echo-quill-alchemist

Train a model on a novel, then continue writing in that novel's world — with memory of characters, places, and recent events. Load this skill whenever the user asks to train/fine-tune/learn the style of a novel or .txt, mentions Echo-Quill / Alchemist / 炼金 / 双塔, asks to open the training dashboard, asks how the system works, asks to stop or restart the training services, asks to write/continue text (with or without memory), asks to train an additional chapter or continuation they wrote, or asks to roll back to the original-novel state. The skill manages a local FastAPI backend + Vite/React dashboard, dependency installation, .env scaffolding, training, memory-mode continuation, pure-style continuation, and rollback — the user supplies a novel path or a passage, the skill handles everything underneath.