360 Secure SkillHub
360 安全技能中心

simplified-to-traditional

该技能将 ai-thoughts/docs/ 中的简体文章（<slug>-zh-hans.md）转为繁体双胞胎（<slug>-zh-hant.md），以简为源、用 OpenCC s2twp 转换，并自动注入双语交叉链接。脚本 sync_tra.py 支持 --check 校验两版同步；需审核一简对多繁歧义字及引文用字，必要时在 OVERRIDES 中固定修正。仅按用户点名文章、经批准后运行，不自动执行、不回溯旧文。

j3ffyang
j3ffyang
热度 23.79
最近更新 2026-10-11 22:00:28
分类：效率依赖：代码执行 / 读写文件应用场景：翻译 / 文件转换 / 工作流

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add simplified-to-traditional
Metadata
name
simplified-to-traditional
description
Convert a Simplified-Chinese article in ai-thoughts/docs/ into its Traditional-Chinese twin, and keep the two in sync. The Simplified article is the source of truth and lives as <YYMMDD-slug>-zh-hans.md; the Traditional twin is generated to <YYMMDD-slug>-zh-hant.md with OpenCC (s2twp) plus a permanent cross-link. Use when the user asks to make, sync, or update the Traditional (繁體) version of a Chinese article — "convert to traditional", "做繁體版", "同步繁简", "zh-hant", "繁體". Never runs automatically; acts only on an explicitly named article. Future articles only — do NOT backfill existing files unless the user explicitly orders it.

Simplified → Traditional (zh-hans → zh-hant)

Convert one Simplified-Chinese article into Traditional Chinese, and keep the pair in sync.

Approval gate — read before anything else

  • Never convert or write without explicit user approval. This skill converts on demand only; it does not run automatically and does not pick articles on its own.
  • Wait for the user to name the article. If they don't, do nothing but ask which one.
  • Before writing, confirm the plan: the source file, the exact output path, and whether an existing -zh-hant.md should be overwritten. Never overwrite silently.
  • Scope: future articles only. In this repo a bare docs/<YYMMDD-slug>.md is the English article and the legacy -chn.md files are the existing Simplified-Chinese articles — neither is a source here unless the user explicitly orders it.

Conventions

  • Filenames. Simplified article = docs/<YYMMDD-slug>-zh-hans.md; Traditional twin = docs/<YYMMDD-slug>-zh-hant.md. (In ai-thoughts a bare .md is English and the legacy -chn.md files are out of scope.)
  • Source of truth = the Simplified file. The Traditional file is derived from it.
  • Traditional variant. OpenCC s2twp (Taiwan, phrase-aware) — the best default for these literary articles.
  • Cross-link. Exactly one line, directly under the H1 and above the infographic; Chinese label; a permanent absolute URL on the j3ffyang account:
    • In the Simplified file: > 繁体版：[<simplified title>](https://github.com/j3ffyang/ai-thoughts/blob/main/docs/<slug>-zh-hant.md)
    • In the Traditional file: > 簡體版：[<traditional title>](https://github.com/j3ffyang/ai-thoughts/blob/main/docs/<slug>-zh-hans.md)
    • The label is written in the file's own script (Simplified file → 繁体版; Traditional file → 簡體版) so neither file mixes scripts.
  • Infographic stays Simplified and is shared by both files (same ../imgs/... path). The README lists the Simplified article only.

Procedure

  1. Wait for the user to name the article, then locate docs/<YYMMDD-slug>-zh-hans.md (in this repo the source must be -zh-hans.md — a bare .md is English).
  2. Confirm the plan: source path, output path (-zh-hant.md), and overwrite permission if the output already exists.
  3. If an existing -zh-hant.md is present, stop and ask before overwriting — unless the user asked to update it.
  4. Run the sync script: 
    python3 .opencode/skills/simplified-to-traditional/scripts/sync_tra.py docs/<slug>-zh-hans.md
    It (a) injects/updates the Simplified file's 繁體版 cross-link, (b) converts the body with opencc -c s2twp, (c) injects the Traditional file's 简体版 cross-link, and (d) writes docs/<slug>-zh-hant.md.
  5. Review the generated Traditional against the Simplified, and clear the script's lint. OpenCC can pick the wrong form for a 一简对多繁 character (see "OpenCC pitfalls" below) and can rewrite characters inside a quote. sync_tra.py prints a ⚠ … ambiguous form(s) to verify list after every run — check each against the cited edition (or the intended meaning) and pin any genuine fix in OVERRIDES. Surface anything else suspicious to the user; do not silently hand-fix — a hand-edit is drift and will fail the next --check.
  6. Verify the pair is in sync: 
    python3 .opencode/skills/simplified-to-traditional/scripts/sync_tra.py docs/<slug>-zh-hans.md --check
    Exit 0 = in sync; exit 1 lists what is stale.
  7. Report: source, output, and any review notes.

Keeping the two in sync

  • After editing the Simplified article, re-run the sync to refresh its Traditional twin.
  • --check regenerates the Traditional in memory and diffs it against the file (ignoring the cross-link lines); non-zero exit means out of sync.
  • Do not treat the Traditional file as an editable primary. Edit the Simplified and re-sync. If a term is genuinely mis-converted by OpenCC, tell the user and decide together, then pin it as a (from, to) pair in scripts/sync_tra.py → OVERRIDES — it is applied after OpenCC on every re-sync, so it survives (a hand-edit would not).

OpenCC pitfalls (一简对多繁)

opencc -c s2twp is deterministic but context-blind: when one Simplified character maps to several Traditional ones, it can pick the wrong one for a rare, literary, or technical term. Two failure modes:

  • Ambiguous characters. e.g. 云（说）→ 雲（cloud）; 干 → 乾（dry）/ 幹（do）. Rare terms err most. sync_tra.py prints a lint of every risky form（雲 幹 髮 複 …）after each run — verify each against the cited edition and pin a fix in OVERRIDES if wrong.
  • Register normalization. s2twp may rewrite characters its Taiwan dictionary prefers（e.g. 哄→鬨, 背→揹）. Acceptable in running prose, but wrong inside a verbatim quote taken from another edition.

Quotes must match the cited edition

Never delegate quoted primary text to a converter. For every verbatim quote, compare its Traditional characters with the edition named in the article's 引用来源; where s2twp differs, pin the (from, to) pair in OVERRIDES. Quoted text is otherwise frozen — do not change its characters or punctuation.

Pinning fixes (OVERRIDES)

OVERRIDES in sync_tra.py is applied to the converted Traditional text after OpenCC, in order, on every re-sync. Add a pair only when OpenCC is genuinely wrong for this corpus; keep the list small and specific. Never hand-edit -zh-hant.md — a hand-edit is drift and fails --check.

Verification checklist

  • Traditional file exists at docs/<slug>-zh-hant.md.
  • Both files carry the correct cross-link line directly under the H1 (Chinese label, j3ffyang URL, pointing at the other file).
  • sync_tra.py <source> --check exits 0.
  • The script's ambiguous-form lint was reviewed; genuine miscodings pinned in OVERRIDES (never hand-edited), and quotes verified against the cited edition.
  • The infographic reference points at the shared (Simplified) image; no broken ../imgs/.
  • The Simplified file's body is otherwise unchanged (only the cross-link line added).
  • Nothing committed (commit only on explicit approval).

Error handling

  • Source not found: list docs/*-zh-hans.md and ask which to use; do not substitute a bare (English) or -chn.md file.
  • Output already exists: stop and ask before overwriting.
  • opencc missing or wrong dict: report and stop; do not guess a conversion.
  • Ambiguous source name (neither -zh-hans.md nor an explicitly ordered -chn.md): ask.