Search papers, labs, and topics across Lattice.
RuleForge is an automated system that generates web vulnerability detection rules from structured Nuclei templates, addressing the challenge of keeping pace with the high volume of newly disclosed CVEs. The system employs a novel LLM-as-a-judge confidence validation system, evaluating rules for both sensitivity and specificity. This validation approach achieves an AUROC of 0.75 and reduces false positives by 67% compared to synthetic-test-only validation in production, demonstrating the effectiveness of LLMs in cybersecurity tasks when combined with systematic feedback and human-in-the-loop validation.
LLMs can automatically generate web vulnerability detection rules with surprisingly high accuracy, but only with careful validation and human oversight to mitigate overconfidence.
Security teams face a challenge: the volume of newly disclosed Common Vulnerabilities and Exposures (CVEs) far exceeds the capacity to manually develop detection mechanisms. In 2025, the National Vulnerability Database published over 48,000 new vulnerabilities, motivating the need for automation. We present RuleForge, an AWS internal system that automatically generates detection rules--JSON-based patterns that identify malicious HTTP requests exploiting specific vulnerabilities--from structured Nuclei templates describing CVE details. Nuclei templates provide standardized, YAML-based vulnerability descriptions that serve as the structured input for our rule generation process. This paper focuses on RuleForge's architecture and operational deployment for CVE-related threat detection, with particular emphasis on our novel LLM-as-a-judge (Large Language Model as judge) confidence validation system and systematic feedback integration mechanism. This validation approach evaluates candidate rules across two dimensions--sensitivity (avoiding false negatives) and specificity (avoiding false positives)--achieving AUROC of 0.75 and reducing false positives by 67% compared to synthetic-test-only validation in production. Our 5x5 generation strategy (five parallel candidates with up to five refinement attempts each) combined with continuous feedback loops enables systematic quality improvement. We also present extensions enabling rule generation from unstructured data sources and demonstrate a proof-of-concept agentic workflow for multi-event-type detection. Our lessons learned highlight critical considerations for applying LLMs to cybersecurity tasks, including overconfidence mitigation and the importance of domain expertise in both prompt design and quality review of generated rules through human-in-the-loop validation.
Amazon Science