Osegi
Security·1 min read

Cisco adds Anthropic's Claude Mythos to AI Defense for red teaming

Cisco will use Anthropic's Claude Mythos in AI Defense to plan attacks on customers' own AI apps and agents, with tight controls and no open prompt access.

Cisco Live stage presentation with the Cisco logo beside a graphic of an AI security shield testing software agents.
Image: TNW | Anthropic

Cisco will add Anthropic's Claude Mythos to AI Defense, the company announced at its Cisco Live event in Melbourne. The model will plan attacks against customers' own AI applications and agents, a practice known as red teaming, to find weak spots before hackers do. It will come first to select enterprise customers, with no date set.

Tightly fenced access

Customers will not be able to type prompts into Mythos directly. The model works only on a fixed job, against systems the customer owns and has registered for testing. Every test is tied to a named person, recorded for audit and can be stopped at any time. According to Cisco's blog post, Mythos cannot scan the wider internet.

In July, Cisco researchers used the model to push 15 chatbots past their guardrails. "Enterprises are increasingly building AI agents into their core operations and want to test those agents thoroughly before putting them into production," said Michael Moore, Anthropic's cybersecurity lead.

Security vendors line up

Cisco is not alone. Palo Alto Networks launched always-on testing built on Mythos and OpenAI's cyber model in September, and IBM said on Monday it plans to add Mythos to its Concert Protect tool in early 2027.

The approach suits Anthropic, which puts Mythos inside partners' products rather than distributing the model itself. It released Mythos 5.2 on Monday, still only for vetted organisations. The controls matter: in April, an unauthorised group reached Mythos through a contractor's access.

Splunk updates

Cisco also updated its Splunk tools that track what AI agents cost and grade the quality of their work. A cost tool called Tokenomics enters private preview in December.

Apps worth a look

SUMO: the number puzzle

Tap the right tiles, add them up, hit the target. Easy to learn, hard to master. Free, for all ages.

Tags:ciscoanthropicclaude mythosai defensered teamingai agents

More from this day

See all →

More in: Security

See all →