Anthropic has integrated its advanced Claude Mythos 5 model into Claude Security, enabling enterprise teams to perform vulnerability scanning without directly accessing the model.
What is Claude Security?
As of August 21, 2026, Claude Security scans utilize the Mythos-class model, previously available only to vetted defenders through Project Glasswing.
How It Works
- Connects to a GitHub repository
- Traces data flows across files
- Provides findings categorized by Common Weakness Enumeration (CWE)
- Includes confidence and severity ratings
- Suggests patches with adversarial verification to minimize false positives
Key Features
- No separate model add-on required
- Billed as standard token usage under existing enterprise plan
- Patching conducted separately through Claude Code
- Findings can be exported via Slack, Jira, CSV, or Markdown
Defender Advantage Fund
Anthropic also launched the Defender Advantage Fund (0xDAF), providing 5 million in Claude credits aimed at organizations focused on securing open-source software.
Source: Anthropic, August 2026