Morning Edition · Sunday, August 23, 2026Published at 2:19 AM EDT · New York

Anthropic Runs Claude Security Scans on Mythos 5 and Adds a $35 Million Credit Fund for Open-Source Defenders

Enterprise customers get repository scanning with findings classified by weakness type, severity and confidence, delivered through a scan interface rather than direct access to the model.

Anthropic moved Claude Security scanning onto Mythos 5, its most capable model, in public beta for Claude Enterprise customers. An administrator enables the product in the console, a user selects a repository, and the scan returns findings…