zhtw-mcp-linguistic-linter

Lint Traditional Chinese text for Taiwan MoE vocabulary, punctuation, and character standards.

7|2|Updated May 16, 2026
One-click install
npx skills add https://github.com/reason-machines/mcp-skills --skill zhtw-mcp-linguistic-linter-reason-machines
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: zhtw-mcp-linguistic-linter
Source: https://github.com/reason-machines/mcp-skills/tree/main/skills/zhtw-mcp-linguistic-linter
Command: npx skills add https://github.com/reason-machines/mcp-skills --skill zhtw-mcp-linguistic-linter-reason-machines

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI-generated and cross-strait-influenced Chinese text often drifts into Mainland Chinese (zh-CN) vocabulary, punctuation, and character shapes, violating Taiwan Ministry of Education standards. This Skill catches that regional drift automatically before content reaches readers. ## Core Features & Use Cases - Cross-strait vocabulary linting: Detects 1100+ Mainland terms (軟件→軟體, 內存→記憶體) and enforces Taiwan-style punctuation and MoE character variants. - Auto-fix and quality gates: Applies lexical-safe corrections in place and rejects documents exceeding a configurable error threshold, suitable for CI/CD pipelines. - AI writing detection: Flags filler phrases, semantic safety words, and passive-voice density in AI-generated Chinese content. - Use Case: A documentation team adds a GitHub Actions step that runs the linter with the strict profile on every pull request touching Markdown files, blocking merges when zh-TW errors are found. ## Quick Start Ask the assistant to check this paragraph for Mainland Chinese terms and Taiwan MoE standard violations, then apply safe auto-fixes.

Frequently Asked Questions about zhtw-mcp-linguistic-linter

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I check Traditional Chinese text for Mainland Chinese terms?▼

Pass the text to the zhtw MCP tool or run zhtw-mcp lint on a file. The linter reports each Mainland term with line, column, and a Taiwan-standard suggestion such as 軟件→軟體 or 內存→記憶體.

How to enforce Taiwan MoE standards in CI/CD pipelines?▼

Install the zhtw-mcp binary in your CI job and run zhtw-mcp lint with --max-errors 0 on changed Markdown files. The command exits with a failure verdict when any zh-TW violation is found, blocking the merge.

Does the linter support auto-fixing Traditional Chinese punctuation?▼

Yes, run with --fix or set fix_mode to lexical_safe in the MCP tool call. It converts half-width punctuation to full-width and replaces Mainland vocabulary, returning corrected text without altering ambiguous terms.

Can I convert Simplified Chinese to Traditional Chinese with it?▼

Yes, the s2t subcommand converts Simplified Chinese files or stdin input to Traditional Chinese, with an optional strict profile. Example: zhtw-mcp s2t input.md -o output.md.

What is the difference between base and strict linting profiles?▼

The base profile covers cross-strait vocabulary, punctuation, casing, grammar, and political terms. The strict profile adds character variant enforcement (裏→裡) and full MoE standard compliance on top of base.

Why does the linter skip code blocks in Markdown files?▼

When content_type is set to markdown, the linter excludes fenced code blocks, inline code spans, HTML blocks, and YAML frontmatter from scanning. This prevents false positives on source code and configuration content.