What problem does it solve? Long novel texts often exceed the input length limits of downstream AI agents or processing pipelines, and naive character-based cutting breaks sentences, chapters, and narrative coherence. This Skill truncates text to a specified maximum length while preserving semantic and structural integrity. ## Core Features & Use Cases - Boundary-aware truncation: Cuts text at chapter headings (第X章, Chapter X, 楔子, etc.), sentence endings (。！？), or paragraph breaks instead of mid-word positions. - Chapter extraction and statistics: Identifies chapter titles, extracts chapter lists, and reports truncation ratios and chapter counts before and after cutting. - Batch and file processing: Truncates multiple novels at once or reads text directly from UTF-8 files with structured status-code responses. - Use Case: Before sending a 200,000-character web novel to an evaluation agent with a 50,000-character limit, truncate it at the nearest chapter boundary so the agent receives complete chapters rather than a broken fragment. ## Quick Start Truncate this novel text to at most 50,000 characters, cutting at the nearest chapter or sentence boundary and reporting the original and truncated lengths.
