Skip to main content
Glama
Prompt Guard Injection Sanitizer Skill

Prompt Guard Injection Sanitizer Skill

by Alpha-Park

Prompt Guard Injection Sanitizer Skill for Autonomous Agents

MCP Compatible Python 3.9+ License: MIT Zero External Dependencies

Enterprise-grade defensive input firewall for LLM autonomous agents. Inspects, filters, and neutralizes prompt injections, indirect jailbreaks, and instructions override exploits before passing untrusted text into prompts.

Key Features

  • Exploit & Jailbreak Detection: Flags attempts to override system prompts or bypass ethical boundary instructions.

  • Neutralization: Replaces hostile command sequences with safe sanitized placeholders.

  • Zero Overhead: Blazing fast regex and heuristic matching using the Python Standard Library.

Related MCP server: Prompt Rejector MCP

MCP Server Integration

python mcp_server.py

Available MCP Tools

  • sanitize_prompt: Analyze and cleanse text of malicious injection attacks.

  • inspect_threats: Scan text and list identified security threat patterns.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      B
      maintenance
      Protects AI agents from prompt injection attacks, jailbreak attempts, and common web vulnerabilities by screening untrusted input through semantic LLM analysis and static pattern matching.
      71 npm
      2
      ISC
    • A
      license
      Not graded
      quality
      B
      maintenance
      Enables MCP clients and autonomous agents to screen incoming prompts for system overrides, delimiter breakouts, and encoded payload evasion, helping block prompt injection and jailbreak attempts before execution.
      8
      MIT
    • A
      license
      Not graded
      quality
      B
      maintenance
      Enables users to scan untrusted text, files, and directory trees for prompt-injection patterns and receive severity-ranked, explainable verdicts before content reaches an agent. It runs locally and returns clean, review, or blocked verdicts for safer agent input.
      1
      MIT