severity1

severity1 / open-guard-engine

Public

Agent-agnostic security engine for AI coding assistants

11
0
100% credibility
Found Feb 04, 2026 at 10 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Go
AI Summary

Open Guard is a security scanner that protects AI coding assistants by checking prompts and commands for malicious injections, harmful actions, and unsafe content through multiple detection layers.

How It Works

1
🛡️ Worry about AI safety

You love using AI helpers for coding but hear stories of sneaky bad prompts that could harm your projects.

2
📥 Find and grab the protector

You discover Open Guard, a free tool that watches over your AI chats to keep things safe.

3
✨ Set it up super easy

Download the app, make it ready with a quick step, and it's good to go on your computer.

4
⚙️ Pick your safety level

Choose if you want quick checks, ask-me checks, or just notes – it fits how careful you want to be.

5
🔍 Check your AI request

Paste what you want to ask the AI coding helper, and run the protector to scan it.

6
✅ See the clear verdict

Instantly learn if it's safe to send or needs to be stopped, with a simple reason why.

🎉 Code safely forever

Now your projects stay protected while the AI helps you build faster and smarter.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 10 to 11 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is open-guard-engine?

Open-guard-engine is an agent-agnostic security engine for AI coding assistants, scanning prompts and tool inputs for prompt injection, malicious commands, and harmful content. Written in Go, it delivers a standalone binary that processes stdin text and outputs JSON decisions like allow or block, integrating seamlessly with tools via hooks. Developers pipe inputs through `open-guard analyze` for fast threat detection across layers: regex patterns, agent analysis via Claude or Ollama, and LLM safety checks.

Why is it gaining traction?

It stands out with 75-100% detection on known and novel attacks, zero false positives on safe prompts, and configurable modes (strict, confirm, permissive) that fit real workflows. The open guard engine supports local Ollama for privacy or Claude for accuracy, making it flexible without vendor lock-in. Easy binary installs and a demo GIF show immediate value for securing AI assistants without slowing down coding.

Who should use this?

Security-conscious devs using Claude Code, Cursor, or similar AI coding assistants that execute Bash, Read/Write, or other tools. Teams handling untrusted repos or enterprise codebases prone to prompt injection via generated commands. Anyone piping AI outputs to shell needs this guard before execution.

Verdict

Early-stage with 10 stars and 1.0% credibility score, but solid docs, tests, and binary releases make it worth trying for pioneers. Grab the release if you're securing AI tools today—scale up as it matures.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.