Aguara

About

Aguara is an MCP server designed to empower AI agents with robust security scanning capabilities. In a rapidly evolving landscape where autonomous AI agents discover, install, and execute third-party code without prior security review, Aguara fills a critical gap. It acts as a dedicated security advisor, enabling agents to proactively scan for prompt injection, credential theft, remote code execution, data exfiltration, and supply chain attacks, ensuring informed decisions before integrating new tools or configurations into their environment. All scans run locally, deterministically, and without network access, completing in milliseconds.

Key Features

• Provides structured reports with severity-rated findings, matched patterns, and line numbers.
• Operates locally and deterministically, requiring no network access or cloud dependencies.
• Comprehensive threat coverage with 138 rules across 15 categories (e.g., Prompt injection, Credential leak, Exfiltration).
• Scans text content (skill descriptions, tool definitions) for security threats.
• 0 GitHub stars
• Analyzes MCP server configurations for dangerous patterns and insecure settings.

Use Cases

• AI agents evaluating whether to install a new skill or MCP server from a registry.
• Pre-installation security review of skill descriptions and tool definitions by AI agents.
• Auditing MCP server configurations for potential vulnerabilities and insecure settings.