SleuthCo

Security proxy for AI agents. Scans every message for prompt injection, PII, and secrets. Defense-in-depth: Go proxy + iptables firewall + eBPF kernel monitor. YAML policy engine, audit logging, 5 AI agents with RAG knowledge bases.

18
2
100% credibility
Found Mar 02, 2026 at 18 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Go
AI Summary

ClawShield is a security proxy that sits between users and AI agents, scanning messages for prompt injections, leaked personal info, and secrets before they reach the AI model.

How It Works

1
🔍 Discover Safe AI Helpers

You hear about ClawShield, a shield that lets you chat with smart AI agents without worries about sneaky tricks or leaked info.

2
📥 Grab and Start It

Download the ready-to-go package and launch everything with a single easy command — no hassle.

3
🔗 Link Your AI Brain

Paste in a quick connection code from your AI service so the agents can think and respond.

4
🚀 Open Your Secure Chat Hub

A friendly dashboard pops up in your browser with five ready-to-chat AI experts for coding, security, cloud, comms, and research.

5
💬 Chat Safely with Agents

Pick an agent like the coding whiz or security guru, ask questions, and get helpful answers — all scanned for safety behind the scenes.

6
📊 Peek at Security Logs

Check the built-in logs anytime to see what was blocked or cleaned up, keeping everything private.

Your Protected AI Team is Live

Enjoy worry-free super-smart helpers that stay secure, never leaking secrets or falling for tricks.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 18 to 18 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 clawshield-public?

Clawshield-public is an open source security proxy server built in Go that sits between users and AI agents like OpenClaw, scanning every message for prompt injection, PII leaks, secrets, and vulnerabilities before they hit the model. It delivers defense-in-depth with a core proxy, optional iptables firewall for egress control, and eBPF kernel monitoring for behavioral anomalies, all configurable via a YAML policy engine. Users get a dashboard, Prometheus metrics, SQLite audit logging with a CLI viewer, and a one-command Docker quickstart that spins up five ready-to-use AI agents with RAG knowledge bases.

Why is it gaining traction?

It stands out as a security proxy meaning tailored for AI agents, combining github security scanning-like checks (prompt injection, secrets) with network/kernel layers that adapt via a cross-layer event bus—block a suspicious IP in iptables if eBPF spots port scanning. Developers love the YAML github security policy for fine-grained control (allow/block/redact per tool), full audit trails for compliance, and seamless integration with Anthropic, OpenAI, or local models like LM Studio. Low-latency scanning and built-in agents make it a drop-in upgrade over raw gateways.

Who should use this?

Security engineers securing AI agent deployments in production, especially teams using OpenClaw or similar multi-agent gateways exposed to untrusted inputs. DevOps handling github security projects or actions pipelines with LLM routing will appreciate the proxy security check and iptables/eBPF for zero-trust egress. AI builders prototyping with agents need its audit logging and policy enforcement to catch leaks early.

Verdict

Try it for AI agent security—solid YAML policies and layered defenses address real gaps in open LLM gateways, despite 18 stars and 1.0% credibility signaling early maturity. Docs are crisp with Docker/prod deploys, but expect to tweak tests for custom scanners; pair with github security advisories for full coverage.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.