{
  "id": "wiz-ai-driven-cloud-2026-10",
  "name": "Wiz CIRT AI-driven multi-cloud intrusions",
  "date_reported": "2026-10-19",
  "researcher": "Eden Abergil, Zohar Kaplan, Kurt Giacchino",
  "org": "Wiz CIRT",
  "sources": [
    {
      "url": "https://www.wiz.io/blog/anatomy-of-ai-driven-cloud-attacks"
    }
  ],
  "magicsword_coverage": [],
  "borderline": false,
  "borderline_note": null,
  "ai_role": "attacker_uses_ai_as_operator",
  "ai_role_secondary": null,
  "ai_role_detail": null,
  "technique_summary": "AI agent ran full cloud operations: initial access, credential theft, pivots across GCP/AWS/K8s/CI-CD/IdP, debugging its own scripts. Loud (thousands of actions/hour), pentest-named resources, numeric task IDs with -b/-c retries, commented generated scripts. Lived off SSM, SA token minting, IdP admin creation, GitOps manifests, runner-log exfil.",
  "initial_access": "Confirmed Java RCE in exposed Kafka UI among multiple probed doors; same source IP scanned 600+ other Wiz customer environments",
  "ai_tooling_and_models": [
    "Unnamed AI agent system (Wiz: agent ran the whole operation)"
  ],
  "other_tooling": [
    "Kafka UI exploit",
    "SSM",
    "GCP SA token creation",
    "IdP admin APIs",
    "CI/CD pipelines",
    "GitOps/K8s manifests",
    "webhook.site",
    "IMDS credential harvest"
  ],
  "victims_and_sectors": "Large multi-cloud organization (Wiz customer); opportunistic scanning across many other customers",
  "regions": [
    "unknown"
  ],
  "attribution_or_cluster": "unknown",
  "scale_and_metrics": "Hundreds–thousands of actions per hour; 600+ environments scanned from same IP",
  "prevention_angle": "Unify multi-platform logs; runtime visibility on runners/VMs/pods; restrict cross-system trust; rotate credentials on compromise; correlate loud agent tells (pentest- prefixes, task IDs)",
  "related_magicsword_projects": [],
  "source_notes": null,
  "last_reviewed": "2026-10-20",
  "attack_flow": {
    "nodes": [
      {
        "id": "step-1",
        "title": "Exposed Kafka UI is exploited",
        "description": "The report confirms a Java RCE route among multiple probed entry points.",
        "kind": "access",
        "source_urls": [
          "https://www.wiz.io/blog/anatomy-of-ai-driven-cloud-attacks"
        ]
      },
      {
        "id": "step-2",
        "title": "Agent executes and debugs cloud operations",
        "description": "The automation generates and retries scripts at high volume.",
        "kind": "ai",
        "source_urls": [
          "https://www.wiz.io/blog/anatomy-of-ai-driven-cloud-attacks"
        ]
      },
      {
        "id": "step-3",
        "title": "Credentials enable cross-service pivots",
        "description": "Reported paths span cloud accounts, Kubernetes, identity and CI/CD.",
        "kind": "execution",
        "source_urls": [
          "https://www.wiz.io/blog/anatomy-of-ai-driven-cloud-attacks"
        ]
      },
      {
        "id": "step-4",
        "title": "Data leaves through operational channels",
        "description": "The research describes runner-log exfiltration and abuse of legitimate cloud-management capabilities.",
        "kind": "impact",
        "source_urls": [
          "https://www.wiz.io/blog/anatomy-of-ai-driven-cloud-attacks"
        ]
      }
    ],
    "edges": [
      {
        "source": "step-1",
        "target": "step-2"
      },
      {
        "source": "step-2",
        "target": "step-3"
      },
      {
        "source": "step-3",
        "target": "step-4"
      }
    ]
  },
  "atlas_techniques": [
    {
      "id": "AML.T0016.002",
      "rationale": "The supplied source explicitly describes generative-model tooling assisting the reported activity; this mapping does not establish additional autonomy or victim impact.",
      "source_urls": [
        "https://www.wiz.io/blog/anatomy-of-ai-driven-cloud-attacks"
      ]
    }
  ],
  "category": "adversary-use-of-ai",
  "ai_involvement_confidence": "confirmed",
  "detections": [
    {
      "name": "Review attack-specific indicators",
      "format": "guidance",
      "notes": "Monitor for the techniques described in Wiz CIRT AI-driven multi-cloud intrusions. Refer to sources for specific IOCs and TTPs."
    }
  ]
}