audrey 1.1.1 → 1.1.3
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 +27 -0
- package/README.md +3 -1
- package/benchmarks/output/adapter-self-test/guardbench-adapter-self-test.json +4 -4
- package/benchmarks/output/external/guardbench-external-dry-run.json +1 -1
- package/benchmarks/output/external/guardbench-external-evidence.json +1 -1
- package/benchmarks/output/guardbench-conformance-card.json +9 -9
- package/benchmarks/output/guardbench-raw.json +125 -117
- package/benchmarks/output/guardbench-summary.json +207 -191
- package/benchmarks/output/leaderboard/guardbench-leaderboard.json +5 -5
- package/benchmarks/output/leaderboard/guardbench-leaderboard.md +2 -2
- package/benchmarks/output/submission-bundle/guardbench-conformance-card.json +9 -9
- package/benchmarks/output/submission-bundle/guardbench-raw.json +125 -117
- package/benchmarks/output/submission-bundle/guardbench-summary.json +207 -191
- package/benchmarks/output/submission-bundle/submission-manifest.json +10 -10
- package/benchmarks/output/submission-bundle/validation-report.json +1 -1
- package/benchmarks/output/summary.json +70 -70
- package/dist/mcp-server/config.d.ts +1 -1
- package/dist/mcp-server/config.js +1 -1
- package/dist/mcp-server/index.d.ts.map +1 -1
- package/dist/mcp-server/index.js +68 -18
- package/dist/mcp-server/index.js.map +1 -1
- package/dist/src/audrey.d.ts +2 -0
- package/dist/src/audrey.d.ts.map +1 -1
- package/dist/src/audrey.js +2 -0
- package/dist/src/audrey.js.map +1 -1
- package/dist/src/autopilot.d.ts +3 -1
- package/dist/src/autopilot.d.ts.map +1 -1
- package/dist/src/autopilot.js +150 -37
- package/dist/src/autopilot.js.map +1 -1
- package/dist/src/capsule.d.ts +2 -0
- package/dist/src/capsule.d.ts.map +1 -1
- package/dist/src/capsule.js +43 -3
- package/dist/src/capsule.js.map +1 -1
- package/dist/src/controller.d.ts.map +1 -1
- package/dist/src/controller.js +7 -3
- package/dist/src/controller.js.map +1 -1
- package/dist/src/events.d.ts +23 -6
- package/dist/src/events.d.ts.map +1 -1
- package/dist/src/events.js +80 -13
- package/dist/src/events.js.map +1 -1
- package/dist/src/preflight.d.ts.map +1 -1
- package/dist/src/preflight.js +16 -3
- package/dist/src/preflight.js.map +1 -1
- package/dist/src/project.d.ts +17 -0
- package/dist/src/project.d.ts.map +1 -0
- package/dist/src/project.js +56 -0
- package/dist/src/project.js.map +1 -0
- package/dist/src/redact.d.ts.map +1 -1
- package/dist/src/redact.js +20 -0
- package/dist/src/redact.js.map +1 -1
- package/docs/paper/07-evaluation.md +5 -5
- package/docs/paper/audrey-paper-v1.md +5 -5
- package/docs/paper/evidence-ledger.md +1 -1
- package/docs/paper/output/arxiv/arxiv-manifest.json +4 -4
- package/docs/paper/output/arxiv/main.tex +5 -5
- package/docs/paper/output/arxiv-compile-report.json +3 -3
- package/docs/paper/output/submission-bundle/README.md +3 -1
- package/docs/paper/output/submission-bundle/benchmarks/output/adapter-self-test/guardbench-adapter-self-test.json +4 -4
- package/docs/paper/output/submission-bundle/benchmarks/output/external/guardbench-external-dry-run.json +1 -1
- package/docs/paper/output/submission-bundle/benchmarks/output/external/guardbench-external-evidence.json +1 -1
- package/docs/paper/output/submission-bundle/benchmarks/output/guardbench-conformance-card.json +9 -9
- package/docs/paper/output/submission-bundle/benchmarks/output/guardbench-raw.json +125 -117
- package/docs/paper/output/submission-bundle/benchmarks/output/guardbench-summary.json +207 -191
- package/docs/paper/output/submission-bundle/benchmarks/output/leaderboard/guardbench-leaderboard.json +5 -5
- package/docs/paper/output/submission-bundle/benchmarks/output/leaderboard/guardbench-leaderboard.md +2 -2
- package/docs/paper/output/submission-bundle/benchmarks/output/submission-bundle/submission-manifest.json +10 -10
- package/docs/paper/output/submission-bundle/benchmarks/output/submission-bundle/validation-report.json +1 -1
- package/docs/paper/output/submission-bundle/benchmarks/output/summary.json +62 -62
- package/docs/paper/output/submission-bundle/docs/paper/07-evaluation.md +5 -5
- package/docs/paper/output/submission-bundle/docs/paper/audrey-paper-v1.md +5 -5
- package/docs/paper/output/submission-bundle/docs/paper/evidence-ledger.md +1 -1
- package/docs/paper/output/submission-bundle/docs/paper/output/arxiv/arxiv-manifest.json +4 -4
- package/docs/paper/output/submission-bundle/docs/paper/output/arxiv/main.tex +5 -5
- package/docs/paper/output/submission-bundle/docs/paper/output/arxiv-compile-report.json +3 -3
- package/docs/paper/output/submission-bundle/package.json +2 -1
- package/docs/paper/output/submission-bundle/paper-submission-manifest.json +30 -30
- package/package.json +2 -1
|
@@ -1,17 +1,17 @@
|
|
|
1
1
|
{
|
|
2
2
|
"suite": "GuardBench comparative",
|
|
3
|
-
"generatedAt": "2026-07-
|
|
3
|
+
"generatedAt": "2026-07-13T15:07:31.178Z",
|
|
4
4
|
"manifestVersion": "0.2.0",
|
|
5
5
|
"provenance": {
|
|
6
|
-
"generatedAt": "2026-07-
|
|
7
|
-
"gitSha": "
|
|
6
|
+
"generatedAt": "2026-07-13T15:07:31.179Z",
|
|
7
|
+
"gitSha": "36842cfb0cad7ca5da864231a7e265229ba56fbd",
|
|
8
8
|
"gitDirty": false,
|
|
9
9
|
"node": "v24.18.0",
|
|
10
10
|
"v8": "13.6.233.17-node.50",
|
|
11
11
|
"platform": "linux",
|
|
12
12
|
"arch": "x64",
|
|
13
13
|
"osRelease": "6.17.0-1018-azure",
|
|
14
|
-
"cpuModel": "
|
|
14
|
+
"cpuModel": "AMD EPYC 7763 64-Core Processor",
|
|
15
15
|
"cpuCount": 4,
|
|
16
16
|
"totalMemoryGb": 15.61,
|
|
17
17
|
"embeddingProvider": "mock",
|
|
@@ -33,17 +33,19 @@
|
|
|
33
33
|
"decisionCorrect": true,
|
|
34
34
|
"riskScore": 0.9,
|
|
35
35
|
"passed": true,
|
|
36
|
-
"latencyMs":
|
|
37
|
-
"evidenceCount":
|
|
36
|
+
"latencyMs": 12.274,
|
|
37
|
+
"evidenceCount": 3,
|
|
38
38
|
"evidenceIds": [
|
|
39
|
-
"
|
|
40
|
-
"failure:Bash:2026-07-
|
|
39
|
+
"01KXE09KRM3F7XNRBG7KRBC2PP",
|
|
40
|
+
"failure:Bash:2026-07-13T15:07:26.869Z",
|
|
41
|
+
"01KXE09KRP13YT2CWDQSQDF6KB"
|
|
41
42
|
],
|
|
42
43
|
"recommendedActions": [
|
|
43
44
|
"Do not repeat the exact failed action until the prior error is understood or the command is changed.",
|
|
44
|
-
"Before re-running Bash, check what changed since the last failure."
|
|
45
|
+
"Before re-running Bash, check what changed since the last failure.",
|
|
46
|
+
"Mitigate this remembered risk before proceeding."
|
|
45
47
|
],
|
|
46
|
-
"summary": "Blocked: this exact Bash action failed before. Caution:
|
|
48
|
+
"summary": "Blocked: this exact Bash action failed before. Caution: 2 memory signals, 2 medium severity found before acting.",
|
|
47
49
|
"recallErrors": [],
|
|
48
50
|
"leakedSecrets": [],
|
|
49
51
|
"hasEvidenceForDecision": true,
|
|
@@ -59,7 +61,7 @@
|
|
|
59
61
|
"decisionCorrect": false,
|
|
60
62
|
"riskScore": 0,
|
|
61
63
|
"passed": false,
|
|
62
|
-
"latencyMs": 0.
|
|
64
|
+
"latencyMs": 0.077,
|
|
63
65
|
"evidenceCount": 0,
|
|
64
66
|
"evidenceIds": [],
|
|
65
67
|
"recommendedActions": [],
|
|
@@ -79,10 +81,10 @@
|
|
|
79
81
|
"decisionCorrect": false,
|
|
80
82
|
"riskScore": 0.55,
|
|
81
83
|
"passed": false,
|
|
82
|
-
"latencyMs": 0.
|
|
84
|
+
"latencyMs": 0.305,
|
|
83
85
|
"evidenceCount": 1,
|
|
84
86
|
"evidenceIds": [
|
|
85
|
-
"
|
|
87
|
+
"01KXE09KT14FCCQ17VECTYQ62H"
|
|
86
88
|
],
|
|
87
89
|
"recommendedActions": [
|
|
88
90
|
"Check the recent failed event before repeating a similar action."
|
|
@@ -103,10 +105,10 @@
|
|
|
103
105
|
"decisionCorrect": false,
|
|
104
106
|
"riskScore": 0.35,
|
|
105
107
|
"passed": false,
|
|
106
|
-
"latencyMs":
|
|
108
|
+
"latencyMs": 1.087,
|
|
107
109
|
"evidenceCount": 1,
|
|
108
110
|
"evidenceIds": [
|
|
109
|
-
"
|
|
111
|
+
"01KXE09KTHDVHXH9QY977Y3VBQ"
|
|
110
112
|
],
|
|
111
113
|
"recommendedActions": [
|
|
112
114
|
"Treat retrieved memory as advisory context."
|
|
@@ -127,7 +129,7 @@
|
|
|
127
129
|
"decisionCorrect": false,
|
|
128
130
|
"riskScore": 0,
|
|
129
131
|
"passed": false,
|
|
130
|
-
"latencyMs": 0.
|
|
132
|
+
"latencyMs": 0.615,
|
|
131
133
|
"evidenceCount": 0,
|
|
132
134
|
"evidenceIds": [],
|
|
133
135
|
"recommendedActions": [],
|
|
@@ -154,10 +156,10 @@
|
|
|
154
156
|
"decisionCorrect": true,
|
|
155
157
|
"riskScore": 0.85,
|
|
156
158
|
"passed": true,
|
|
157
|
-
"latencyMs":
|
|
159
|
+
"latencyMs": 3.943,
|
|
158
160
|
"evidenceCount": 1,
|
|
159
161
|
"evidenceIds": [
|
|
160
|
-
"
|
|
162
|
+
"01KXE09KVH81YSF0V7QA4ZDERQ"
|
|
161
163
|
],
|
|
162
164
|
"recommendedActions": [
|
|
163
165
|
"Do not proceed until the high-severity memory warning is addressed.",
|
|
@@ -179,7 +181,7 @@
|
|
|
179
181
|
"decisionCorrect": false,
|
|
180
182
|
"riskScore": 0,
|
|
181
183
|
"passed": false,
|
|
182
|
-
"latencyMs": 0.
|
|
184
|
+
"latencyMs": 0.006,
|
|
183
185
|
"evidenceCount": 0,
|
|
184
186
|
"evidenceIds": [],
|
|
185
187
|
"recommendedActions": [],
|
|
@@ -199,10 +201,10 @@
|
|
|
199
201
|
"decisionCorrect": true,
|
|
200
202
|
"riskScore": 0.85,
|
|
201
203
|
"passed": true,
|
|
202
|
-
"latencyMs": 0.
|
|
204
|
+
"latencyMs": 0.43,
|
|
203
205
|
"evidenceCount": 1,
|
|
204
206
|
"evidenceIds": [
|
|
205
|
-
"
|
|
207
|
+
"01KXE09KWN7GCDW06T5K69GYMG"
|
|
206
208
|
],
|
|
207
209
|
"recommendedActions": [
|
|
208
210
|
"Review retrieved memory before acting."
|
|
@@ -223,10 +225,10 @@
|
|
|
223
225
|
"decisionCorrect": true,
|
|
224
226
|
"riskScore": 0.85,
|
|
225
227
|
"passed": true,
|
|
226
|
-
"latencyMs": 0.
|
|
228
|
+
"latencyMs": 0.633,
|
|
227
229
|
"evidenceCount": 1,
|
|
228
230
|
"evidenceIds": [
|
|
229
|
-
"
|
|
231
|
+
"01KXE09KX3RM8CFMH2W52JXVR5"
|
|
230
232
|
],
|
|
231
233
|
"recommendedActions": [
|
|
232
234
|
"Review retrieved memory before acting."
|
|
@@ -247,7 +249,7 @@
|
|
|
247
249
|
"decisionCorrect": false,
|
|
248
250
|
"riskScore": 0,
|
|
249
251
|
"passed": false,
|
|
250
|
-
"latencyMs": 0.
|
|
252
|
+
"latencyMs": 0.528,
|
|
251
253
|
"evidenceCount": 0,
|
|
252
254
|
"evidenceIds": [],
|
|
253
255
|
"recommendedActions": [],
|
|
@@ -274,15 +276,17 @@
|
|
|
274
276
|
"decisionCorrect": true,
|
|
275
277
|
"riskScore": 0.55,
|
|
276
278
|
"passed": true,
|
|
277
|
-
"latencyMs": 4.
|
|
278
|
-
"evidenceCount":
|
|
279
|
+
"latencyMs": 4.913,
|
|
280
|
+
"evidenceCount": 2,
|
|
279
281
|
"evidenceIds": [
|
|
280
|
-
"failure:Bash:2026-07-
|
|
282
|
+
"failure:Bash:2026-07-13T15:07:27.044Z",
|
|
283
|
+
"01KXE09KY55M1KXB4ET1PRF571"
|
|
281
284
|
],
|
|
282
285
|
"recommendedActions": [
|
|
283
|
-
"Before re-running Bash, check what changed since the last failure."
|
|
286
|
+
"Before re-running Bash, check what changed since the last failure.",
|
|
287
|
+
"Mitigate this remembered risk before proceeding."
|
|
284
288
|
],
|
|
285
|
-
"summary": "Caution:
|
|
289
|
+
"summary": "Caution: 2 memory signals, 2 medium severity found before acting.",
|
|
286
290
|
"recallErrors": [],
|
|
287
291
|
"leakedSecrets": [],
|
|
288
292
|
"hasEvidenceForDecision": true,
|
|
@@ -298,7 +302,7 @@
|
|
|
298
302
|
"decisionCorrect": false,
|
|
299
303
|
"riskScore": 0,
|
|
300
304
|
"passed": false,
|
|
301
|
-
"latencyMs": 0.
|
|
305
|
+
"latencyMs": 0.007,
|
|
302
306
|
"evidenceCount": 0,
|
|
303
307
|
"evidenceIds": [],
|
|
304
308
|
"recommendedActions": [],
|
|
@@ -318,10 +322,10 @@
|
|
|
318
322
|
"decisionCorrect": true,
|
|
319
323
|
"riskScore": 0.55,
|
|
320
324
|
"passed": true,
|
|
321
|
-
"latencyMs": 0.
|
|
325
|
+
"latencyMs": 0.102,
|
|
322
326
|
"evidenceCount": 1,
|
|
323
327
|
"evidenceIds": [
|
|
324
|
-
"
|
|
328
|
+
"01KXE09KZB2M0RM2W1H16XTN15"
|
|
325
329
|
],
|
|
326
330
|
"recommendedActions": [
|
|
327
331
|
"Check the recent failed event before repeating a similar action."
|
|
@@ -342,10 +346,10 @@
|
|
|
342
346
|
"decisionCorrect": true,
|
|
343
347
|
"riskScore": 0.35,
|
|
344
348
|
"passed": true,
|
|
345
|
-
"latencyMs": 0.
|
|
349
|
+
"latencyMs": 0.542,
|
|
346
350
|
"evidenceCount": 1,
|
|
347
351
|
"evidenceIds": [
|
|
348
|
-
"
|
|
352
|
+
"01KXE09KZVKXTT4EEPJTQ7BB7G"
|
|
349
353
|
],
|
|
350
354
|
"recommendedActions": [
|
|
351
355
|
"Treat retrieved memory as advisory context."
|
|
@@ -366,7 +370,7 @@
|
|
|
366
370
|
"decisionCorrect": false,
|
|
367
371
|
"riskScore": 0,
|
|
368
372
|
"passed": false,
|
|
369
|
-
"latencyMs": 0.
|
|
373
|
+
"latencyMs": 0.498,
|
|
370
374
|
"evidenceCount": 0,
|
|
371
375
|
"evidenceIds": [],
|
|
372
376
|
"recommendedActions": [],
|
|
@@ -393,15 +397,17 @@
|
|
|
393
397
|
"decisionCorrect": true,
|
|
394
398
|
"riskScore": 0.55,
|
|
395
399
|
"passed": true,
|
|
396
|
-
"latencyMs":
|
|
397
|
-
"evidenceCount":
|
|
400
|
+
"latencyMs": 4.607,
|
|
401
|
+
"evidenceCount": 2,
|
|
398
402
|
"evidenceIds": [
|
|
399
|
-
"failure:Bash:2026-07-
|
|
403
|
+
"failure:Bash:2026-07-13T15:07:27.131Z",
|
|
404
|
+
"01KXE09M0W99EZNCCVB9T6CYVC"
|
|
400
405
|
],
|
|
401
406
|
"recommendedActions": [
|
|
402
|
-
"Before re-running Bash, check what changed since the last failure."
|
|
407
|
+
"Before re-running Bash, check what changed since the last failure.",
|
|
408
|
+
"Mitigate this remembered risk before proceeding."
|
|
403
409
|
],
|
|
404
|
-
"summary": "Caution:
|
|
410
|
+
"summary": "Caution: 2 memory signals, 2 medium severity found before acting.",
|
|
405
411
|
"recallErrors": [],
|
|
406
412
|
"leakedSecrets": [],
|
|
407
413
|
"hasEvidenceForDecision": true,
|
|
@@ -417,7 +423,7 @@
|
|
|
417
423
|
"decisionCorrect": false,
|
|
418
424
|
"riskScore": 0,
|
|
419
425
|
"passed": false,
|
|
420
|
-
"latencyMs": 0.
|
|
426
|
+
"latencyMs": 0.012,
|
|
421
427
|
"evidenceCount": 0,
|
|
422
428
|
"evidenceIds": [],
|
|
423
429
|
"recommendedActions": [],
|
|
@@ -437,10 +443,10 @@
|
|
|
437
443
|
"decisionCorrect": true,
|
|
438
444
|
"riskScore": 0.55,
|
|
439
445
|
"passed": true,
|
|
440
|
-
"latencyMs": 0.
|
|
446
|
+
"latencyMs": 0.09,
|
|
441
447
|
"evidenceCount": 1,
|
|
442
448
|
"evidenceIds": [
|
|
443
|
-
"
|
|
449
|
+
"01KXE09M22ENW4TDBZQCD5D14W"
|
|
444
450
|
],
|
|
445
451
|
"recommendedActions": [
|
|
446
452
|
"Check the recent failed event before repeating a similar action."
|
|
@@ -461,10 +467,10 @@
|
|
|
461
467
|
"decisionCorrect": true,
|
|
462
468
|
"riskScore": 0.35,
|
|
463
469
|
"passed": true,
|
|
464
|
-
"latencyMs": 0.
|
|
470
|
+
"latencyMs": 0.519,
|
|
465
471
|
"evidenceCount": 1,
|
|
466
472
|
"evidenceIds": [
|
|
467
|
-
"
|
|
473
|
+
"01KXE09M2HDYKTN8652470CT9V"
|
|
468
474
|
],
|
|
469
475
|
"recommendedActions": [
|
|
470
476
|
"Treat retrieved memory as advisory context."
|
|
@@ -485,7 +491,7 @@
|
|
|
485
491
|
"decisionCorrect": false,
|
|
486
492
|
"riskScore": 0,
|
|
487
493
|
"passed": false,
|
|
488
|
-
"latencyMs": 0.
|
|
494
|
+
"latencyMs": 0.483,
|
|
489
495
|
"evidenceCount": 0,
|
|
490
496
|
"evidenceIds": [],
|
|
491
497
|
"recommendedActions": [],
|
|
@@ -512,15 +518,15 @@
|
|
|
512
518
|
"decisionCorrect": true,
|
|
513
519
|
"riskScore": 0.2,
|
|
514
520
|
"passed": true,
|
|
515
|
-
"latencyMs":
|
|
521
|
+
"latencyMs": 12.263,
|
|
516
522
|
"evidenceCount": 2,
|
|
517
523
|
"evidenceIds": [
|
|
518
|
-
"
|
|
519
|
-
"
|
|
524
|
+
"01KXE09M3MDMK9ZED7TX1N9YCJ",
|
|
525
|
+
"01KXE09M3HD3936DV3SY6KN54G"
|
|
520
526
|
],
|
|
521
527
|
"recommendedActions": [
|
|
522
528
|
"This exact action has succeeded since its last failure; proceed with normal validation.",
|
|
523
|
-
"
|
|
529
|
+
"Mitigate this remembered risk before proceeding."
|
|
524
530
|
],
|
|
525
531
|
"summary": "Allowed: this exact Bash action has succeeded since the prior failure. Caution: 1 memory signal, 1 medium severity found before acting.",
|
|
526
532
|
"recallErrors": [],
|
|
@@ -538,7 +544,7 @@
|
|
|
538
544
|
"decisionCorrect": true,
|
|
539
545
|
"riskScore": 0,
|
|
540
546
|
"passed": true,
|
|
541
|
-
"latencyMs": 0.
|
|
547
|
+
"latencyMs": 0.007,
|
|
542
548
|
"evidenceCount": 0,
|
|
543
549
|
"evidenceIds": [],
|
|
544
550
|
"recommendedActions": [],
|
|
@@ -558,10 +564,10 @@
|
|
|
558
564
|
"decisionCorrect": false,
|
|
559
565
|
"riskScore": 0.55,
|
|
560
566
|
"passed": false,
|
|
561
|
-
"latencyMs": 0.
|
|
567
|
+
"latencyMs": 0.089,
|
|
562
568
|
"evidenceCount": 1,
|
|
563
569
|
"evidenceIds": [
|
|
564
|
-
"
|
|
570
|
+
"01KXE09M4Y1SNNTHQHKGPY1BD9"
|
|
565
571
|
],
|
|
566
572
|
"recommendedActions": [
|
|
567
573
|
"Check the recent failed event before repeating a similar action."
|
|
@@ -582,10 +588,10 @@
|
|
|
582
588
|
"decisionCorrect": false,
|
|
583
589
|
"riskScore": 0.35,
|
|
584
590
|
"passed": false,
|
|
585
|
-
"latencyMs": 0.
|
|
591
|
+
"latencyMs": 0.622,
|
|
586
592
|
"evidenceCount": 1,
|
|
587
593
|
"evidenceIds": [
|
|
588
|
-
"
|
|
594
|
+
"01KXE09M5EAZWHPSDT7R639VBV"
|
|
589
595
|
],
|
|
590
596
|
"recommendedActions": [
|
|
591
597
|
"Treat retrieved memory as advisory context."
|
|
@@ -606,7 +612,7 @@
|
|
|
606
612
|
"decisionCorrect": true,
|
|
607
613
|
"riskScore": 0,
|
|
608
614
|
"passed": true,
|
|
609
|
-
"latencyMs": 0.
|
|
615
|
+
"latencyMs": 0.507,
|
|
610
616
|
"evidenceCount": 0,
|
|
611
617
|
"evidenceIds": [],
|
|
612
618
|
"recommendedActions": [],
|
|
@@ -633,7 +639,7 @@
|
|
|
633
639
|
"decisionCorrect": true,
|
|
634
640
|
"riskScore": 0.85,
|
|
635
641
|
"passed": true,
|
|
636
|
-
"latencyMs":
|
|
642
|
+
"latencyMs": 4.395,
|
|
637
643
|
"evidenceCount": 1,
|
|
638
644
|
"evidenceIds": [
|
|
639
645
|
"recall:episodic:recall.vector_counts"
|
|
@@ -659,7 +665,7 @@
|
|
|
659
665
|
"decisionCorrect": false,
|
|
660
666
|
"riskScore": 0,
|
|
661
667
|
"passed": false,
|
|
662
|
-
"latencyMs": 0.
|
|
668
|
+
"latencyMs": 0.008,
|
|
663
669
|
"evidenceCount": 0,
|
|
664
670
|
"evidenceIds": [],
|
|
665
671
|
"recommendedActions": [],
|
|
@@ -679,10 +685,10 @@
|
|
|
679
685
|
"decisionCorrect": true,
|
|
680
686
|
"riskScore": 0.85,
|
|
681
687
|
"passed": true,
|
|
682
|
-
"latencyMs": 0.
|
|
688
|
+
"latencyMs": 0.164,
|
|
683
689
|
"evidenceCount": 1,
|
|
684
690
|
"evidenceIds": [
|
|
685
|
-
"
|
|
691
|
+
"01KXE09M7NN3RTZSMYC3NQ84AA"
|
|
686
692
|
],
|
|
687
693
|
"recommendedActions": [
|
|
688
694
|
"Review retrieved memory before acting."
|
|
@@ -703,7 +709,7 @@
|
|
|
703
709
|
"decisionCorrect": false,
|
|
704
710
|
"riskScore": 0.55,
|
|
705
711
|
"passed": false,
|
|
706
|
-
"latencyMs": 0.
|
|
712
|
+
"latencyMs": 0.373,
|
|
707
713
|
"evidenceCount": 0,
|
|
708
714
|
"evidenceIds": [],
|
|
709
715
|
"recommendedActions": [
|
|
@@ -731,7 +737,7 @@
|
|
|
731
737
|
"decisionCorrect": false,
|
|
732
738
|
"riskScore": 0,
|
|
733
739
|
"passed": false,
|
|
734
|
-
"latencyMs": 0.
|
|
740
|
+
"latencyMs": 0.439,
|
|
735
741
|
"evidenceCount": 0,
|
|
736
742
|
"evidenceIds": [],
|
|
737
743
|
"recommendedActions": [],
|
|
@@ -758,11 +764,11 @@
|
|
|
758
764
|
"decisionCorrect": true,
|
|
759
765
|
"riskScore": 0.85,
|
|
760
766
|
"passed": true,
|
|
761
|
-
"latencyMs": 2.
|
|
767
|
+
"latencyMs": 2.962,
|
|
762
768
|
"evidenceCount": 2,
|
|
763
769
|
"evidenceIds": [
|
|
764
770
|
"recall:fts:recall.fts_lookup",
|
|
765
|
-
"
|
|
771
|
+
"01KXE09M92D89KNBQK0BQFN2Y7"
|
|
766
772
|
],
|
|
767
773
|
"recommendedActions": [
|
|
768
774
|
"Do not proceed until the high-severity memory warning is addressed.",
|
|
@@ -785,7 +791,7 @@
|
|
|
785
791
|
"decisionCorrect": false,
|
|
786
792
|
"riskScore": 0,
|
|
787
793
|
"passed": false,
|
|
788
|
-
"latencyMs": 0.
|
|
794
|
+
"latencyMs": 0.008,
|
|
789
795
|
"evidenceCount": 0,
|
|
790
796
|
"evidenceIds": [],
|
|
791
797
|
"recommendedActions": [],
|
|
@@ -805,10 +811,10 @@
|
|
|
805
811
|
"decisionCorrect": false,
|
|
806
812
|
"riskScore": 0.35,
|
|
807
813
|
"passed": false,
|
|
808
|
-
"latencyMs": 0.
|
|
814
|
+
"latencyMs": 0.148,
|
|
809
815
|
"evidenceCount": 1,
|
|
810
816
|
"evidenceIds": [
|
|
811
|
-
"
|
|
817
|
+
"01KXE09MA3ESCT1B6MG4ZMB0Q3"
|
|
812
818
|
],
|
|
813
819
|
"recommendedActions": [
|
|
814
820
|
"Treat retrieved memory as advisory context."
|
|
@@ -829,10 +835,10 @@
|
|
|
829
835
|
"decisionCorrect": false,
|
|
830
836
|
"riskScore": 0.35,
|
|
831
837
|
"passed": false,
|
|
832
|
-
"latencyMs": 0.
|
|
838
|
+
"latencyMs": 0.408,
|
|
833
839
|
"evidenceCount": 1,
|
|
834
840
|
"evidenceIds": [
|
|
835
|
-
"
|
|
841
|
+
"01KXE09MAHS4CCJWTECVYGKJNQ"
|
|
836
842
|
],
|
|
837
843
|
"recommendedActions": [
|
|
838
844
|
"Treat retrieved memory as advisory context."
|
|
@@ -853,7 +859,7 @@
|
|
|
853
859
|
"decisionCorrect": false,
|
|
854
860
|
"riskScore": 0.55,
|
|
855
861
|
"passed": false,
|
|
856
|
-
"latencyMs": 0.
|
|
862
|
+
"latencyMs": 0.17,
|
|
857
863
|
"evidenceCount": 0,
|
|
858
864
|
"evidenceIds": [],
|
|
859
865
|
"recommendedActions": [
|
|
@@ -888,17 +894,19 @@
|
|
|
888
894
|
"decisionCorrect": true,
|
|
889
895
|
"riskScore": 0.9,
|
|
890
896
|
"passed": true,
|
|
891
|
-
"latencyMs":
|
|
892
|
-
"evidenceCount":
|
|
897
|
+
"latencyMs": 4.536,
|
|
898
|
+
"evidenceCount": 3,
|
|
893
899
|
"evidenceIds": [
|
|
894
|
-
"
|
|
895
|
-
"failure:Bash:2026-07-
|
|
900
|
+
"01KXE09MBGSBCHBAMKCDPKPW1N",
|
|
901
|
+
"failure:Bash:2026-07-13T15:07:27.472Z",
|
|
902
|
+
"01KXE09MBGSBCHBAMKCDPKPW1P"
|
|
896
903
|
],
|
|
897
904
|
"recommendedActions": [
|
|
898
905
|
"Do not repeat the exact failed action until the prior error is understood or the command is changed.",
|
|
899
|
-
"Before re-running Bash, check what changed since the last failure."
|
|
906
|
+
"Before re-running Bash, check what changed since the last failure.",
|
|
907
|
+
"Mitigate this remembered risk before proceeding."
|
|
900
908
|
],
|
|
901
|
-
"summary": "Blocked: this exact Bash action failed before. Caution:
|
|
909
|
+
"summary": "Blocked: this exact Bash action failed before. Caution: 2 memory signals, 2 medium severity found before acting.",
|
|
902
910
|
"recallErrors": [],
|
|
903
911
|
"leakedSecrets": [],
|
|
904
912
|
"hasEvidenceForDecision": true,
|
|
@@ -914,7 +922,7 @@
|
|
|
914
922
|
"decisionCorrect": false,
|
|
915
923
|
"riskScore": 0,
|
|
916
924
|
"passed": false,
|
|
917
|
-
"latencyMs": 0.
|
|
925
|
+
"latencyMs": 0.007,
|
|
918
926
|
"evidenceCount": 0,
|
|
919
927
|
"evidenceIds": [],
|
|
920
928
|
"recommendedActions": [],
|
|
@@ -934,10 +942,10 @@
|
|
|
934
942
|
"decisionCorrect": false,
|
|
935
943
|
"riskScore": 0.55,
|
|
936
944
|
"passed": false,
|
|
937
|
-
"latencyMs": 0.
|
|
945
|
+
"latencyMs": 0.075,
|
|
938
946
|
"evidenceCount": 1,
|
|
939
947
|
"evidenceIds": [
|
|
940
|
-
"
|
|
948
|
+
"01KXE09MCKN9T68MQ2SR7S74MX"
|
|
941
949
|
],
|
|
942
950
|
"recommendedActions": [
|
|
943
951
|
"Check the recent failed event before repeating a similar action."
|
|
@@ -958,10 +966,10 @@
|
|
|
958
966
|
"decisionCorrect": false,
|
|
959
967
|
"riskScore": 0.35,
|
|
960
968
|
"passed": false,
|
|
961
|
-
"latencyMs": 0.
|
|
969
|
+
"latencyMs": 0.445,
|
|
962
970
|
"evidenceCount": 1,
|
|
963
971
|
"evidenceIds": [
|
|
964
|
-
"
|
|
972
|
+
"01KXE09MD21TX971RHKPGH6J0S"
|
|
965
973
|
],
|
|
966
974
|
"recommendedActions": [
|
|
967
975
|
"Treat retrieved memory as advisory context."
|
|
@@ -982,7 +990,7 @@
|
|
|
982
990
|
"decisionCorrect": false,
|
|
983
991
|
"riskScore": 0,
|
|
984
992
|
"passed": false,
|
|
985
|
-
"latencyMs": 0.
|
|
993
|
+
"latencyMs": 0.413,
|
|
986
994
|
"evidenceCount": 0,
|
|
987
995
|
"evidenceIds": [],
|
|
988
996
|
"recommendedActions": [],
|
|
@@ -1009,11 +1017,11 @@
|
|
|
1009
1017
|
"decisionCorrect": true,
|
|
1010
1018
|
"riskScore": 0.85,
|
|
1011
1019
|
"passed": true,
|
|
1012
|
-
"latencyMs":
|
|
1020
|
+
"latencyMs": 4.009,
|
|
1013
1021
|
"evidenceCount": 2,
|
|
1014
1022
|
"evidenceIds": [
|
|
1015
|
-
"
|
|
1016
|
-
"
|
|
1023
|
+
"01KXE09ME0E710DBZCDRM6ZAPK",
|
|
1024
|
+
"01KXE09ME1SDSF3N9ER3MEAYRW"
|
|
1017
1025
|
],
|
|
1018
1026
|
"recommendedActions": [
|
|
1019
1027
|
"Do not proceed until the high-severity memory warning is addressed.",
|
|
@@ -1035,7 +1043,7 @@
|
|
|
1035
1043
|
"decisionCorrect": false,
|
|
1036
1044
|
"riskScore": 0,
|
|
1037
1045
|
"passed": false,
|
|
1038
|
-
"latencyMs": 0.
|
|
1046
|
+
"latencyMs": 0.008,
|
|
1039
1047
|
"evidenceCount": 0,
|
|
1040
1048
|
"evidenceIds": [],
|
|
1041
1049
|
"recommendedActions": [],
|
|
@@ -1055,11 +1063,11 @@
|
|
|
1055
1063
|
"decisionCorrect": true,
|
|
1056
1064
|
"riskScore": 0.85,
|
|
1057
1065
|
"passed": true,
|
|
1058
|
-
"latencyMs": 0.
|
|
1066
|
+
"latencyMs": 0.185,
|
|
1059
1067
|
"evidenceCount": 2,
|
|
1060
1068
|
"evidenceIds": [
|
|
1061
|
-
"
|
|
1062
|
-
"
|
|
1069
|
+
"01KXE09MF4FRFAGWDMZYWSSYF3",
|
|
1070
|
+
"01KXE09MF382A2BJNVBV44BH5Q"
|
|
1063
1071
|
],
|
|
1064
1072
|
"recommendedActions": [
|
|
1065
1073
|
"Review retrieved memory before acting."
|
|
@@ -1080,11 +1088,11 @@
|
|
|
1080
1088
|
"decisionCorrect": true,
|
|
1081
1089
|
"riskScore": 0.85,
|
|
1082
1090
|
"passed": true,
|
|
1083
|
-
"latencyMs": 0.
|
|
1091
|
+
"latencyMs": 0.514,
|
|
1084
1092
|
"evidenceCount": 2,
|
|
1085
1093
|
"evidenceIds": [
|
|
1086
|
-
"
|
|
1087
|
-
"
|
|
1094
|
+
"01KXE09MFK7PV5JRT20T18GH2Q",
|
|
1095
|
+
"01KXE09MFMTZF8990X4FNN72R6"
|
|
1088
1096
|
],
|
|
1089
1097
|
"recommendedActions": [
|
|
1090
1098
|
"Review retrieved memory before acting."
|
|
@@ -1105,7 +1113,7 @@
|
|
|
1105
1113
|
"decisionCorrect": false,
|
|
1106
1114
|
"riskScore": 0,
|
|
1107
1115
|
"passed": false,
|
|
1108
|
-
"latencyMs": 0.
|
|
1116
|
+
"latencyMs": 0.477,
|
|
1109
1117
|
"evidenceCount": 0,
|
|
1110
1118
|
"evidenceIds": [],
|
|
1111
1119
|
"recommendedActions": [],
|
|
@@ -1132,22 +1140,22 @@
|
|
|
1132
1140
|
"decisionCorrect": true,
|
|
1133
1141
|
"riskScore": 0.85,
|
|
1134
1142
|
"passed": true,
|
|
1135
|
-
"latencyMs":
|
|
1143
|
+
"latencyMs": 21.585,
|
|
1136
1144
|
"evidenceCount": 13,
|
|
1137
1145
|
"evidenceIds": [
|
|
1138
|
-
"
|
|
1139
|
-
"
|
|
1140
|
-
"
|
|
1141
|
-
"
|
|
1142
|
-
"
|
|
1143
|
-
"
|
|
1144
|
-
"
|
|
1145
|
-
"
|
|
1146
|
-
"
|
|
1147
|
-
"
|
|
1148
|
-
"
|
|
1149
|
-
"
|
|
1150
|
-
"
|
|
1146
|
+
"01KXE09N3T9YNYY0NTHA51W62G",
|
|
1147
|
+
"01KXE09N3KVDJJS43TSNP6ENVC",
|
|
1148
|
+
"01KXE09N2CE328SXRXJCH3Z74A",
|
|
1149
|
+
"01KXE09N26TY6RT4C8JST58PXD",
|
|
1150
|
+
"01KXE09N1MNFD0KGFQQR6JNEWS",
|
|
1151
|
+
"01KXE09N1FJESNKB9AYKCVWN0R",
|
|
1152
|
+
"01KXE09N071JZ8HGQ0V4TC4PF7",
|
|
1153
|
+
"01KXE09MZ1ZTJ6T6KVNCBTRJW2",
|
|
1154
|
+
"01KXE09MWBZ49Q7MT4A1DHEMRY",
|
|
1155
|
+
"01KXE09MPH6KST4EQA6MD9MJCA",
|
|
1156
|
+
"01KXE09MN0SN0T0B9Q4GJFFZ79",
|
|
1157
|
+
"01KXE09MKRBKW8GV17QGFFJERT",
|
|
1158
|
+
"01KXE09MJQHKHMH87DPWAPJ852"
|
|
1151
1159
|
],
|
|
1152
1160
|
"recommendedActions": [
|
|
1153
1161
|
"Do not proceed until the high-severity memory warning is addressed.",
|
|
@@ -1190,10 +1198,10 @@
|
|
|
1190
1198
|
"decisionCorrect": true,
|
|
1191
1199
|
"riskScore": 0.85,
|
|
1192
1200
|
"passed": true,
|
|
1193
|
-
"latencyMs": 0.
|
|
1201
|
+
"latencyMs": 0.516,
|
|
1194
1202
|
"evidenceCount": 1,
|
|
1195
1203
|
"evidenceIds": [
|
|
1196
|
-
"
|
|
1204
|
+
"01KXE09PG3YKMX3PNE2KPWDQX4"
|
|
1197
1205
|
],
|
|
1198
1206
|
"recommendedActions": [
|
|
1199
1207
|
"Review retrieved memory before acting."
|
|
@@ -1214,14 +1222,14 @@
|
|
|
1214
1222
|
"decisionCorrect": false,
|
|
1215
1223
|
"riskScore": 0.35,
|
|
1216
1224
|
"passed": false,
|
|
1217
|
-
"latencyMs": 1.
|
|
1225
|
+
"latencyMs": 1.502,
|
|
1218
1226
|
"evidenceCount": 5,
|
|
1219
1227
|
"evidenceIds": [
|
|
1220
|
-
"
|
|
1221
|
-
"
|
|
1222
|
-
"
|
|
1223
|
-
"
|
|
1224
|
-
"
|
|
1228
|
+
"01KXE09PX05XK1MNQM6HTSJRDX",
|
|
1229
|
+
"01KXE09PRMM0S934H24B8J9RS3",
|
|
1230
|
+
"01KXE09PS93CENDK7GY7FMJ0F1",
|
|
1231
|
+
"01KXE09PNN1Z71Z9SF29F1PNHS",
|
|
1232
|
+
"01KXE09Q1YJF9QHNDRKJ2HNRFG"
|
|
1225
1233
|
],
|
|
1226
1234
|
"recommendedActions": [
|
|
1227
1235
|
"Treat retrieved memory as advisory context."
|
|
@@ -1242,7 +1250,7 @@
|
|
|
1242
1250
|
"decisionCorrect": false,
|
|
1243
1251
|
"riskScore": 0,
|
|
1244
1252
|
"passed": false,
|
|
1245
|
-
"latencyMs": 0.
|
|
1253
|
+
"latencyMs": 0.803,
|
|
1246
1254
|
"evidenceCount": 0,
|
|
1247
1255
|
"evidenceIds": [],
|
|
1248
1256
|
"recommendedActions": [],
|
|
@@ -1257,7 +1265,7 @@
|
|
|
1257
1265
|
}
|
|
1258
1266
|
],
|
|
1259
1267
|
"artifactRedactionSweep": {
|
|
1260
|
-
"checkedAt": "2026-07-
|
|
1268
|
+
"checkedAt": "2026-07-13T15:07:31.207Z",
|
|
1261
1269
|
"filesChecked": [
|
|
1262
1270
|
"benchmarks/output/guardbench-manifest.json",
|
|
1263
1271
|
"benchmarks/output/guardbench-raw.json",
|