{
  "id": "datadog-bedrock-key-testing-2025",
  "name": "Exposed AWS keys mass-tested for Bedrock LLMjacking",
  "date_reported": "2025-06-01",
  "researcher": "Datadog Security Labs",
  "org": "Datadog",
  "sources": [
    {
      "url": "https://securitylabs.datadoghq.com/articles/beyond-valid-credentials-how-exposed-aws-keys-are-tested-for-amazon-bedrock-access/"
    }
  ],
  "magicsword_coverage": [],
  "borderline": false,
  "borderline_note": null,
  "ai_role": "ai_credential_theft",
  "ai_role_secondary": null,
  "ai_role_detail": null,
  "technique_summary": "Datadog documents automated testing of exposed AWS keys specifically for Amazon Bedrock access (ListFoundationModels / Invoke / Converse patterns), showing LLMjacking reconnaissance at internet scale beyond generic AWS abuse.",
  "initial_access": "unknown",
  "ai_tooling_and_models": [
    "Amazon Bedrock"
  ],
  "other_tooling": [
    "Stolen AKIA keys"
  ],
  "victims_and_sectors": "unknown",
  "regions": [
    "unknown"
  ],
  "attribution_or_cluster": "unknown",
  "scale_and_metrics": "unknown",
  "prevention_angle": "Secret scanning; CloudTrail detections for Bedrock enum+invoke from unexpected keys",
  "related_magicsword_projects": [],
  "source_notes": null,
  "last_reviewed": "2026-10-20",
  "attack_flow": {
    "nodes": [
      {
        "id": "step-1",
        "title": "Exposed AWS keys are collected",
        "description": "Attackers obtain credentials exposed outside the legitimate account boundary.",
        "kind": "access",
        "source_urls": [
          "https://securitylabs.datadoghq.com/articles/beyond-valid-credentials-how-exposed-aws-keys-are-tested-for-amazon-bedrock-access/"
        ]
      },
      {
        "id": "step-2",
        "title": "Automated key tests run",
        "description": "Requests check model listing, invocation and conversation permissions.",
        "kind": "execution",
        "source_urls": [
          "https://securitylabs.datadoghq.com/articles/beyond-valid-credentials-how-exposed-aws-keys-are-tested-for-amazon-bedrock-access/"
        ]
      },
      {
        "id": "step-3",
        "title": "Bedrock access is identified",
        "description": "Testing distinguishes keys that can use model inference services.",
        "kind": "ai",
        "source_urls": [
          "https://securitylabs.datadoghq.com/articles/beyond-valid-credentials-how-exposed-aws-keys-are-tested-for-amazon-bedrock-access/"
        ]
      },
      {
        "id": "step-4",
        "title": "Per-key abuse is not quantified",
        "description": "The supplied entry establishes reconnaissance at scale, rather than a bill or victim count for every key.",
        "kind": "unknown",
        "source_urls": [
          "https://securitylabs.datadoghq.com/articles/beyond-valid-credentials-how-exposed-aws-keys-are-tested-for-amazon-bedrock-access/"
        ]
      }
    ],
    "edges": [
      {
        "source": "step-1",
        "target": "step-2"
      },
      {
        "source": "step-2",
        "target": "step-3"
      },
      {
        "source": "step-3",
        "target": "step-4"
      }
    ]
  },
  "atlas_techniques": [
    {
      "id": "AML.T0012",
      "rationale": "Exposed AWS credentials are tested for model-service access.",
      "source_urls": [
        "https://securitylabs.datadoghq.com/articles/beyond-valid-credentials-how-exposed-aws-keys-are-tested-for-amazon-bedrock-access/"
      ]
    }
  ],
  "category": "attacks-on-ai",
  "ai_involvement_confidence": "confirmed",
  "detections": [
    {
      "name": "Review attack-specific indicators",
      "format": "guidance",
      "notes": "Monitor for the techniques described in Exposed AWS keys mass-tested for Bedrock LLMjacking. Refer to sources for specific IOCs and TTPs."
    }
  ]
}