Anthropic's Mythos Bug-Finding AI Is Being Built Into Cisco and IBM Security Tools

Anthropic has begun extending access to Mythos, its bug-finding AI model, by allowing Cisco and IBM to build it into their security products. Cisco announced its integration at an event in Melbourne. The red teaming tools in its AI Defence suite, which test AI models, agents and the applications built on them for weaknesses, will be able to call Mythos to probe customers' setups. Cisco says the tool is now "super-charged by Mythos". Mythos will be another option alongside the AI-powered red teaming the product already uses, and it will be available under existing licences at no extra cost.

Hours later, IBM said it will add Mythos to its Concert tool, which uses AI to scan applications for risk and weaknesses. IBM says it will use the model to uncover weaknesses traditional scanners miss, determine which pose real risk, and generate fixes for developer approval.

Customers get the model's functionality without being given the model itself. This matters because, as The Register notes, Mythos could help attackers exploit software flaws if it fell into the wrong hands. Anthropic first offered it to vetted organisations through Project Glasswing and has since widened access through other controlled programs. Anthropic's Natalie Mead said there is a long queue of buyers. She did not say whether similar deals exist with other companies, and The Register has asked Anthropic whether these integrations signal broader access.