ShopPrompt Injection Shield
Skillv1.0.0

Prompt Injection Shield

Every channel your agent reads is an attack surface. Close them.

Hermes AgentOpenClawpython
$9one-time
Buy through Claw Prime Marketplace: $9

What It Is

Your AI agent reads emails, replies to tweets, processes web forms, and talks to other agents. Every one of those channels is an attack surface.

Prompt injection is no longer theoretical. Attackers embed hidden instructions in normal-looking text, and your agent follows them. Social engineering through legitimate channels can talk an agent into handing over credentials with no embedded code at all.

This is the same protection system running in production at Claw Prime AI. It covers email intake, Twitter and social mentions, web forms, content queues, MCP integrations, and inter-agent messages.

Three response protocols. CRITICAL patterns get blocked and logged. SUSPICIOUS patterns get sanitized — zero-width characters stripped, unicode normalized — and flagged. Clean input passes through inside a trust wrapper.

Defense in depth. Email gets three independent layers: webhook intake sanitization, cron agent pattern matching, and framework-level unicode scanning. Each catches what the others miss.

Regex-based and framework-agnostic. Built for Hermes, drops into any pipeline.

Features

30+ detection patterns across 8 attack categories — instruction overrides, markup injection, prompt extraction, tool poisoning, multi-agent exploits, multimodal injection, MCP data injection, social engineering heuristics
Three response protocols — block critical, sanitize suspicious, wrap clean input in a trust boundary
MCP trust boundary — treats every MCP server response as untrusted external input, defending against Agentjacking
Agent social engineering defense — triggers identity verification before any credential or sensitive action
Defense in depth — 3 independent layers on the email path alone
Integration checklist — exact hooks for content pages, client CRM, lead capture, inbound email, social handlers
Weekly updates — new attack vectors from CVE databases, OWASP reports, and security research land in your copy

What's Included

  • ✓SKILL.md — the full detection and response protocol
  • ✓30+ regex detection patterns across 8 attack categories
  • ✓Sanitization routines (zero-width strip, unicode normalization)
  • ✓Integration checklist with exact wiring points
  • ✓INSTALL.md

Ready to deploy?

Get Prompt Injection Shield and start building today.

Buy through Claw Prime Marketplace: $9