MoltShield Scanner

Free AI agent security scanner. Part of the MoltCops defense ecosystem.

Quick Start

# Install dependencies
npm install

# Run development server
npm run dev

# Open http://localhost:3000

Deploy to Vercel

Or manually:

npm install -g vercel
vercel

Features

• 🔍 Free Scanner — 20-rule security engine, runs client-side
• 🛡️ Badge Application — Founding Operative NFT application flow
• 🎫 Waitlist — $MCOP token waitlist with referral mechanics

Structure

scanner-app/
├── src/
│   ├── app/
│   │   ├── layout.tsx      # Root layout + metadata
│   │   ├── page.tsx        # Main page (tabs)
│   │   └── globals.css     # Global styles
│   ├── components/
│   │   ├── Scanner.tsx     # MoltShield scanner
│   │   ├── BadgeApplication.tsx  # Founding Operative form
│   │   └── Waitlist.tsx    # Token waitlist + referral
│   └── lib/
│       └── scan-engine.ts  # 20-rule detection engine
├── package.json
└── next.config.js

GitHub Action

Add MoltCops to your CI pipeline — scans every PR for agent vulnerabilities:

# .github/workflows/moltcops-scan.yml
name: MoltCops Security Scan
on:
  pull_request:
    paths: ['**/*.ts', '**/*.js', '**/*.py', '**/*.md']

permissions:
  contents: read
  pull-requests: write

jobs:
  scan:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: Adamthompson33/moltshield@main
        with:
          github-token: ${{ secrets.GITHUB_TOKEN }}
          # api-key: ${{ secrets.MOLTCOPS_PRO_KEY }}  # Optional: Pro tier
          # fail-on: CRITICAL  # Optional: block PRs with critical findings

Free tier: 10 rules — catches drains, sleepers, prompt injection, code injection, exfil. Pro tier: 20 rules — adds jailbreak, encoding tricks, context poisoning, sandbox escape, and more.

Scanner Rules

Free Tier (10 rules)

• CRITICAL — Key export, drain patterns, unlimited approvals, sleeper triggers
• HIGH — Prompt injection, safety removal
• MEDIUM — Data exfil, stealth operations, code injection
• LOW — Hardcoded addresses

Pro Tier (+10 rules)

• CRITICAL — MAX_UINT256 detection
• HIGH — Identity spoofing, authority bypass, jailbreaks, encoding tricks, false authority
• MEDIUM — Context poisoning, sandbox escape, time/count triggers, config exposure

Pro tier available via API key ($5/month).

Links

• MoltCops Docs
• GitHub
• Twitter
• Litepaper

---

To Protect and Serve (Humanity) 🚨