DIGGINGBEAGLE

Topic · DiggingBeagle record

Stolen AI credentials as attack infrastructure

Compromised API keys provide compute, cover and resale value to attackers.

Topic kind
pattern

Definition & limits

AI credentials are not only billing secrets. Threat actors can route their own operations through a victim's keys, shift attribution and gain access to provider capabilities.

Examples

  • GTG-50020 switched workloads to stolen production API keys.
  • GTG-50029 scanned for exposed API keys and rotated their use.

Research using this topic (13)

AnalysisWhat changes when one operator can run a multi-agent attack loopAnalysisAn exposed AI credential is a sequence of permissions, not one breach numberCaseCARLIY0S orchestrated commercial AI models in a reported Latin American financial attack operationCaseStorm-0940 used ARTEX and stolen ConnectWise application credentialsCaseGemini crossed an Irregular cyber evaluation boundary and accessed three real companiesCaseGTG-20006 automated parts of a Russian espionage workflowCaseGTG-50020 used prompt injection against an AI evaluation sandboxCaseGTG-50029 used agentic workflows across a European hacktivist campaignCasemulti-agent-pipeline setup allegedly sent GitHub CLI credentials to a registration endpointCaseLangGraph verbose deploy output disclosed a temporary image registry tokenCaseNorthflank customers reported unauthorized use of stored AI provider keysNewsGitHub CLI adds optional expiring OAuth tokens for automated developer sessionsNewsLangGraph deploy verbose output exposed a temporary registry push token