AI news ·
Cisco Adds Claude Mythos to Controlled AI Defense Testing
Cisco tested an AI attack-path tool in a controlled defense setting, with Splunk tracking costs and quality to keep the process governed and auditable.
Cisco has integrated the Claude Mythos system into a controlled defense testing environment. The system generates contextual attack paths that operate exclusively within registered targets, using authenticated operators and bounded sessions. Each session includes stop controls to maintain operational limits.
Splunk observability is used to track both cost and output quality during these exercises. The combined setup provides a model for governed offensive-security automation, where testing parameters and spending remain visible and auditable throughout the process.