This research demonstrates how Large Language Models can extract formal specifications from software documents, significantly reducing the manual effort required for automated security testing.

• Transforms natural language documents into formal specifications with high accuracy and reduced human labor
• Enables more efficient automated testing workflows for security validation
• Allows for easier maintenance of specifications during system updates
• Bridges the gap between human-readable documentation and machine-verifiable test requirements

This innovation matters for security teams by enabling more comprehensive testing coverage with fewer resources, ultimately helping organizations detect vulnerabilities earlier in the development lifecycle while maintaining rigorous specification standards.

Extracting Formal Specifications from Documents Using LLMs for Automated Testing