Skip to content

ducdg88/ducpt-skills / skills/vietnamese-copy-polish

vietnamese-copy-polish

Verified

Polish Vietnamese marketing copy so it reads like a native writer and fits each platform. Checks missing diacritics, long sentences, filler, banned punctuation (en dash and em dash), weak hooks and missing calls to action, then rewrites for Facebook, TikTok, YouTube, Zalo, email or a website. Use when the user asks to "sửa văn", "viết lại cho hay", "chuốt caption", "kiểm tra bài đăng", "viết content tiếng Việt", or shares Vietnamese copy that feels machine translated.

FileKindSize
SKILL.mdskill md2.7 KB
scripts/check_copy.pyscript3.7 KB

scripts/check_copy.py

View on GitHub
#!/usr/bin/env python3
"""Lint Vietnamese copy: dashes, long sentences, missing diacritics, CTA, length.

Usage:
    python check_copy.py FILE [--platform facebook]

Exit code 1 when a hard error (dash characters) is found. Standard library only.
"""
import argparse
import re
import sys
import unicodedata
from collections import Counter

TARGETS = {
    "facebook": (400, 1500), "tiktok": (80, 300), "youtube-title": (40, 70),
    "youtube-description": (300, 1500), "zalo": (150, 600), "email": (500, 2000), "web": (800, 10 ** 9),
}
DASHES = {chr(0x2013): "en dash", chr(0x2014): "em dash"}
CTA_HINTS = ["comment", "còm", "bình luận", "inbox", "nhắn", "đăng ký", "click", "bấm", "liên hệ",
             "gọi", "tải", "xem thêm", "link", "reply", "trả lời", "để lại", "đặt", "mua", "nhận"]
STOP = {"và", "là", "của", "có", "cho", "một", "những", "các", "được", "không", "này", "với", "thì",
        "mình", "bạn", "đã", "để", "trong", "khi", "người", "the", "a", "to", "of", "and"}
VN_LETTERS = set("ăâđêôơưĂÂĐÊÔƠƯ")


def has_diacritic(word):
    if any(ch in VN_LETTERS for ch in word):
        return True
    decomposed = unicodedata.normalize("NFD", word)
    return any(unicodedata.category(ch) == "Mn" for ch in decomposed)


def lint(text, platform):
    errors, warnings, info = [], [], []
    for ch, name in DASHES.items():
        for m in re.finditer(ch, text):
            line = text.count("\n", 0, m.start()) + 1
            errors.append(f"line {line}: {name} found, replace with comma, colon or full stop")

    sentences = [s.strip() for s in re.split(r"(?<=[.!?…])\s+|\n+", text) if s.strip()]
    for s in sentences:
        n = len(s.split())
        if n > 35:
            warnings.append(f"long sentence ({n} words): {s[:60]}...")

    words = re.findall(r"[^\W\d_]+", text)
    if words:
        plain = [w for w in words if w.isascii() and len(w) > 1]
        ratio = len([w for w in plain if not has_diacritic(w)]) / len(words)
        info.append(f"words without diacritics: {ratio:.0%}")
        if ratio > 0.45:
            warnings.append("many words have no diacritics; text may be missing dấu or be mostly English")
        counts = Counter(w.lower() for w in words if w.lower() not in STOP and len(w) > 2)
        for word, n in counts.most_common(3):
            if n >= 5 and n / len(words) > 0.03:
                warnings.append(f"'{word}' repeated {n} times")

    lower = text.lower()
    if platform not in ("youtube-title",) and not any(h in lower for h in CTA_HINTS):
        warnings.append("no clear call to action found")

    lo, hi = TARGETS[platform]
    length = len(text.strip())
    info.append(f"length: {length} chars (target {lo} to {hi if hi < 10 ** 9 else 'any'})")
    if length < lo or length > hi:
        warnings.append(f"length {length} outside {platform} target")
    return errors, warnings, info


def main(argv=None):
    ap = argparse.ArgumentParser(description="Lint Vietnamese copy.")
    ap.add_argument("file")
    ap.add_argument("--platform", default="facebook", choices=sorted(TARGETS))
    args = ap.parse_args(argv)
    if hasattr(sys.stdout, "reconfigure"):
        sys.stdout.reconfigure(encoding="utf-8")
    try:
        with open(args.file, encoding="utf-8-sig") as fh:
            text = fh.read()
    except OSError as exc:
        sys.exit(f"Cannot read {args.file}: {exc}")
    errors, warnings, info = lint(text, args.platform)
    for line in info:
        print("INFO ", line)
    for line in warnings:
        print("WARN ", line)
    for line in errors:
        print("ERROR", line)
    print(f"\n{len(errors)} errors, {len(warnings)} warnings")
    sys.exit(1 if errors else 0)


if __name__ == "__main__":
    main()
slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack."Stars: 198kUpdated Claude CodeCodexVerifiedInstall
ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt.Stars: 194kUpdated Claude CodeCodexVerifiedInstall