Vibeleaderboard
← Back to Tools
Index / tool

GuardRail

github.com/fvdhmbai/guardrail
Visit github.com
Category
Developer Tools
Rank
No. 1319Tools index
Listed in
#1 Review code for security
Pricing
Freemium
Platform
cli
Type
TOOL
Latest release
v0.3.1
Date

About

GuardRail is an open-source, MIT-licensed set of guards that hook into AI coding agents like Claude Code to block dangerous shell commands and file writes (e.g. rm -rf on protected paths, destructive SQL, force pushes, secret leaks) before they execute, logging every block. A paid Pro tier adds PII output scanning and EU AI Act compliance reporting for teams.

What it can do

  • Block destructive shell commands (e.g. rm -rf on protected paths, force pushes) before execution

    Shell command from AI coding agent → Blocked/allowed command decision

  • Block destructive SQL statements before execution

    SQL command from AI coding agent → Blocked/allowed command decision

  • Block file writes that leak secrets

    File write operation from AI coding agent → Blocked/allowed write decision

  • Log every blocked command or action

    Blocked command/event → Log entry

  • Scan AI output for personally identifiable information (Pro tier)

    AI-generated output text → PII scan results

Why it made the leaderboard

Stops an AI coding agent from executing a destructive command like a force-push to main or an unscoped DELETE before it happens, with a self-test proving the guards actually work.

Intel on GuardRail

More in Intel

Tags

ai-agentssecurityguardrailsclaude-codepre-executiondevsecopscliopen-source

Tech Stack

Node.js

Comments (0)

No comments yet

Editorially curated, with community endorsements as a secondary signal. Corrections welcome.