claude-security Beta: Cost, Output, Accuracy — 189 Agents, 2 Hours, Measured

A measured first-hand report on Anthropic's beta claude-security plugin for Claude Code. What it costs in wall-clock time, what the output actually looks like, how far you can trust it, and what to do with the findings — with the raw numbers from the run.

Read Original

Related

Dev.to tutorial 27m ago

The Skill Audit

After 12 issues of careful work, the prompt library audit reveals more living in memory than anyone expected. Here's what got found, converted, and deprecated.