KXDZ7CX2VQB8W0AK1XY6N7W5"
1185
1197
  ],
1186
1198
  "recommendedActions": [
1187
1199
  "Treat retrieved memory as advisory context."
@@ -1202,7 +1214,7 @@
1202
1214
  "decisionCorrect": false,
1203
1215
  "riskScore": 0,
1204
1216
  "passed": false,
1205
- "latencyMs": 0.42,
1217
+ "latencyMs": 0.458,
1206
1218
  "evidenceCount": 0,
1207
1219
  "evidenceIds": [],
1208
1220
  "recommendedActions": [],
@@ -1229,15 +1241,17 @@
1229
1241
  "decisionCorrect": true,
1230
1242
  "riskScore": 0.55,
1231
1243
  "passed": true,
1232
- "latencyMs": 3.352,
1233
- "evidenceCount": 1,
1244
+ "latencyMs": 4.253,
1245
+ "evidenceCount": 2,
1234
1246
  "evidenceIds": [
1235
- "failure:Bash:2026-07-10T15:00:38.199Z"
1247
+ "failure:Bash:2026-07-13T14:48:45.755Z",
1248
+ "01KXDZ7CXVVQDPT9JT83RBRG57"
1236
1249
  ],
1237
1250
  "recommendedActions": [
1238
- "Before re-running Bash, check what changed since the last failure."
1251
+ "Before re-running Bash, check what changed since the last failure.",
1252
+ "Mitigate this remembered risk before proceeding."
1239
1253
  ],
1240
- "summary": "Caution: 1 memory signal, 1 medium severity found before acting.",
1254
+ "summary": "Caution: 2 memory signals, 2 medium severity found before acting.",
1241
1255
  "recallErrors": [],
1242
1256
  "leakedSecrets": [],
1243
1257
  "hasEvidenceForDecision": true,
@@ -1253,7 +1267,7 @@
1253
1267
  "decisionCorrect": false,
1254
1268
  "riskScore": 0,
1255
1269
  "passed": false,
1256
- "latencyMs": 0.015,
1270
+ "latencyMs": 0.008,
1257
1271
  "evidenceCount": 0,
1258
1272
  "evidenceIds": [],
1259
1273
  "recommendedActions": [],
@@ -1273,10 +1287,10 @@
1273
1287
  "decisionCorrect": true,
1274
1288
  "riskScore": 0.55,
1275
1289
  "passed": true,
1276
- "latencyMs": 0.062,
1290
+ "latencyMs": 0.069,
1277
1291
  "evidenceCount": 1,
1278
1292
  "evidenceIds": [
1279
- "01KX68PZPK06NCBYPWW90767AS"
1293
+ "01KXDZ7CYQ5BYW6GC0PFAPF5NP"
1280
1294
  ],
1281
1295
  "recommendedActions": [
1282
1296
  "Check the recent failed event before repeating a similar action."
@@ -1297,10 +1311,10 @@
1297
1311
  "decisionCorrect": true,
1298
1312
  "riskScore": 0.35,
1299
1313
  "passed": true,
1300
- "latencyMs": 0.449,
1314
+ "latencyMs": 0.465,
1301
1315
  "evidenceCount": 1,
1302
1316
  "evidenceIds": [
1303
- "01KX68PZPZPY9E2NHCQ8KNDDRB"
1317
+ "01KXDZ7CZ4XMHVMV4MQ65G76KK"
1304
1318
  ],
1305
1319
  "recommendedActions": [
1306
1320
  "Treat retrieved memory as advisory context."
@@ -1321,7 +1335,7 @@
1321
1335
  "decisionCorrect": false,
1322
1336
  "riskScore": 0,
1323
1337
  "passed": false,
1324
- "latencyMs": 0.414,
1338
+ "latencyMs": 0.423,
1325
1339
  "evidenceCount": 0,
1326
1340
  "evidenceIds": [],
1327
1341
  "recommendedActions": [],
@@ -1348,15 +1362,15 @@
1348
1362
  "decisionCorrect": true,
1349
1363
  "riskScore": 0.2,
1350
1364
  "passed": true,
1351
- "latencyMs": 3.713,
1365
+ "latencyMs": 11.953,
1352
1366
  "evidenceCount": 2,
1353
1367
  "evidenceIds": [
1354
- "01KX68PZQV0YJ1HSP588W4MNEF",
1355
- "failure:Bash:2026-07-10T15:00:38.264Z"
1368
+ "01KXDZ7D00R86YJ387KXYCAGQY",
1369
+ "01KXDZ7CZYQ6ZA41GTZECZYCQ8"
1356
1370
  ],
1357
1371
  "recommendedActions": [
1358
1372
  "This exact action has succeeded since its last failure; proceed with normal validation.",
1359
- "Before re-running Bash, check what changed since the last failure."
1373
+ "Mitigate this remembered risk before proceeding."
1360
1374
  ],
1361
1375
  "summary": "Allowed: this exact Bash action has succeeded since the prior failure. Caution: 1 memory signal, 1 medium severity found before acting.",
1362
1376
  "recallErrors": [],
@@ -1374,7 +1388,7 @@
1374
1388
  "decisionCorrect": true,
1375
1389
  "riskScore": 0,
1376
1390
  "passed": true,
1377
- "latencyMs": 0.008,
1391
+ "latencyMs": 0.005,
1378
1392
  "evidenceCount": 0,
1379
1393
  "evidenceIds": [],
1380
1394
  "recommendedActions": [],
@@ -1394,10 +1408,10 @@
1394
1408
  "decisionCorrect": false,
1395
1409
  "riskScore": 0.55,
1396
1410
  "passed": false,
1397
- "latencyMs": 0.079,
1411
+ "latencyMs": 0.074,
1398
1412
  "evidenceCount": 1,
1399
1413
  "evidenceIds": [
1400
- "01KX68PZRQTVN9GPVQCFTZ3GVF"
1414
+ "01KXDZ7D15SRJYZ8GRVY6YET3F"
1401
1415
  ],
1402
1416
  "recommendedActions": [
1403
1417
  "Check the recent failed event before repeating a similar action."
@@ -1418,10 +1432,10 @@
1418
1432
  "decisionCorrect": false,
1419
1433
  "riskScore": 0.35,
1420
1434
  "passed": false,
1421
- "latencyMs": 0.436,
1435
+ "latencyMs": 0.595,
1422
1436
  "evidenceCount": 1,
1423
1437
  "evidenceIds": [
1424
- "01KX68PZS5761176FZDQVD6B7S"
1438
+ "01KXDZ7D1J2WAKA39WF2ACKRDD"
1425
1439
  ],
1426
1440
  "recommendedActions": [
1427
1441
  "Treat retrieved memory as advisory context."
@@ -1442,7 +1456,7 @@
1442
1456
  "decisionCorrect": true,
1443
1457
  "riskScore": 0,
1444
1458
  "passed": true,
1445
- "latencyMs": 0.408,
1459
+ "latencyMs": 0.41,
1446
1460
  "evidenceCount": 0,
1447
1461
  "evidenceIds": [],
1448
1462
  "recommendedActions": [],
@@ -1469,15 +1483,15 @@
1469
1483
  "decisionCorrect": true,
1470
1484
  "riskScore": 0.85,
1471
1485
  "passed": true,
1472
- "latencyMs": 3.501,
1486
+ "latencyMs": 3.697,
1473
1487
  "evidenceCount": 1,
1474
1488
  "evidenceIds": [
1475
1489
  "recall:episodic:recall.vector_counts"
1476
1490
  ],
1477
1491
  "recommendedActions": [
1478
1492
  "Do not proceed until the high-severity memory warning is addressed.",
1479
- "Run npx audrey status and npx audrey reembed before depending on memory.",
1480
- "Run npx audrey status and repair the degraded recall path before relying on Guard."
1493
+ "Run audrey status and audrey reembed before depending on memory.",
1494
+ "Run audrey status and repair the degraded recall path before relying on Guard."
1481
1495
  ],
1482
1496
  "summary": "Blocked: 2 memory signals, 2 high severity found before acting.",
1483
1497
  "recallErrors": [],
@@ -1495,7 +1509,7 @@
1495
1509
  "decisionCorrect": false,
1496
1510
  "riskScore": 0,
1497
1511
  "passed": false,
1498
- "latencyMs": 0.005,
1512
+ "latencyMs": 0.006,
1499
1513
  "evidenceCount": 0,
1500
1514
  "evidenceIds": [],
1501
1515
  "recommendedActions": [],
@@ -1515,10 +1529,10 @@
1515
1529
  "decisionCorrect": true,
1516
1530
  "riskScore": 0.85,
1517
1531
  "passed": true,
1518
- "latencyMs": 0.14,
1532
+ "latencyMs": 0.148,
1519
1533
  "evidenceCount": 1,
1520
1534
  "evidenceIds": [
1521
- "01KX68PZTYTW8DWT8920EV3JCG"
1535
+ "01KXDZ7D3BDX746H56BJNSE6Z6"
1522
1536
  ],
1523
1537
  "recommendedActions": [
1524
1538
  "Review retrieved memory before acting."
@@ -1539,7 +1553,7 @@
1539
1553
  "decisionCorrect": false,
1540
1554
  "riskScore": 0.55,
1541
1555
  "passed": false,
1542
- "latencyMs": 0.311,
1556
+ "latencyMs": 0.292,
1543
1557
  "evidenceCount": 0,
1544
1558
  "evidenceIds": [],
1545
1559
  "recommendedActions": [
@@ -1567,7 +1581,7 @@
1567
1581
  "decisionCorrect": false,
1568
1582
  "riskScore": 0,
1569
1583
  "passed": false,
1570
- "latencyMs": 0.367,
1584
+ "latencyMs": 0.388,
1571
1585
  "evidenceCount": 0,
1572
1586
  "evidenceIds": [],
1573
1587
  "recommendedActions": [],
@@ -1594,15 +1608,15 @@
1594
1608
  "decisionCorrect": true,
1595
1609
  "riskScore": 0.85,
1596
1610
  "passed": true,
1597
- "latencyMs": 2.331,
1611
+ "latencyMs": 2.763,
1598
1612
  "evidenceCount": 2,
1599
1613
  "evidenceIds": [
1600
1614
  "recall:fts:recall.fts_lookup",
1601
- "01KX68PZW4P9SM0YA35MZ5DCH2"
1615
+ "01KXDZ7D4F9H6PC43S0C6FCN6W"
1602
1616
  ],
1603
1617
  "recommendedActions": [
1604
1618
  "Do not proceed until the high-severity memory warning is addressed.",
1605
- "Run npx audrey status and repair the degraded recall path before relying on Guard.",
1619
+ "Run audrey status and repair the degraded recall path before relying on Guard.",
1606
1620
  "Apply this must-follow rule before acting."
1607
1621
  ],
1608
1622
  "summary": "Blocked: 2 memory signals, 2 high severity found before acting.",
@@ -1641,10 +1655,10 @@
1641
1655
  "decisionCorrect": false,
1642
1656
  "riskScore": 0.35,
1643
1657
  "passed": false,
1644
- "latencyMs": 0.104,
1658
+ "latencyMs": 0.107,
1645
1659
  "evidenceCount": 1,
1646
1660
  "evidenceIds": [
1647
- "01KX68PZWYQ9JQXFYW0BF5VXYR"
1661
+ "01KXDZ7D5A65W8JASMQHJDB9ZH"
1648
1662
  ],
1649
1663
  "recommendedActions": [
1650
1664
  "Treat retrieved memory as advisory context."
@@ -1665,10 +1679,10 @@
1665
1679
  "decisionCorrect": false,
1666
1680
  "riskScore": 0.35,
1667
1681
  "passed": false,
1668
- "latencyMs": 0.343,
1682
+ "latencyMs": 0.37,
1669
1683
  "evidenceCount": 1,
1670
1684
  "evidenceIds": [
1671
- "01KX68PZXAEV3750MQ9YR6QWGM"
1685
+ "01KXDZ7D5NRMHVW0E774P15SSF"
1672
1686
  ],
1673
1687
  "recommendedActions": [
1674
1688
  "Treat retrieved memory as advisory context."
@@ -1689,7 +1703,7 @@
1689
1703
  "decisionCorrect": false,
1690
1704
  "riskScore": 0.55,
1691
1705
  "passed": false,
1692
- "latencyMs": 0.128,
1706
+ "latencyMs": 0.135,
1693
1707
  "evidenceCount": 0,
1694
1708
  "evidenceIds": [],
1695
1709
  "recommendedActions": [
@@ -1724,17 +1738,19 @@
1724
1738
  "decisionCorrect": true,
1725
1739
  "riskScore": 0.9,
1726
1740
  "passed": true,
1727
- "latencyMs": 3.325,
1728
- "evidenceCount": 2,
1741
+ "latencyMs": 4.107,
1742
+ "evidenceCount": 3,
1729
1743
  "evidenceIds": [
1730
- "01KX68PZY4C1NNRZ4AMPJ2HRCS",
1731
- "failure:Bash:2026-07-10T15:00:38.468Z"
1744
+ "01KXDZ7D6E5JHP9RXZV2XE758W",
1745
+ "failure:Bash:2026-07-13T14:48:46.030Z",
1746
+ "01KXDZ7D6FR4DVYEP3RE8N47M1"
1732
1747
  ],
1733
1748
  "recommendedActions": [
1734
1749
  "Do not repeat the exact failed action until the prior error is understood or the command is changed.",
1735
- "Before re-running Bash, check what changed since the last failure."
1750
+ "Before re-running Bash, check what changed since the last failure.",
1751
+ "Mitigate this remembered risk before proceeding."
1736
1752
  ],
1737
- "summary": "Blocked: this exact Bash action failed before. Caution: 1 memory signal, 1 medium severity found before acting.",
1753
+ "summary": "Blocked: this exact Bash action failed before. Caution: 2 memory signals, 2 medium severity found before acting.",
1738
1754
  "recallErrors": [],
1739
1755
  "leakedSecrets": [],
1740
1756
  "hasEvidenceForDecision": true,
@@ -1770,10 +1786,10 @@
1770
1786
  "decisionCorrect": false,
1771
1787
  "riskScore": 0.55,
1772
1788
  "passed": false,
1773
- "latencyMs": 0.061,
1789
+ "latencyMs": 0.062,
1774
1790
  "evidenceCount": 1,
1775
1791
  "evidenceIds": [
1776
- "01KX68PZZ01F99MPCJ0NNPD8SA"
1792
+ "01KXDZ7D7B6KSNRGXDM48CCF2B"
1777
1793
  ],
1778
1794
  "recommendedActions": [
1779
1795
  "Check the recent failed event before repeating a similar action."
@@ -1794,10 +1810,10 @@
1794
1810
  "decisionCorrect": false,
1795
1811
  "riskScore": 0.35,
1796
1812
  "passed": false,
1797
- "latencyMs": 0.349,
1813
+ "latencyMs": 0.376,
1798
1814
  "evidenceCount": 1,
1799
1815
  "evidenceIds": [
1800
- "01KX68PZZC81TCG6X2PHYDNH8Y"
1816
+ "01KXDZ7D7RGKJ0SM9QT3H9WTKG"
1801
1817
  ],
1802
1818
  "recommendedActions": [
1803
1819
  "Treat retrieved memory as advisory context."
@@ -1818,7 +1834,7 @@
1818
1834
  "decisionCorrect": false,
1819
1835
  "riskScore": 0,
1820
1836
  "passed": false,
1821
- "latencyMs": 0.345,
1837
+ "latencyMs": 0.361,
1822
1838
  "evidenceCount": 0,
1823
1839
  "evidenceIds": [],
1824
1840
  "recommendedActions": [],
@@ -1845,11 +1861,11 @@
1845
1861
  "decisionCorrect": true,
1846
1862
  "riskScore": 0.85,
1847
1863
  "passed": true,
1848
- "latencyMs": 2.929,
1864
+ "latencyMs": 3.483,
1849
1865
  "evidenceCount": 2,
1850
1866
  "evidenceIds": [
1851
- "01KX68Q005CD1A3WSXT60H59YR",
1852
- "01KX68Q0073S22ZE9SP0ZXH3WB"
1867
+ "01KXDZ7D8HCCD19F4MX5T8TQW6",
1868
+ "01KXDZ7D8GSJYEYEA0PQA2CFEQ"
1853
1869
  ],
1854
1870
  "recommendedActions": [
1855
1871
  "Do not proceed until the high-severity memory warning is addressed.",
@@ -1871,7 +1887,7 @@
1871
1887
  "decisionCorrect": false,
1872
1888
  "riskScore": 0,
1873
1889
  "passed": false,
1874
- "latencyMs": 0.004,
1890
+ "latencyMs": 0.005,
1875
1891
  "evidenceCount": 0,
1876
1892
  "evidenceIds": [],
1877
1893
  "recommendedActions": [],
@@ -1891,11 +1907,11 @@
1891
1907
  "decisionCorrect": true,
1892
1908
  "riskScore": 0.85,
1893
1909
  "passed": true,
1894
- "latencyMs": 0.112,
1910
+ "latencyMs": 0.11,
1895
1911
  "evidenceCount": 2,
1896
1912
  "evidenceIds": [
1897
- "01KX68Q012APPJZRT9713GF2RG",
1898
- "01KX68Q011FPXTBDCB09N8DESP"
1913
+ "01KXDZ7D9D2T5HC5363DW62Q5Y",
1914
+ "01KXDZ7D9CB9AN9NTPT76ER3FS"
1899
1915
  ],
1900
1916
  "recommendedActions": [
1901
1917
  "Review retrieved memory before acting."
@@ -1916,11 +1932,11 @@
1916
1932
  "decisionCorrect": true,
1917
1933
  "riskScore": 0.85,
1918
1934
  "passed": true,
1919
- "latencyMs": 0.417,
1935
+ "latencyMs": 0.436,
1920
1936
  "evidenceCount": 2,
1921
1937
  "evidenceIds": [
1922
- "01KX68Q01FSCQM4T04XN25MK09",
1923
- "01KX68Q01DGS5D8FWBPXSTDM5N"
1938
+ "01KXDZ7D9TWCABY6SX2PRFF3GC",
1939
+ "01KXDZ7D9RRHJGZX0SGG6KW02G"
1924
1940
  ],
1925
1941
  "recommendedActions": [
1926
1942
  "Review retrieved memory before acting."
@@ -1941,7 +1957,7 @@
1941
1957
  "decisionCorrect": false,
1942
1958
  "riskScore": 0,
1943
1959
  "passed": false,
1944
- "latencyMs": 0.358,
1960
+ "latencyMs": 0.364,
1945
1961
  "evidenceCount": 0,
1946
1962
  "evidenceIds": [],
1947
1963
  "recommendedActions": [],
@@ -1968,22 +1984,22 @@
1968
1984
  "decisionCorrect": true,
1969
1985
  "riskScore": 0.85,
1970
1986
  "passed": true,
1971
- "latencyMs": 20.658,
1987
+ "latencyMs": 21.196,
1972
1988
  "evidenceCount": 13,
1973
1989
  "evidenceIds": [
1974
- "01KX68Q0KMT3CNSVDB8361Y1P9",
1975
- "01KX68Q0J3WJC7S9PKS6R1P66X",
1976
- "01KX68Q0HKXX65SX762BNK9EW3",
1977
- "01KX68Q0GTQFFX4X1SM68P5X2T",
1978
- "01KX68Q0G3M9M4ZZAG7H47DD6Y",
1979
- "01KX68Q0DJMHF9A3DVEQZMGN5E",
1980
- "01KX68Q0BTQKBASY41SRJWGA4W",
1981
- "01KX68Q0BC3SDK0J1QKVV1VYHN",
1982
- "01KX68Q0A7SAR678VQG7PEGBY5",
1983
- "01KX68Q0936RX5YKWWW3782KJ9",
1984
- "01KX68Q06M2AHVDX2XDXF1GTWT",
1985
- "01KX68Q063XP49JPKJZBEXDDZ7",
1986
- "01KX68Q05NZ6CY360M6MYJ3YKV"
1990
+ "01KXDZ7DVW999JP92H6F6TV4Q7",
1991
+ "01KXDZ7DTEMQ67D5HHJD5A4Q4P",
1992
+ "01KXDZ7DRHKN6TRS0FD99EXHRR",
1993
+ "01KXDZ7DQDBDNYA0CB6026EN7J",
1994
+ "01KXDZ7DP9VB7FAQMF83XT32QB",
1995
+ "01KXDZ7DMTNF8ZWMG97CX7M3YD",
1996
+ "01KXDZ7DM0CM23Y6ETWAB3M3KM",
1997
+ "01KXDZ7DJGG0YRG6SYVP00ZC21",
1998
+ "01KXDZ7DJCQJV1NCSEF680GCCQ",
1999
+ "01KXDZ7DFMRHWTY6X42JC6C1FG",
2000
+ "01KXDZ7DEDR8KPS7T6E8QZM1BY",
2001
+ "01KXDZ7DDP9V0DGWT3YJADCSM3",
2002
+ "01KXDZ7DBFNRKSF2KH1FBXGDJV"
1987
2003
  ],
1988
2004
  "recommendedActions": [
1989
2005
  "Do not proceed until the high-severity memory warning is addressed.",
@@ -2026,10 +2042,10 @@
2026
2042
  "decisionCorrect": true,
2027
2043
  "riskScore": 0.85,
2028
2044
  "passed": true,
2029
- "latencyMs": 0.46,
2045
+ "latencyMs": 0.427,
2030
2046
  "evidenceCount": 1,
2031
2047
  "evidenceIds": [
2032
- "01KX68Q1VS6FYC6CSZKBNPN87N"
2048
+ "01KXDZ7F3CRSBGB5NWMYJGJG6C"
2033
2049
  ],
2034
2050
  "recommendedActions": [
2035
2051
  "Review retrieved memory before acting."
@@ -2050,14 +2066,14 @@
2050
2066
  "decisionCorrect": false,
2051
2067
  "riskScore": 0.35,
2052
2068
  "passed": false,
2053
- "latencyMs": 1.908,
2069
+ "latencyMs": 1.589,
2054
2070
  "evidenceCount": 5,
2055
2071
  "evidenceIds": [
2056
- "01KX68Q290ECWQNTHHY5GH7RGB",
2057
- "01KX68Q2AMG4D8CTM7XEAA1P59",
2058
- "01KX68Q28XSK36RBVB8MNTKBG5",
2059
- "01KX68Q222MGWCVEC1BM3R2MAK",
2060
- "01KX68Q2AVD1DFNMN6GGWG221S"
2072
+ "01KXDZ7FMS8GJ0E0VZDPTNMFWA",
2073
+ "01KXDZ7FEEQ4N77YQ6N07WW0GV",
2074
+ "01KXDZ7FDQABHDN1HWRRKPKJ7E",
2075
+ "01KXDZ7FC5XZMKTACKRSH258TR",
2076
+ "01KXDZ7FN452RQXNN7FBJ9EX1K"
2061
2077
  ],
2062
2078
  "recommendedActions": [
2063
2079
  "Treat retrieved memory as advisory context."
@@ -2078,7 +2094,7 @@
2078
2094
  "decisionCorrect": false,
2079
2095
  "riskScore": 0,
2080
2096
  "passed": false,
2081
- "latencyMs": 0.716,
2097
+ "latencyMs": 0.739,
2082
2098
  "evidenceCount": 0,
2083
2099
  "evidenceIds": [],
2084
2100
  "recommendedActions": [],
@@ -2093,7 +2109,7 @@
2093
2109
  }
2094
2110
  ],
2095
2111
  "artifactRedactionSweep": {
2096
- "checkedAt": "2026-07-10T15:00:41.807Z",
2112
+ "checkedAt": "2026-07-13T14:48:49.340Z",
2097
2113
  "filesChecked": [
2098
2114
  "benchmarks/output/guardbench-manifest.json",
2099
2115
  "benchmarks/output/guardbench-raw.json",