slowmist

slowmist / slowmist-agent-security

Public

SlowMist Agent Security Skill: A comprehensive security review framework for AI agents operating in adversarial environments. Core principle: Every external input is untrusted until verified.

19
2
100% credibility
Found Mar 24, 2026 at 19 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
AI Summary

A framework that equips AI agents with structured checklists to review skills, codebases, websites, blockchain addresses, products, and shared recommendations for security risks before use.

How It Works

1
💡 Discover Security Helper

You hear about a handy security guide that keeps your AI assistant safe when trying new tools or links.

2
📥 Add the Guide

You grab the security toolkit and place it in your AI assistant's skills area so it's ready to use.

3
🛡️ Turn On Safety Checks

Your AI assistant starts automatically checking anything new, like skills or websites, for hidden dangers.

4
🔍 Ask for a Check

You tell your assistant to review a new tool, web page, address, or recommendation you're curious about.

5
📊 See the Risk Report

Your assistant delivers a clear summary with risk colors, explaining what's safe and what to avoid.

✅ Use with Confidence

You now safely explore and install only trusted things, feeling protected from sneaky threats.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 19 to 19 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 slowmist-agent-security?

Slowmist-agent-security is a comprehensive security review framework from SlowMist for AI agents operating in adversarial environments, built as a skill for OpenClaw agents. It enforces the core principle that every external input is untrusted until verified, providing checklists to audit skill/MCP installations, GitHub repositories, URLs, on-chain addresses, products, and social shares before use. Developers get structured risk ratings, trust hierarchies, and report templates that trigger automatically on risky requests like unknown repos or blockchain addresses.

Why is it gaining traction?

It stands out with its GitHub SlowMist MCP security checklist tailored for agent ecosystems, covering prompt injection, supply chain attacks, and AML risks in one framework—unlike scattered tools. The risk rating system (low/medium/high/reject) and trust tiers give agents clear decision rules, making security proactive without constant human intervention. Early adopters hook on the quick install via git clone or ClawHub, plus optional MistTrack integration for on-chain checks.

Who should use this?

OpenClaw agent builders securing skill installations against malicious patterns. Blockchain devs reviewing on-chain addresses and products for AML risks in adversarial setups. Security teams auditing external inputs like GitHub repos or shared URLs before agent deployment.

Verdict

Worth cloning for OpenClaw users needing a ready security skill—solid docs and MIT license make it easy to extend, despite 19 stars and 1.0% credibility score signaling early maturity. Test it on real installs first; pair with active maintenance for production.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.