New Claude Security Tool · Automate What Academy

New Claude Security Tool

Anthropic is launching a research preview of Claude Code Security, an AI tool designed to think like a human researcher to find and fix deep-seated vulnerabilities that traditional scanners usually miss.

Claude Opus 4.6 recently uncovered over 500 vulnerabilities in open-source code that had been hidden for decades.
The tool moves beyond simple rule-based scanning by reasoning through business logic and data flows just like a pro security analyst.
Every finding goes through a multi-stage verification process to filter out false positives and assign a severity rating.
It suggests targeted software patches for human review, meaning we stay in control while the AI does the heavy lifting.
This shifts the advantage to defenders, allowing teams to patch weaknesses before attackers can use AI to exploit them.

Do you guys think AI-driven security will eventually make traditional manual code audits obsolete?

1 comment