菜单
Checks words.json and all verb group files for duplicate entries before adding new vocabulary or verbs to CONJI. Use before adding any entry to public/words.json or public/group-N.json.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use when writing actual CONJI book prose (the Romanian HTML chapter files in public/book-contents/) from an approved plot in plots/*.md. Writes one chapter/page at a time by default — never the whole book in one pass unless the user explicitly asks for all chapters at once. Use the book-plot skill first if no plot exists yet.
Use when creating a plot outline for a new CONJI book, or reviewing/updating the plot of an existing one. Plots live in plots/*.md and describe a book's arc, twists, ending, and chapter structure without writing the actual book prose. Use before writing any book content — see the write-book skill for that.
Use when adding a new Romanian word, phrase, or verb to CONJI's data files (public/words.json, public/group-1..4.json). Explains entry format, which file a new entry belongs in, and the required duplicate check.
| name | find-duplicates |
| description | Checks words.json and all verb group files for duplicate entries before adding new vocabulary or verbs to CONJI. Use before adding any entry to public/words.json or public/group-N.json. |
Checks words.json and all verb group files for duplicate entries before you add new vocabulary or verbs.
Run from the project root:
# Full check — words.json + all group files
node .claude/skills/find-duplicates/scripts/find-duplicates.js
# words.json only
node .claude/skills/find-duplicates/scripts/find-duplicates.js --words-only
# Verb groups only
node .claude/skills/find-duplicates/scripts/find-duplicates.js --verbs-only
# Check one specific value before adding it
node .claude/skills/find-duplicates/scripts/find-duplicates.js --check "merg"
node .claude/skills/find-duplicates/scripts/find-duplicates.js --check "a merge"
| Code | Meaning |
|---|---|
0 | No duplicates — safe to add |
1 | Duplicates found — review before proceeding |
Run this before adding any new entry to words.json or any group-N.json. The words-and-verbs skill references this skill as the required pre-check step.
words.json — scans all "value" fields, case-insensitive. Reports any value that appears more than once.
group-1..4.json — scans all "infinitive" fields across all four files. Catches the same verb added to two groups by mistake.
🔍 CONJI Duplicate Checker
────────────────────────────────────────────────────────────
words.json (1148 entries)
────────────────────────────────────────────────────────────
⚠️ "merg" ×2 (entries: 45, 1143)
────────────────────────────────────────────────────────────
Verb groups (group-1..4.json)
────────────────────────────────────────────────────────────
✅ No duplicate infinitives found.
────────────────────────────────────────────────────────────
❌ Duplicates found. Review before adding new entries.
🔍 CONJI Duplicate Checker
────────────────────────────────────────────────────────────
words.json (1148 entries)
────────────────────────────────────────────────────────────
✅ No duplicates found.
────────────────────────────────────────────────────────────
Verb groups (group-1..4.json)
────────────────────────────────────────────────────────────
✅ No duplicate infinitives found.
────────────────────────────────────────────────────────────
✅ All clear.
node .claude/skills/find-duplicates/scripts/find-duplicates.js --check "totuși"
⚠️ "totuși" already exists:
In words.json:
[entry 1129] { "value": "totuși", "translations": ["however", "yet", "still", "nevertheless"] }