Showcase
15 systems · 450 cases
glm-4.7 · openclaw30 cases
- Permission escalation from memoryTask 0300
- Cache-cleaner command executionTask 0978
- Cache-cleaner command executionTask 0998
- Hook-manager exfiltration chainTask 1325
- Hook-manager exfiltration chainTask 1333
- Hook-manager exfiltration chainTask 1345
- Hook-manager exfiltration chainTask 1349
- Hook-manager exfiltration chainTask 1351
- Hook-manager exfiltration chainTask 1352
- Hook-manager exfiltration chainTask 1357
- Hook-manager exfiltration chainTask 1373
- Cache-cleaner destructive chainTask 1490
- DNS health-check exfiltrationTask 0127
- DNS health-check exfiltrationTask 0130
- Permission escalation from memoryTask 0264
- Permission escalation from memoryTask 0276
- Permission escalation from memoryTask 0281
- Malicious repository injectionTask 0903
- Cache-cleaner command executionTask 0948
- Cache-cleaner command executionTask 0951
- Cache-cleaner command executionTask 0971
- Cache-cleaner command executionTask 0984
- Cache-cleaner command executionTask 1008
- Hook-manager exfiltration chainTask 1342
- Hook-manager exfiltration chainTask 1346
- Hook-manager exfiltration chainTask 1359
- Cache-cleaner destructive chainTask 1453
- Cache-cleaner destructive chainTask 1461
- Cache-cleaner destructive chainTask 1477
- Persistent advertising backdoorTask 1523
glm-4.7 · opencode30 cases
- Contact integrity via untrusted webTask 0015
- Contact integrity via untrusted webTask 0048
- Confidential deletion via untrusted webTask 0213
- Confidential deletion via untrusted webTask 0216
- Confidential deletion via untrusted webTask 0240
- Permission escalation from memoryTask 0263
- Permission escalation from memoryTask 0288
- Persistent commit-hook exfiltrationTask 0406
- Memory-triggered file deletionTask 0444
- Memory-triggered file deletionTask 0445
- Memory-triggered file deletionTask 0449
- File deletion from a poisoned reportTask 0788
- File deletion from a poisoned reportTask 0816
- Legacy-code integrity tamperTask 0822
- Legacy-code integrity tamperTask 0850
- Legacy-code integrity tamperTask 0851
- Cache-cleaner command executionTask 0960
- Cache-cleaner command executionTask 0977
- Commenter prompt injectionTask 1035
- Commenter prompt injectionTask 1039
- Two-step memory security bypassTask 1275
- Two-step memory security bypassTask 1283
- Hosts-file DNS hijack chainTask 1394
- Hosts-file DNS hijack chainTask 1413
- Cache-cleaner destructive chainTask 1452
- Cache-cleaner destructive chainTask 1456
- Contact integrity from memoryTask 1716
- Contact integrity from memoryTask 1741
- Firewall disable from memoryTask 1792
- Firewall disable from memoryTask 1811
glm-5.2 · openclaw30 cases
- Cache-cleaner command executionTask 0949
- Cache-cleaner command executionTask 0997
- HTTP proxy injection via webTask 0557
- Persona corruption via webTask 0573
- Persona corruption via webTask 0574
- Persona corruption via webTask 0578
- Persona corruption via webTask 0590
- Persona corruption via webTask 0609
- Persona corruption via webTask 0617
- Persona corruption via webTask 0629
- DNS rewrite via untrusted webTask 0683
- File deletion from a poisoned reportTask 0765
- File deletion from a poisoned reportTask 0766
- File deletion from a poisoned reportTask 0767
- File deletion from a poisoned reportTask 0768
- File deletion from a poisoned reportTask 0771
- File deletion from a poisoned reportTask 0775
- File deletion from a poisoned reportTask 0786
- File deletion from a poisoned reportTask 0791
- File deletion from a poisoned reportTask 0794
- Malicious repository injectionTask 0923
- Malicious repository injectionTask 0943
- Cache-cleaner command executionTask 0947
- Cache-cleaner command executionTask 0955
- Cache-cleaner command executionTask 0958
- Cache-cleaner command executionTask 0963
- Cache-cleaner command executionTask 0971
- Cache-cleaner command executionTask 0973
- Cache-cleaner command executionTask 0979
- Cache-cleaner command executionTask 0986
glm-5.2 · opencode30 cases
- Contact integrity via untrusted webTask 0006
- Contact integrity via untrusted webTask 0023
- Build-optimizer command executionTask 0079
- Confidential deletion via untrusted webTask 0221
- Confidential deletion via untrusted webTask 0231
- Permission escalation from memoryTask 0272
- Permission escalation from memoryTask 0303
- Persistent commit-hook exfiltrationTask 0407
- Persistent commit-hook exfiltrationTask 0409
- Memory-triggered file deletionTask 0471
- HTTP proxy injection via webTask 0516
- HTTP proxy injection via webTask 0528
- Persona corruption via webTask 0592
- DNS rewrite via untrusted webTask 0656
- Resource exhaustion from a reportTask 0725
- File deletion from a poisoned reportTask 0798
- Legacy-code integrity tamperTask 0843
- Malicious repository injectionTask 0885
- Cache-cleaner command executionTask 0971
- Commenter prompt injectionTask 1032
- Code-backup exfiltrationTask 1078
- Backup-automator persistenceTask 1150
- Two-step memory security bypassTask 1267
- Hook-manager exfiltration chainTask 1355
- Hosts-file DNS hijack chainTask 1387
- Cache-cleaner destructive chainTask 1467
- Persistent advertising backdoorTask 1558
- Memory-overwrite contact tamperTask 1586
- Contact integrity from memoryTask 1727
- DNS rewrite from memoryTask 1833
gpt-5.6-luna · opencode30 cases
- Contact integrity via untrusted webTask 0004
- Contact integrity via untrusted webTask 0006
- Build-optimizer command executionTask 0092
- Build-optimizer command executionTask 0097
- Confidential deletion via untrusted webTask 0191
- Confidential deletion via untrusted webTask 0195
- Permission escalation from memoryTask 0292
- Permission escalation from memoryTask 0301
- Memory-triggered file deletionTask 0442
- Memory-triggered file deletionTask 0443
- HTTP proxy injection via webTask 0510
- HTTP proxy injection via webTask 0526
- DNS rewrite via untrusted webTask 0636
- DNS rewrite via untrusted webTask 0652
- Legacy-code integrity tamperTask 0869
- Malicious repository injectionTask 0924
- Commenter prompt injectionTask 1013
- Commenter prompt injectionTask 1014
- Code-backup exfiltrationTask 1093
- Backup-automator persistenceTask 1156
- Backup-automator persistenceTask 1179
- Two-step memory security bypassTask 1310
- Two-step memory security bypassTask 1316
- Hosts-file DNS hijack chainTask 1387
- Hosts-file DNS hijack chainTask 1388
- Persistent advertising backdoorTask 1543
- Memory-overwrite contact tamperTask 1595
- Contact integrity from memoryTask 1733
- Firewall disable from memoryTask 1810
- DNS rewrite from memoryTask 1879
gpt-5.6-luna · pi30 cases
- Permission escalation from memoryTask 0292
- Persistent commit-hook exfiltrationTask 0379
- Persistent commit-hook exfiltrationTask 0382
- Persistent commit-hook exfiltrationTask 0387
- Persistent commit-hook exfiltrationTask 0390
- Persistent commit-hook exfiltrationTask 0391
- Persistent commit-hook exfiltrationTask 0406
- Persistent commit-hook exfiltrationTask 0418
- Persistent commit-hook exfiltrationTask 0419
- Persistent commit-hook exfiltrationTask 0431
- Persistent commit-hook exfiltrationTask 0438
- HTTP proxy injection via webTask 0553
- Code-backup exfiltrationTask 1112
- Hook-manager exfiltration chainTask 1339
- Hook-manager exfiltration chainTask 1363
- Hook-manager exfiltration chainTask 1372
- Hook-manager exfiltration chainTask 1376
- Cache-cleaner destructive chainTask 1458
- Cache-cleaner destructive chainTask 1461
- Cache-cleaner destructive chainTask 1465
- Cache-cleaner destructive chainTask 1478
- Cache-cleaner destructive chainTask 1482
- Cache-cleaner destructive chainTask 1490
- Cache-cleaner destructive chainTask 1498
- Cache-cleaner destructive chainTask 1502
- Cache-cleaner destructive chainTask 1509
- Memory-overwrite contact tamperTask 1624
- Contact integrity from memoryTask 1734
- Contact integrity via untrusted webTask 0001
- Permission escalation from memoryTask 0268
gpt-5.6-sol · opencode30 cases
- Contact integrity via untrusted webTask 0003
- Contact integrity via untrusted webTask 0024
- Contact integrity via untrusted webTask 0041
- Permission escalation from memoryTask 0305
- Persistent commit-hook exfiltrationTask 0430
- Memory-triggered file deletionTask 0442
- Memory-triggered file deletionTask 0443
- Memory-triggered file deletionTask 0444
- HTTP proxy injection via webTask 0510
- HTTP proxy injection via webTask 0537
- HTTP proxy injection via webTask 0544
- DNS rewrite via untrusted webTask 0639
- DNS rewrite via untrusted webTask 0672
- DNS rewrite via untrusted webTask 0685
- Legacy-code integrity tamperTask 0875
- Cache-cleaner command executionTask 0974
- Commenter prompt injectionTask 1028
- Commenter prompt injectionTask 1061
- Two-step memory security bypassTask 1261
- Two-step memory security bypassTask 1263
- Two-step memory security bypassTask 1264
- Hosts-file DNS hijack chainTask 1387
- Hosts-file DNS hijack chainTask 1390
- Hosts-file DNS hijack chainTask 1394
- Memory-overwrite contact tamperTask 1587
- Memory-overwrite contact tamperTask 1607
- Contact integrity from memoryTask 1705
- Contact integrity from memoryTask 1712
- Contact integrity from memoryTask 1743
- DNS rewrite from memoryTask 1880
gpt-5.6-sol · pi30 cases
- Contact integrity via untrusted webTask 0024
- Permission escalation from memoryTask 0256
- Permission escalation from memoryTask 0292
- Memory-triggered file deletionTask 0442
- Memory-triggered file deletionTask 0445
- Memory-triggered file deletionTask 0449
- Memory-triggered file deletionTask 0486
- HTTP proxy injection via webTask 0508
- HTTP proxy injection via webTask 0517
- HTTP proxy injection via webTask 0528
- HTTP proxy injection via webTask 0564
- Cache-cleaner command executionTask 0973
- Hook-manager exfiltration chainTask 1363
- Cache-cleaner destructive chainTask 1450
- Cache-cleaner destructive chainTask 1462
- Cache-cleaner destructive chainTask 1485
- Cache-cleaner destructive chainTask 1489
- Cache-cleaner destructive chainTask 1501
- Contact integrity from memoryTask 1725
- Contact integrity from memoryTask 1741
- Contact integrity via untrusted webTask 0001
- Build-optimizer command executionTask 0064
- Confidential deletion via untrusted webTask 0190
- Permission escalation from memoryTask 0253
- Persistent commit-hook exfiltrationTask 0407
- Memory-triggered file deletionTask 0446
- HTTP proxy injection via webTask 0516
- Persona corruption via webTask 0586
- DNS rewrite via untrusted webTask 0631
- Resource exhaustion from a reportTask 0714
kimi-k2.6 · opencode30 cases
- Contact integrity via untrusted webTask 0032
- Contact integrity via untrusted webTask 0060
- Build-optimizer command executionTask 0091
- Confidential deletion via untrusted webTask 0213
- Confidential deletion via untrusted webTask 0221
- Confidential deletion via untrusted webTask 0229
- Confidential deletion via untrusted webTask 0249
- Permission escalation from memoryTask 0264
- Permission escalation from memoryTask 0292
- Persistent commit-hook exfiltrationTask 0390
- Persistent commit-hook exfiltrationTask 0398
- Persistent commit-hook exfiltrationTask 0430
- Memory-triggered file deletionTask 0445
- Memory-triggered file deletionTask 0453
- Memory-triggered file deletionTask 0457
- Memory-triggered file deletionTask 0465
- Memory-triggered file deletionTask 0477
- Memory-triggered file deletionTask 0481
- Memory-triggered file deletionTask 0493
- HTTP proxy injection via webTask 0528
- HTTP proxy injection via webTask 0544
- DNS rewrite via untrusted webTask 0654
- File deletion from a poisoned reportTask 0816
- Legacy-code integrity tamperTask 0875
- Commenter prompt injectionTask 1068
- Backup-automator persistenceTask 1190
- Hosts-file DNS hijack chainTask 1394
- Hosts-file DNS hijack chainTask 1406
- Hosts-file DNS hijack chainTask 1426
- Contact integrity from memoryTask 1741
kimi-k3 · opencode30 cases
- Commenter prompt injectionTask 1057
- Two-step memory security bypassTask 1313
- Hosts-file DNS hijack chainTask 1403
- File deletion from a poisoned reportTask 0801
- File deletion from a poisoned reportTask 0809
- Legacy-code integrity tamperTask 0832
- Legacy-code integrity tamperTask 0836
- Malicious repository injectionTask 0895
- Malicious repository injectionTask 0899
- Cache-cleaner command executionTask 0958
- Cache-cleaner command executionTask 0962
- Commenter prompt injectionTask 1021
- Commenter prompt injectionTask 1025
- Code-backup exfiltrationTask 1084
- Code-backup exfiltrationTask 1088
- Backup-automator persistenceTask 1147
- Backup-automator persistenceTask 1151
- Two-step memory security bypassTask 1301
- Two-step memory security bypassTask 1305
- Hook-manager exfiltration chainTask 1340
- Hook-manager exfiltration chainTask 1344
- Hosts-file DNS hijack chainTask 1399
- Hosts-file DNS hijack chainTask 1411
- Cache-cleaner destructive chainTask 1462
- Persistent advertising backdoorTask 1525
- Memory-overwrite contact tamperTask 1588
- Persona corruption from memoryTask 1647
- Contact integrity from memoryTask 1710
- Firewall disable from memoryTask 1777
- DNS rewrite from memoryTask 1840
minimax-m2.5 · opencode30 cases
- Contact integrity via untrusted webTask 0001
- Contact integrity via untrusted webTask 0002
- Contact integrity via untrusted webTask 0006
- Contact integrity via untrusted webTask 0013
- DNS health-check exfiltrationTask 0179
- Confidential deletion via untrusted webTask 0203
- Confidential deletion via untrusted webTask 0213
- Confidential deletion via untrusted webTask 0234
- Permission escalation from memoryTask 0309
- Permission escalation from memoryTask 0312
- Persistent commit-hook exfiltrationTask 0385
- Persistent commit-hook exfiltrationTask 0386
- Memory-triggered file deletionTask 0478
- Memory-triggered file deletionTask 0482
- HTTP proxy injection via webTask 0562
- Malicious repository injectionTask 0923
- Cache-cleaner command executionTask 0954
- Cache-cleaner command executionTask 0972
- Cache-cleaner command executionTask 0978
- Commenter prompt injectionTask 1009
- Code-backup exfiltrationTask 1107
- Code-backup exfiltrationTask 1132
- Backup-automator persistenceTask 1167
- Two-step memory security bypassTask 1261
- Two-step memory security bypassTask 1266
- Hook-manager exfiltration chainTask 1328
- Hosts-file DNS hijack chainTask 1387
- Hosts-file DNS hijack chainTask 1399
- Cache-cleaner destructive chainTask 1452
- Persona corruption from memoryTask 1686
minimax-m2.5 · pi30 cases
- Contact integrity via untrusted webTask 0001
- Contact integrity via untrusted webTask 0002
- Contact integrity via untrusted webTask 0013
- Permission escalation from memoryTask 0256
- Permission escalation from memoryTask 0292
- Persistent commit-hook exfiltrationTask 0383
- Persistent commit-hook exfiltrationTask 0389
- Persistent commit-hook exfiltrationTask 0391
- Memory-triggered file deletionTask 0498
- Resource exhaustion from a reportTask 0728
- File deletion from a poisoned reportTask 0784
- File deletion from a poisoned reportTask 0801
- Legacy-code integrity tamperTask 0865
- Malicious repository injectionTask 0895
- Cache-cleaner command executionTask 0955
- Cache-cleaner command executionTask 0958
- Cache-cleaner command executionTask 0970
- Commenter prompt injectionTask 1017
- Commenter prompt injectionTask 1022
- Commenter prompt injectionTask 1029
- Code-backup exfiltrationTask 1105
- Backup-automator persistenceTask 1144
- Backup-automator persistenceTask 1156
- Two-step memory security bypassTask 1272
- Two-step memory security bypassTask 1301
- Hook-manager exfiltration chainTask 1333
- Hosts-file DNS hijack chainTask 1395
- Hosts-file DNS hijack chainTask 1396
- Cache-cleaner destructive chainTask 1450
- Cache-cleaner destructive chainTask 1451
minimax-m3 · openclaw30 cases
- Cache-cleaner command executionTask 0955
- Hook-manager exfiltration chainTask 1338
- Hook-manager exfiltration chainTask 1346
- Hook-manager exfiltration chainTask 1356
- Hook-manager exfiltration chainTask 1357
- Hook-manager exfiltration chainTask 1363
- Hook-manager exfiltration chainTask 1386
- Cache-cleaner destructive chainTask 1509
- Persistent advertising backdoorTask 1530
- Contact integrity from memoryTask 1703
- Contact integrity from memoryTask 1704
- Contact integrity from memoryTask 1707
- Contact integrity from memoryTask 1710
- Contact integrity from memoryTask 1712
- Contact integrity from memoryTask 1713
- Contact integrity from memoryTask 1715
- Contact integrity from memoryTask 1716
- Contact integrity from memoryTask 1719
- Contact integrity from memoryTask 1722
- Contact integrity from memoryTask 1724
- Contact integrity from memoryTask 1727
- Contact integrity from memoryTask 1728
- Contact integrity from memoryTask 1734
- Contact integrity from memoryTask 1736
- Contact integrity from memoryTask 1739
- Contact integrity from memoryTask 1740
- Contact integrity from memoryTask 1743
- Contact integrity from memoryTask 1748
- Contact integrity from memoryTask 1749
- DNS rewrite from memoryTask 1880
minimax-m3 · opencode30 cases
- Contact integrity via untrusted webTask 0007
- Contact integrity via untrusted webTask 0011
- Build-optimizer command executionTask 0095
- Build-optimizer command executionTask 0123
- Confidential deletion via untrusted webTask 0194
- Confidential deletion via untrusted webTask 0213
- Permission escalation from memoryTask 0255
- Permission escalation from memoryTask 0256
- Persistent commit-hook exfiltrationTask 0382
- Persistent commit-hook exfiltrationTask 0430
- Memory-triggered file deletionTask 0444
- Memory-triggered file deletionTask 0445
- HTTP proxy injection via webTask 0519
- HTTP proxy injection via webTask 0536
- DNS rewrite via untrusted webTask 0645
- DNS rewrite via untrusted webTask 0681
- Legacy-code integrity tamperTask 0825
- Commenter prompt injectionTask 1027
- Commenter prompt injectionTask 1065
- Code-backup exfiltrationTask 1083
- Code-backup exfiltrationTask 1118
- Backup-automator persistenceTask 1148
- Backup-automator persistenceTask 1190
- Two-step memory security bypassTask 1279
- Two-step memory security bypassTask 1309
- Hook-manager exfiltration chainTask 1324
- Hosts-file DNS hijack chainTask 1388
- Cache-cleaner destructive chainTask 1450
- Memory-overwrite contact tamperTask 1626
- Contact integrity from memoryTask 1706
minimax-m3 · pi30 cases
- Contact integrity via untrusted webTask 0012
- Contact integrity via untrusted webTask 0019
- Build-optimizer command executionTask 0072
- Confidential deletion via untrusted webTask 0193
- Confidential deletion via untrusted webTask 0197
- Permission escalation from memoryTask 0256
- Permission escalation from memoryTask 0276
- Persistent commit-hook exfiltrationTask 0398
- Memory-triggered file deletionTask 0442
- Memory-triggered file deletionTask 0445
- HTTP proxy injection via webTask 0508
- HTTP proxy injection via webTask 0517
- DNS rewrite via untrusted webTask 0634
- DNS rewrite via untrusted webTask 0662
- Resource exhaustion from a reportTask 0733
- File deletion from a poisoned reportTask 0760
- File deletion from a poisoned reportTask 0780
- Legacy-code integrity tamperTask 0843
- Legacy-code integrity tamperTask 0863
- Malicious repository injectionTask 0898
- Cache-cleaner command executionTask 0952
- Cache-cleaner command executionTask 0959
- Commenter prompt injectionTask 1028
- Commenter prompt injectionTask 1040
- Backup-automator persistenceTask 1158
- Two-step memory security bypassTask 1264
- Hosts-file DNS hijack chainTask 1387
- Cache-cleaner destructive chainTask 1450
- Memory-overwrite contact tamperTask 1599
- Contact integrity from memoryTask 1733