UTC

AML.CS0073

Autonomous Claude Agents Compromise Three Real Organizations During Cyber Evaluations

Type
Incident
When
2026-04 to 2026-07
Actor, as ATLAS names it
Autonomous AI agents: Claude Opus 4.7, Claude Mythos 5, and an internal Anthropic research model
Target
3 organizations
Techniques, in order
  1. 1Autonomous Attack-Path AdaptationAI Attack Adaptation · AML.T0117
  2. 2Autonomous ReconnaissanceReconnaissance · AML.T0116
  3. 3Exploit Public-Facing ApplicationInitial Access · AML.T0049
  4. 4Exploitation for Credential AccessCredential Access · AML.T0106
  5. 5Develop Capabilities: MalwareResource Development · AML.T0017.003
  6. 6Establish AccountsResource Development · AML.T0021
  7. 7Stage CapabilitiesResource Development · AML.T0079
  8. 8Supply Chain Compromise: SoftwareInitial Access · AML.T0010.001
  9. 9Command and Scripting InterpreterExecution · AML.T0050
  10. 10Exfiltration via Cyber MeansExfiltration · AML.T0025
  11. 11Valid AccountsLateral Movement · AML.T0012
  12. 12Active ScanningReconnaissance · AML.T0006
  13. 13Unsecured CredentialsCredential Access · AML.T0055
The full case study on MITRE ATLAS →

From MITRE ATLAS 2026.10 (Apache-2.0). ShieldMap lists who and what was involved and the techniques used, and links to ATLAS for the write-up; it does not copy the steps. Attribution is ATLAS's, not ShieldPi's.