mjha-arya/DOC-foundry: Adversarial multi-agent detection engineering

DOC-foundry is an adversarial multi-agent detection engineering framework that turns threats into validated Sigma detection rules. It uses four LLM role agents to propose rules while employing a deterministic referee and proof engine to evaluate false positives, test adversarial mutations, and compute coverage. The pipeline produces reproducible rule bundles accompanied by a structured review history and feedback trail. Open article →