{
  "id": "llmjacking-deepseek-sysdig-2025-02",
  "name": "LLMjacking expands to DeepSeek",
  "date_reported": "2025-02-07",
  "researcher": "Sysdig Threat Research Team",
  "org": "Sysdig",
  "sources": [
    {
      "url": "https://www.sysdig.com/blog/llmjacking-targets-deepseek"
    }
  ],
  "magicsword_coverage": [],
  "borderline": false,
  "borderline_note": null,
  "ai_role": "ai_credential_theft",
  "ai_role_secondary": null,
  "ai_role_detail": null,
  "technique_summary": "Within days of DeepSeek virality, LLMjacking proxies added DeepSeek alongside OpenAI, AWS, Azure. Sysdig observed dozen+ proxy servers on stolen credentials; notes Microsoft lawsuit against actors abusing GenAI services (DALL-E abusive content).",
  "initial_access": "unknown",
  "ai_tooling_and_models": [
    "DeepSeek",
    "OpenAI",
    "AWS",
    "Azure"
  ],
  "other_tooling": [],
  "victims_and_sectors": "unknown",
  "regions": [
    "unknown"
  ],
  "attribution_or_cluster": "unknown",
  "scale_and_metrics": "Cloud LLM abuse costs can exceed hundreds of thousands USD monthly (Sysdig)",
  "prevention_angle": "Same as LLMjacking: key hygiene, Bedrock/OpenAI anomaly detection, proxy hunting",
  "related_magicsword_projects": [],
  "source_notes": null,
  "last_reviewed": "2026-10-20",
  "attack_flow": {
    "nodes": [
      {
        "id": "step-1",
        "title": "Original credential theft is undisclosed",
        "description": "The entry begins with already-stolen service credentials.",
        "kind": "unknown",
        "source_urls": [
          "https://www.sysdig.com/blog/llmjacking-targets-deepseek"
        ]
      },
      {
        "id": "step-2",
        "title": "Proxy operators add service support",
        "description": "LLMjacking proxies add DeepSeek to other inference providers.",
        "kind": "execution",
        "source_urls": [
          "https://www.sysdig.com/blog/llmjacking-targets-deepseek"
        ]
      },
      {
        "id": "step-3",
        "title": "Stolen accounts provide inference access",
        "description": "Requests are routed through unauthorized service credentials.",
        "kind": "ai",
        "source_urls": [
          "https://www.sysdig.com/blog/llmjacking-targets-deepseek"
        ]
      },
      {
        "id": "step-4",
        "title": "AI access is resold or abused",
        "description": "The report describes more than a dozen proxy servers using stolen credentials.",
        "kind": "impact",
        "source_urls": [
          "https://www.sysdig.com/blog/llmjacking-targets-deepseek"
        ]
      }
    ],
    "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": "Stolen service credentials provide unauthorized inference access.",
      "source_urls": [
        "https://www.sysdig.com/blog/llmjacking-targets-deepseek"
      ]
    }
  ],
  "category": "attacks-on-ai",
  "ai_involvement_confidence": "confirmed",
  "detections": [
    {
      "name": "Review attack-specific indicators",
      "format": "guidance",
      "notes": "Monitor for the techniques described in LLMjacking expands to DeepSeek. Refer to sources for specific IOCs and TTPs."
    }
  ]
}