What problem does it solve? Legal questions often get shallow or hallucinated AI answers because models skip fact-gathering and cite unverifiable statutes. This Skill enforces a lawyer-style workflow—facts first, then law—with mandatory retrieval grounding so every cited statute and case traces back to a searchable source. ## Core Features & Use Cases - Five-Stage Diagnosis: User profiling, scenario-based fact mining across 8 legal domains (contracts, torts, family, labor, criminal, administrative, corporate, IP), then retrieval, tiered strategy, and actionable output. - Dual Modes with Smart Routing: Quick mode for fast directional answers (RAG-only, capped results) and expert mode with multi-round query reformulation, Tier A/B/C layered reading, and structured citation lists. - Anti-Hallucination Safeguards: Mandatory source tracing, statute validity status labeling, cross-consistency checks, and a 38-item pre-output verification checklist. - Use Case: A user asks "my landlord won't return my deposit"—the Skill classifies it as a contract dispute, asks targeted follow-up questions, searches subscribed legal knowledge bases via RAG (or web search as fallback), and outputs a structured opinion with verified statute citations and next-step actions. ## Quick Start Describe your legal problem in plain language, such as "my employer fired me without cause, can I claim compensation?", and the Skill will guide you through fact confirmation and deliver a sourced legal analysis.
