Release of the Claude Security plugin for multi-agent vulnerability scanning
57s
2
13:30, 23.07.2026
Anthropic launched Claude Security for multi-agent scanning of the codebase right from the local session. Currently, this plugin is only available via paid plans.
Overview of the new release
Claude Security functions by analyzing the architecture, and after building the threat model, all the revealed vulnerabilities are divided among various agents for scanning. Only after proper final verification will all the needed details be included in the final report.
The new plugin can work with a specific commit, a changed branch, a certain component, or an entire repository. In the final report, users get access to the security level, exploitation scenario, confidence score, and some recommendations.
For the determined vulnerabilities, the plugin can generate patches. They are additionally verified and created in a separate copy of the repository. To activate the patch, the user should apply it manually.
After plugin installation, scanning is launched with the claude-security command. For the proper functioning, there is a necessity for Dynamic Workflows support, Python 3.9.6+, and Claude Code 2.1.154+.
Hope you found this article helpful - what do you think? Like and subscribe to our blog to get more practical tips and the latest tech news from HostZealot.
Was this article helpful to you?
Yes
No
VPS popular offers
-12%
Linux
€ 8.33 /mo
Billed annually
-10%
Windows
€ 29.99 /mo
Billed annually
-10%
Linux
€ 21.37 /mo
Billed annually
-10%
Windows
€ 18.65 /mo
Billed annually
-7.2%
Linux
€ 6.89 /mo
Billed annually
-10%
Linux
€ 17.43 /mo
Billed annually
-18.4%
Windows
€ 24 /mo
Billed annually
-10%
Linux
€ 134.99 /mo
Billed annually
-10%
Linux
€ 12.54 /mo
Billed annually
-10%
Linux
€ 9.32 /mo
Billed annually
