Cost harvesting / LLMjacking
Stolen cloud credentials run LLM workloads on the victim's bill
categoryAI/ML and emerging attacksseverityHighlabnot yet measured
Example: up to $46k/day
Classification
16 values: 10 reviewed, 5 imported, 1 derived. What do these mean?
Exploitability
| Attack vector | AI model manipulation |
|---|---|
| Reach (CVSS) | Network |
| Attack complexity | Low |
| Privileges required | Low |
| User interaction | None |
| Attacker skill required | Low |
| Resources required | Medium |
| Likelihood / prevalence | High |
| Maturity | Theoretical |
Impact & scope
| Severity | High |
|---|---|
| Impact | AvailabilityFinancial |
| Platform | AI System |
| Environment | AI system |
| Target asset class | ML model / agent |
| Kill-chain phase (tactic) | Impact |
Detection & response
Mitigated by 4
- AML.M0004 atlas-mitigation
- AML.M0019 atlas-mitigation
- AML.M0035 atlas-mitigation
- AML.M0036 atlas-mitigation
Framework references
- Primary source: https://www.sysdig.com/blog/llmjacking-stolen-cloud-credentials-used-in-new-ai-attack
- atlas: AML.T0034 (primary)
Evidence & sources
- ev-0001 (Cyber Attack Taxonomy) — Name, definition and Example/ID taken verbatim from the 'AI/ML and emerging attacks' table, line 627 of Cyber-Attack-Taxonomy.md