Show HN: Guardrails for Claude Code: blocks rm -RF, reports what loaded

Category: devtools

Tags: claude-code, ai-safety, plugin, developer-tools, security

Score: 6.8/10 (Innovation: 6, Technical: 7, Documentation: 8, Utility: 6)

Guardrails is a Claude Code plugin that enforces hard limits on destructive commands and edits, reports what instruction files and plugins load at session start, and lints CLAUDE.md files against best practices. It provides a granular, configurable security layer for AI-assisted coding, addressing a growing need for safe agentic development.

Target audience: backend devs, DevOps, AI engineers using Claude Code

Repository: https://github.com/ahmed-alstaty/guardrails-plugin · Python · MIT · 2 stars

View on Hacker News