Anthropic launches multi-agent security scanner for Claude Code

Anthropic released the Claude Security plugin for Claude Code in beta, which runs a multi-agent vulnerability scan of a repository from inside an existing Claude Code session. It turns selected findings into patch files that developers review and apply themselves.

Why it matters: This extends Anthropic's coding agent from writing code to auditing it, folding a task that used to require a separate AppSec vendor into the same terminal session developers already use. It fits a broader pattern of coding agents absorbing adjacent workflows like testing and security review rather than staying narrowly scoped to code generation.