LogicGuard is an AI-facing logic-chain skill layer for research, reports, and serious writing. Agents use its subskills to build reusable logic model libraries from papers, books, reports, URLs, and drafts; map claims, evidence, warrants, limits, and rebuttals; audit weak links; and synthesize new paper outlines, technical reports, deck storylines, review responses, or cautious rewrites from existing models without inventing support.