Precise Bilingual Proofreading for Professional Translators
This skill provides a structured, multi-layered audit for translated text, tailored specifically for game localization, fiction, technical documentation, and academic papers. It solves the problem of "silent" translation errors—like terminology drift, AI hallucinations, or gender mismatches—that simple spellcheckers miss.
Advanced Review Strategies
- Monte Carlo Mode: Rapidly assesses large files (millions of words) by identifying high-risk regions and iterating until findings converge, providing a statistical heatmap of quality.
- Split Review Mode: A meticulous line-by-line workflow for critical content, featuring context-aware chunking and direct edit application.
- Smart Terminology Filter: Eliminates false positives by prioritizing long-string glossary matches over substring fragments.
Why This Beats Manual Prompting
Unlike a generic "review this text" prompt, this skill executes a rigorous developer-grade pipeline. It performs widths-based bloat detection, regex-based AI contamination checks, and cross-references character gender maps. The output isn't just advice; it's a structured fix_proposal.md with directly substitutable corrections designed for automated integration into your workflow.
Supported Formats & Tools
Works with any text-based translation format (TXT, TSV, JSON) and supports custom glossaries for strict terminology adherence.