Advanced Engine Benchmark: Dark Cultivation & Reality Weaver Sandbox

An advanced, multi-layered sandbox scenario engineered specifically to stress-test high-token text adventure models, context parsers, and background scripting limits. Built with high-fidelity, explicit narrative directives, this scenario challenges AI systems to maintain deep architectural and identity continuity over extended campaign loops. [ENGINE TESTING METRICS] 1. IDENTITY DUALITY CONTINUITY: Tests the model's capacity to track a single protagonist holding dual, asymmetric identity matrices (Tianhan Cheon Ryu / Yeon Mu Hwan) without collapsing into generic or mechanical persona traits. 2. HYBRID WORLD-BUILDING ARCHITECTURE: Evaluates deep-context rendering across a unified Tech-Murim universe where physical, spiritual, and environmental laws are actively modified in real-time. 3. VISCERAL PROSE COMPILATION: Directs the engine to process raw, high-intensity kinetic combat and intricate biological interactions without defaulting to sanitized, repetitive clichés or safety-filter truncation. 4. LORE DATA ANCHORING: Features dense, optimized custom Story Cards (including Faction, Item, and multi-layered Race entries) designed to test the limits of platform context compression and memory allocation. [CREATOR NOTE] This scenario serves as an open benchmarking standard for advanced authors. Use this file to measure model processing thresholds, script execution parameters, and long-term short-term memory cache stability during deep roleplay runs.