anbeime/skill / skills/native-subtitle-quote-image/native-subtitle-quote-image
native-subtitle-quote-image
Verified将本地视频或用户有权处理的在线视频,经过来源获取、文字稿定位、选题选句、精确取帧、紧凑裁切、拼图和逐张质检,制作成 3:4 或保留画面原比例的视频字幕长图。支持两种明确分开的输出:保留画面内已烧录字幕的原生字幕模式,以及把已审核的时间点与台词绘制到真实视频帧上的脚本字幕模式。用户要求原生字幕截图、字幕帧拼图、YouTube 金句长图、台词截图、不重绘字幕、自定义中文台词,或调整主图比例、字幕区域、台词间隔和美感时使用。
scripts/validate_repo.py
View on GitHub#!/usr/bin/env python3
"""Validate repository packaging invariants without external dependencies."""
import hashlib
import json
import re
import sys
from pathlib import Path
ROOT = Path(__file__).resolve().parents[1]
SKILL_DIR = ROOT / "skills" / "native-subtitle-quote-image"
SKILL_FILE = SKILL_DIR / "SKILL.md"
OPENAI_YAML = SKILL_DIR / "agents" / "openai.yaml"
RENDERER = SKILL_DIR / "scripts" / "native_subtitle_stitch.py"
ENV_CHECK = SKILL_DIR / "scripts" / "check_environment.py"
UPDATE_CHECK = SKILL_DIR / "scripts" / "check_update.py"
VERSION_FILE = SKILL_DIR / "VERSION"
README = ROOT / "README.md"
README_EN = ROOT / "README_EN.md"
README_KO = ROOT / "README_KO.md"
PLUGIN = ROOT / ".codex-plugin" / "plugin.json"
EXPECTED_NAME = "native-subtitle-quote-image"
EXPECTED_VERSION = "2.10.3"
def main():
errors = []
required = [
ROOT / "LICENSE",
README,
README_EN,
README_KO,
ROOT / "assets" / "native-subtitle-quote-image-icon.png",
ROOT / "examples" / "gallery-manifest.json",
ROOT / "assets" / "banner-zh.jpg",
ROOT / "assets" / "banner-en.jpg",
ROOT / "assets" / "banner-ko.jpg",
PLUGIN,
SKILL_FILE,
VERSION_FILE,
OPENAI_YAML,
SKILL_DIR / "requirements.txt",
RENDERER,
ENV_CHECK,
UPDATE_CHECK,
SKILL_DIR / "references" / "yt-dlp-and-transcripts.md",
SKILL_DIR / "references" / "end-to-end-workflow.md",
SKILL_DIR / "references" / "visual-style.md",
SKILL_DIR / "references" / "troubleshooting.md",
ROOT / ".github" / "workflows" / "validate.yml",
]
for path in required:
if not path.is_file():
errors.append(f"缺少文件: {path.relative_to(ROOT)}")
try:
plugin = json.loads(PLUGIN.read_text(encoding="utf-8"))
except (OSError, json.JSONDecodeError) as exc:
errors.append(f"plugin.json 无法解析: {exc}")
plugin = {}
if plugin.get("name") != EXPECTED_NAME:
errors.append("plugin.json name 与 Skill 目录不一致")
if plugin.get("version") != EXPECTED_VERSION:
errors.append(
f"plugin.json version 应为 {EXPECTED_VERSION},实际为 {plugin.get('version')}"
)
if plugin.get("skills") != "./skills/":
errors.append("plugin.json skills 必须指向 ./skills/")
if VERSION_FILE.is_file():
skill_version = VERSION_FILE.read_text(encoding="utf-8").strip()
if skill_version != EXPECTED_VERSION:
errors.append(
f"Skill VERSION 应为 {EXPECTED_VERSION},实际为 {skill_version}"
)
skill_text = SKILL_FILE.read_text(encoding="utf-8") if SKILL_FILE.is_file() else ""
frontmatter = re.match(r"\A---\n(.*?)\n---\n", skill_text, flags=re.DOTALL)
if not frontmatter:
errors.append("SKILL.md 缺少有效 YAML frontmatter")
else:
header = frontmatter.group(1)
if not re.search(rf"^name:\s*{re.escape(EXPECTED_NAME)}\s*$", header, re.MULTILINE):
errors.append("SKILL.md name 不正确")
if not re.search(r"^description:\s*\S.+$", header, re.MULTILINE):
errors.append("SKILL.md description 不能为空")
yaml_text = OPENAI_YAML.read_text(encoding="utf-8") if OPENAI_YAML.is_file() else ""
if f"${EXPECTED_NAME}" not in yaml_text:
errors.append("agents/openai.yaml default_prompt 未引用当前 Skill")
for reference in (
"references/yt-dlp-and-transcripts.md",
"references/end-to-end-workflow.md",
"references/visual-style.md",
"references/troubleshooting.md",
):
if reference not in skill_text:
errors.append(f"SKILL.md 未链接参考文件: {reference}")
public_docs = [
README,
README_EN,
README_KO,
ROOT / "examples" / "README.md",
SKILL_FILE,
SKILL_DIR / "references" / "yt-dlp-and-transcripts.md",
SKILL_DIR / "references" / "end-to-end-workflow.md",
SKILL_DIR / "references" / "visual-style.md",
SKILL_DIR / "references" / "troubleshooting.md",
]
for document in public_docs:
text = document.read_text(encoding="utf-8") if document.is_file() else ""
for target in re.findall(r"\[[^\]]+\]\(([^)]+)\)", text):
clean_target = target.split("#", 1)[0]
if not clean_target or clean_target.startswith(
("http://", "https://", "mailto:", "#")
):
continue
linked = (document.parent / clean_target).resolve()
if not linked.exists():
errors.append(
f"{document.relative_to(ROOT)} 链接不存在: {target}"
)
readme_text = README.read_text(encoding="utf-8") if README.is_file() else ""
for readme in (README, README_EN, README_KO):
text = readme.read_text(encoding="utf-8") if readme.is_file() else ""
for source in re.findall(r'<img\s+[^>]*src="([^"]+)"', text):
if source.startswith(("http://", "https://")):
continue
if not (ROOT / source).is_file():
errors.append(f"{readme.name} 图片不存在: {source}")
readme_languages = {README: "zh", README_EN: "en", README_KO: "ko"}
readme_images = {}
for readme, lang in readme_languages.items():
text = readme.read_text(encoding="utf-8") if readme.is_file() else ""
shown = set(re.findall(r"examples/gallery/[^\"')\s]+", text))
readme_images[lang] = shown
for source in sorted(shown):
if not source.startswith(f"examples/gallery/{lang}/"):
errors.append(f"{readme.name} 只能展示 {lang} 字幕案例: {source}")
manifest_path = ROOT / "examples" / "gallery-manifest.json"
try:
manifest = json.loads(manifest_path.read_text(encoding="utf-8"))
except (OSError, json.JSONDecodeError) as exc:
errors.append(f"gallery-manifest.json 无法解析: {exc}")
manifest = {}
manifest_images = {lang: set() for lang in readme_languages.values()}
for example in manifest.get("examples", []):
lang = example.get("language")
manifest_images.setdefault(lang, set()).add(f"examples/{example.get('image')}")
if example.get("script") and not example["script"].startswith(f"scripts/{lang}/"):
errors.append(f"案例脚本不在对应语言目录: {example.get('id')}")
image_path = ROOT / "examples" / str(example.get("image", ""))
if image_path.is_file() and hashlib.sha256(image_path.read_bytes()).hexdigest() != example.get("sha256"):
errors.append(f"案例图片 SHA-256 与 manifest 不一致: {example.get('id')}")
if lang not in readme_languages.values():
errors.append(f"案例语言无效: {example.get('id')}")
if not str(example.get("image", "")).startswith(f"gallery/{lang}/"):
errors.append(f"案例图片不在对应语言目录: {example.get('id')}")
for key in ("image", "script"):
if example.get(key) and not (ROOT / "examples" / example[key]).is_file():
errors.append(f"案例文件不存在: {example[key]}")
for lang, images in manifest_images.items():
if len(images) != 3:
errors.append(f"{lang} 案例应为 3 张,实际 {len(images)} 张")
if readme_images.get(lang, set()) != images:
errors.append(f"{lang} README 展示的案例与 manifest 不一致")
listed = {path for images in manifest_images.values() for path in images}
for image in sorted((ROOT / "examples" / "gallery").rglob("*.jpg")):
relative = image.relative_to(ROOT).as_posix()
if relative not in listed:
errors.append(f"案例图片未登记在 manifest: {relative}")
if "~/.codex/skills" not in readme_text:
errors.append("README 缺少 Codex 默认 Skill 安装目录")
for document in (README, README_EN, README_KO, SKILL_FILE):
text = document.read_text(encoding="utf-8") if document.is_file() else ""
if "render-script" not in text:
errors.append(f"{document.relative_to(ROOT)} 未说明脚本字幕模式")
for script in (RENDERER, ENV_CHECK, UPDATE_CHECK):
if script.is_file():
try:
compile(script.read_text(encoding="utf-8"), str(script), "exec")
except SyntaxError as exc:
errors.append(f"脚本语法错误 {script.relative_to(ROOT)}: {exc}")
public_text_files = [
*public_docs,
OPENAI_YAML,
RENDERER,
ENV_CHECK,
UPDATE_CHECK,
VERSION_FILE,
PLUGIN,
]
for path in public_text_files:
if path.is_file() and "/Users/" in path.read_text(encoding="utf-8"):
errors.append(f"包含本机绝对路径: {path.relative_to(ROOT)}")
if errors:
print("Repository validation failed:", file=sys.stderr)
for error in errors:
print(f"- {error}", file=sys.stderr)
raise SystemExit(1)
print(f"Repository is valid: {EXPECTED_NAME} v{EXPECTED_VERSION}")
if __name__ == "__main__":
main()
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| 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: 197k | Updated | Claude CodeCodex | Verified | Install |
| 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: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| 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: 194k | Updated | Claude CodeCodex | Verified | Install |