Skip to content
All updates
Weekly AI digest

Securing the agent workforce — and the Mythos 5 standoff

1 min readMultra AI Team

As companies deploy AI agents 'as employees,' a new security layer emerges: NewCore launches to give agents identities — while Anthropic's Mythos 5 stays restricted under export controls.

A quieter week for launches, but a clarifying one for where the agent stack is heading.

Giving AI agents an identity

On June 15, cybersecurity startup NewCore emerged from stealth with $66M to give AI agents identities — authentication, governance, and access control for agents deployed "as employees" at scale. As businesses wire agents into real systems, "who is this agent and what may it touch?" becomes its own product category. Expect identity, audit, and guardrails to follow every serious agent rollout.

Major model news: the Mythos 5 standoff continues

The dominant model-layer story stayed the export-control restriction on Anthropic's frontier Fable 5 and Mythos 5. After access was suspended on June 12, the models remained restricted through the week while Anthropic worked toward a path to restore them (a partial resolution arrived the following week — see the June 28 digest). It's a live preview of how regulation now shapes which top models you can actually build on.

What it means if you're picking tools

If you're deploying agents, plan for governance from day one — and for mission-critical work, prefer broadly-available models with a fallback. See automation and customer support.


Sources: NewCore exits stealth (TechCrunch) · Anthropic on Fable 5 / Mythos 5 access