Compliant LLM

Acerca de

Compliant LLM empowers developers to assess the security and compliance of their AI assistant systems. By testing against prevalent attack vectors such as prompt injection, jailbreaking, and adversarial inputs, it delivers a comprehensive security assessment. The tool facilitates the creation of secure and compliant AI systems, aligning with industry standards through robust testing and detailed reporting.

Características Principales

• Tests agents against top 10 OWASP threats with 8+ attack strategies and 100+ attack types.
• 126 GitHub stars
• Offers an interactive CLI with rich output and a visual dashboard for result analysis.
• Supports advanced configuration via YAML.
• Supports multiple LLM providers via LiteLLM.
• Provides detailed reporting and analysis.

Casos de Uso

• Red-teaming AI assistants to identify vulnerabilities.
• Ensuring compliance with security standards and regulations.
• Improving the robustness of AI systems against adversarial attacks.