Attackers who obtain AWS credentials are actively validating them to determine if they can access Amazon Bedrock services. They use stolen keys to call AWS APIs like GetCallerIdentity, ListFoundationModels, and Converse to check access to Anthropic Claude models and promotional credits. This enables attackers to assess the financial value of credentials for resale in token-jacking markets, where stolen AI model access is sold below retail price. This represents an evolution in credential validation techniques targeting cloud AI services.