Skip to content

NocodeMrLi/mr-li-writer-skill

mr-li-writer

Verified

面向中文内容创作的编辑判断、研究、策划、写作与平台原生交付工作流。根据标题、素材、参考资料、核心想法或模糊灵感,先识别读者任务、材料条件和创新必要性,再按需完成语境扫描、可写性判断与方向校准,并区分发布平台与内容目标,进行事实检索、观点设计、平台原生正文写作、标题策划和对应交付。适用于公众号文章、知乎回答、小红书原生笔记、官网/网页、个人博客、行业解读、实用指南和基于素材的扩展研究写作。

FileKindSize
SKILL.mdexecutableskill md62.0 KB
agents/openai.yamlexecutablescript12.9 KB
assets/gzh-design/LICENSEexecutablescript34.2 KB
assets/gzh-design/NOTICE.mdmarkdown568 B
assets/gzh-design/assets/preview-template.htmlexecutablescript3.8 KB
assets/gzh-design/references/common-components.mdexecutablescript14.3 KB
assets/gzh-design/references/eval-cases.mdexecutablescript5.7 KB
assets/gzh-design/references/format-normalize.mdexecutablescript4.4 KB
assets/gzh-design/references/theme-generator.mdexecutablescript25.3 KB
assets/gzh-design/references/theme-graphite-minimal.mdexecutablescript34.1 KB
assets/gzh-design/references/theme-index.mdexecutablescript3.1 KB
assets/gzh-design/references/theme-moyu-green.mdexecutablescript46.1 KB
assets/gzh-design/references/theme-moyu-ticket.mdexecutablescript24.7 KB
assets/gzh-design/references/theme-olive-journal.mdexecutablescript56.2 KB
assets/gzh-design/references/theme-red-white.mdexecutablescript29.9 KB
assets/gzh-design/references/theme-zen-whitespace.mdexecutablescript23.5 KB
assets/mr-li-writer-promo-poster.pngbinary179.4 KB
assets/mr-li-writer-promo-silent.mp4binary1.2 MB
assets/readme-banner.svgother5.3 KB
assets/readme-showcase/moyu-green-content.jpgbinary281.1 KB
assets/readme-showcase/moyu-green-cover.jpgbinary183.8 KB
assets/readme-showcase/moyu-green-list.jpgbinary203.4 KB
assets/readme-showcase/moyu-green-long.jpgbinary431.9 KB
assets/readme-showcase/olive-journal-content.jpgbinary177.7 KB
assets/readme-showcase/olive-journal-cover.jpgbinary214.2 KB
assets/readme-showcase/olive-journal-ending.jpgbinary154.1 KB
assets/readme-showcase/olive-journal-long.jpgbinary352.3 KB
assets/readme-showcase/red-white-cover.jpgbinary183.1 KB
assets/readme-showcase/red-white-long.jpgbinary405.3 KB
assets/readme-showcase/red-white-section.jpgbinary227.4 KB
assets/readme-showcase/red-white-table.jpgbinary219.4 KB
references/commercial-source-terms.txtmarkdown389 B
references/delivery-protocol.mdmarkdown11.7 KB
references/editorial-evaluation-protocol.mdmarkdown3.1 KB
references/editorial-routing-protocol.mdmarkdown6.2 KB
references/genre-structure-protocol.mdmarkdown7.5 KB
references/high-impact-writing-protocol.mdmarkdown6.2 KB
references/humanize-rules.mdexecutablescript11.5 KB
references/ideation-protocol.mdmarkdown6.7 KB
references/lint-checks.mdmarkdown3.9 KB
references/opening-design-protocol.mdmarkdown9.2 KB
references/platform-native-protocol.mdmarkdown19.3 KB
references/research-protocol.mdexecutablescript13.9 KB
references/title-rules.mdexecutablescript9.2 KB
references/writing-rules.mdexecutablescript18.4 KB
scripts/build_gzh_html.pyexecutablescript70.3 KB
scripts/build_html.pyexecutablescript31.8 KB
scripts/component_lint.pyexecutablescript8.1 KB
scripts/extract_seed.pyexecutablescript5.8 KB
scripts/gzh_component_inventory.pyexecutablescript1.8 KB
scripts/lint_article.pyexecutablescript19.5 KB
scripts/scan_sensitive.pyscript2.0 KB
scripts/source_governance.pyscript4.9 KB
scripts/validate_delivery_bundle.pyexecutablescript18.3 KB
scripts/validate_gzh_html.pyexecutablescript14.4 KB
scripts/validate_research_scope.pyexecutablescript10.0 KB
scripts/validate_task_intake.pyscript27.3 KB
scripts/wrap_gzh_preview.pyexecutablescript3.7 KB

scripts/build_gzh_html.py

View on GitHub
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""Build WeChat Official Account HTML snippets from Markdown.

The 6 built-in styles are adapted from gzh-design-skill (AGPL-3.0):
Moyu Green, Red & White, Graphite Minimal, Zen Whitespace, Moyu Ticket,
and Olive Journal.

The full gzh-design component libraries are vendored under
assets/gzh-design/references/. This script is a deterministic CLI renderer and
quick preview helper; agent-driven final layout should read the selected theme
library and assemble richer article-specific components from that source.
"""

import argparse
import hashlib
import html
import os
import random
import re
import subprocess
import sys

try:
    from source_governance import compile_named_source_regex, has_commercial_source_exposure, source_role_names
except ImportError:
    from scripts.source_governance import compile_named_source_regex, has_commercial_source_exposure, source_role_names


PROCESS_LEAK_PATTERNS = (
    r"(抓取|爬取|采集)(到|自|于|时间|结果|数据|页面|信息)?",
    r"(检索|搜索|查询)(结果|显示|发现|到|出来)",
    r"(我|我们)?(通过|使用|借助).{0,12}(搜索引擎|爬虫|脚本|工具|模型|AI|大模型|提示词|prompt)",
    r"(由|通过).{0,12}(AI|模型|大模型|系统).{0,12}(生成|整理|撰写|输出)",
    r"(资料|数据|信息)(抓取|爬取|采集|清洗|抽取|汇总|整理)(时间|结果|口径)?",
    r"(本次|这次).{0,8}(整理|检索|搜索|抓取|采集).{0,12}(发现|得到|结果)",
    r"(截至|截止)\s*\d{4}\s*年\s*\d{1,2}\s*月\s*(抓取|爬取|采集|检索)",
    r"\d{4}\s*年\s*\d{1,2}\s*月\s*(抓取|爬取|采集|检索)",
)

def configured_commercial_source_regex():
    path = os.path.join(os.path.dirname(os.path.dirname(os.path.abspath(__file__))), "references", "commercial-source-terms.txt")
    try:
        with open(path, encoding="utf-8") as handle:
            terms = [line.strip() for line in handle if line.strip() and not line.lstrip().startswith("#")]
    except FileNotFoundError:
        print("[警告] 商业来源词表缺失: %s;将只使用通用商业来源模式。" % path, file=sys.stderr)
        terms = []
    if not terms:
        return None
    return compile_named_source_regex(terms)


def warn_process_leaks(text, label="正文"):
    for pattern in PROCESS_LEAK_PATTERNS:
        if re.search(pattern, text, re.I):
            print(
                "[警告] %s疑似暴露内容生产/资料处理过程;请改为“信息截至 YYYY-MM-DD”“据官网当前页面”“公开资料显示”等读者口径,避免抓取/爬取/采集/检索结果/AI 生成/提示词等词。"
                % label,
                file=sys.stderr,
            )
            return


def warn_commercial_source_exposure(text, label="正文", task_state=""):
    configured = configured_commercial_source_regex()
    _restricted_names, official_names = source_role_names(task_state)
    if has_commercial_source_exposure(text, configured, official_names):
        print(
            "[警告] %s疑似把商业相关第三方机构写成资料背书;请使用与该事实匹配的官方来源。第三方仅作辅助核对时应匿名写“据相关第三方机构公开信息/公开汇总”;未获官方确认时还须明确待核实。" % label,
            file=sys.stderr,
        )


THEMES = {
    "moyu-green": {
        "name": "摸鱼绿",
        "desc": "教程、测评、清单、工具盘点",
        "main": "#059669",
        "deep": "#065F46",
        "light": "#ECFDF5",
        "lighter": "#D1FAE5",
        "mark": "#A7F3D0",
        "title": "#111827",
        "text": "#374151",
        "muted": "#9CA3AF",
        "line": "#D1D5DB",
        "surface": "#FFFFFF",
        "radius": "12px",
        "shadow": "0 8px 24px -12px rgba(5,150,105,0.38)",
        "mode": "rich-card",
    },
    "red-white": {
        "name": "红白色系",
        "desc": "深度分析、观点、力量感话题",
        "main": "#DC2626",
        "deep": "#991B1B",
        "light": "#FEF2F2",
        "lighter": "#FEE2E2",
        "mark": "#FECACA",
        "title": "#1C1917",
        "text": "#374151",
        "muted": "#9CA3AF",
        "line": "#E5E7EB",
        "surface": "#FFFFFF",
        "radius": "12px",
        "shadow": "0 4px 24px -4px rgba(220,38,38,0.15)",
        "mode": "editorial",
    },
    "graphite-minimal": {
        "name": "石墨极简风",
        "desc": "设计、科技评论、专业观点、高端品牌",
        "main": "#52525B",
        "deep": "#27272A",
        "light": "#F4F4F5",
        "lighter": "#E4E4E7",
        "mark": "#A1A1AA",
        "title": "#18181B",
        "text": "#3F3F46",
        "muted": "#A1A1AA",
        "line": "#E4E4E7",
        "surface": "#FFFFFF",
        "radius": "4px",
        "shadow": "none",
        "mode": "minimal",
    },
    "zen-whitespace": {
        "name": "留白禅意风",
        "desc": "禅意冥想、极简生活、深度随笔、艺术留白",
        "main": "#4A5D52",
        "deep": "#3D5046",
        "light": "#EEF3F0",
        "lighter": "#F7FAF8",
        "mark": "#B5C8BC",
        "title": "#2B2B2B",
        "text": "#525252",
        "muted": "#A3A3A3",
        "line": "#E8E8E8",
        "surface": "#FFFFFF",
        "radius": "0",
        "shadow": "none",
        "mode": "zen",
    },
    "moyu-ticket": {
        "name": "摸鱼票据风",
        "desc": "工具对比、创意评测、票据/门票视觉隐喻",
        "main": "#059669",
        "deep": "#064E3B",
        "light": "#ECFDF5",
        "lighter": "#D1FAE5",
        "mark": "#A7F3D0",
        "title": "#111827",
        "text": "#374151",
        "muted": "#6B7280",
        "line": "#10B981",
        "surface": "#FFFCF3",
        "radius": "3px",
        "shadow": "6px 6px 0 rgba(5,150,105,0.18)",
        "mode": "ticket",
    },
    "olive-journal": {
        "name": "橄榄手记",
        "desc": "内刊手记、深度评测、案例复盘、系统性说明",
        "main": "#ED7B2F",
        "deep": "#1E1F23",
        "light": "#FFF4E8",
        "lighter": "#F8E7D2",
        "mark": "#ED7B2F",
        "title": "#1E1F23",
        "text": "#3F3F38",
        "muted": "#8C877B",
        "line": "#D8D1C4",
        "surface": "#FBF8EF",
        "radius": "10px",
        "shadow": "0 8px 24px -16px rgba(30,31,35,0.35)",
        "mode": "journal",
    },
}

SPECIAL_THEMES = {"auto", "random"}
SCRIPT_DIR = os.path.dirname(os.path.abspath(__file__))
ROOT_DIR = os.path.abspath(os.path.join(SCRIPT_DIR, ".."))
GZH_REF_DIR = os.path.join(ROOT_DIR, "assets", "gzh-design", "references")
COMPONENT_HEADING_RE = re.compile(r"^##\s+组件\s+\d+\s+(.+)$", re.M)
COMPONENT_BLOCK_RE = re.compile(r"^##\s+组件\s+(\d+)\s+(.+?)(?=^##\s+组件\s+\d+\s+|^##\s+完整文章模板骨架|^##\s+多行代码块|^---\s*$|\Z)", re.M | re.S)
HTML_CODE_RE = re.compile(r"```html\s*\n(.*?)```", re.S)
PLACEHOLDER_RE = re.compile(r"{{([^{}]+)}}")


def leaf(text):
    return '<span leaf="">%s</span>' % html.escape(normalize_cn_punctuation(text))


def code_leaf(text):
    """Escape code without rewriting punctuation and preserve every column space."""
    escaped = html.escape(text.replace("\t", "    "))
    return '<span leaf="">%s</span>' % escaped.replace(" ", "&nbsp;")


def normalize_cn_punctuation(text):
    table = str.maketrans({
        ",": ",",
        ";": ";",
        "!": "!",
        "?": "?",
        ":": ":",
        "(": "(",
        ")": ")",
    })
    text = CJK_ASCII_PUNCT.sub(lambda m: m.group(0).translate(table), text)
    text = text.replace('"', "”").replace("'", "’")
    return text


CJK_ASCII_PUNCT = re.compile(r"(?<=[一-鿿㐀-䶿])[,;!?:()]|[,;!?:()](?=[一-鿿㐀-䶿])")


def split_table_row(line):
    text = line.strip()
    if text.startswith("|"):
        text = text[1:]
    if text.endswith("|"):
        text = text[:-1]
    return [cell.strip() for cell in text.split("|")]


def is_table_separator(line):
    cells = split_table_row(line)
    return bool(cells) and all(re.fullmatch(r":?-{3,}:?", cell) for cell in cells)


def starts_table(lines, index):
    return (
        index + 1 < len(lines)
        and "|" in lines[index]
        and is_table_separator(lines[index + 1].strip())
        and len(split_table_row(lines[index])) == len(split_table_row(lines[index + 1]))
    )


def inline_html(text, theme):
    token_re = re.compile(r"(\*\*.+?\*\*|==.+?==|\+\+.+?\+\+|<u>.+?</u>|`.+?`)")
    out = []
    pos = 0
    for match in token_re.finditer(text):
        if match.start() > pos:
            out.append(leaf(text[pos:match.start()]))
        token = match.group(0)
        if token.startswith("**"):
            out.append('<strong style="color:%s;font-weight:800;">%s</strong>' % (theme["deep"], leaf(token[2:-2])))
        elif token.startswith("=="):
            out.append('<span style="border-bottom:2px solid %s;font-weight:600;">%s</span>' % (theme["mark"], leaf(token[2:-2])))
        elif token.startswith("++"):
            out.append('<span style="border-bottom:2px solid %s;font-weight:600;">%s</span>' % (theme["mark"], leaf(token[2:-2])))
        elif token.startswith("<u>"):
            out.append('<span style="border-bottom:2px solid %s;font-weight:600;">%s</span>' % (theme["mark"], leaf(token[3:-4])))
        elif token.startswith("`"):
            out.append(
                '<span style="background:#F1F5F9;color:%s;padding:1px 6px;border-radius:4px;font-family:Consolas,Monaco,monospace;font-size:14px;">%s</span>'
                % (theme["main"], leaf(token[1:-1]))
            )
        pos = match.end()
    if pos < len(text):
        out.append(leaf(text[pos:]))
    return "".join(out)


def parse_blocks(md_text):
    lines = md_text.replace("\r\n", "\n").split("\n")
    blocks = []
    i = 0
    while i < len(lines):
        line = lines[i]
        stripped = line.strip()
        if not stripped:
            i += 1
            continue

        if stripped.startswith("```"):
            lang = stripped.strip("`").strip() or "code"
            i += 1
            code = []
            while i < len(lines) and not lines[i].strip().startswith("```"):
                code.append(lines[i])
                i += 1
            i += 1
            blocks.append(("code", lang, "\n".join(code)))
            continue

        m_img = re.match(r"!\[([^\]]*)\]\(([^)]+)\)", stripped)
        if m_img:
            blocks.append(("image", m_img.group(1).strip(), m_img.group(2).strip()))
            i += 1
            continue

        m_head = re.match(r"^(#{1,6})\s+(.+)$", stripped)
        if m_head:
            blocks.append(("heading", len(m_head.group(1)), m_head.group(2).strip()))
            i += 1
            continue

        if starts_table(lines, i):
            headers = split_table_row(stripped)
            i += 2
            rows = []
            while i < len(lines) and lines[i].strip() and "|" in lines[i]:
                row = split_table_row(lines[i])
                if len(row) < len(headers):
                    row.extend([""] * (len(headers) - len(row)))
                rows.append(row[:len(headers)])
                i += 1
            blocks.append(("table", headers, rows))
            continue

        if stripped.startswith(">"):
            quote = []
            while i < len(lines) and lines[i].strip().startswith(">"):
                quote.append(lines[i].strip().lstrip(">").strip())
                i += 1
            blocks.append(("quote", " ".join(quote)))
            continue

        if re.match(r"^[-*]\s+", stripped):
            items = []
            while i < len(lines) and re.match(r"^[-*]\s+", lines[i].strip()):
                items.append(re.sub(r"^[-*]\s+", "", lines[i].strip()))
                i += 1
            blocks.append(("list", items))
            continue

        if re.match(r"^\d+[.、)]\s*", stripped):
            items = []
            while i < len(lines) and re.match(r"^\d+[.、)]\s*", lines[i].strip()):
                items.append(re.sub(r"^\d+[.、)]\s*", "", lines[i].strip()))
                i += 1
            blocks.append(("olist", items))
            continue

        if re.match(r"^(-{3,}|\*{3,}|={3,})$", stripped):
            blocks.append(("hr",))
            i += 1
            continue

        para = [stripped]
        i += 1
        while i < len(lines) and lines[i].strip() and not re.match(
            r"^(#{1,6}\s|[-*]\s|\d+[.、)]\s*|>|```|!\[|---|\*\*\*)",
            lines[i].strip(),
        ):
            if starts_table(lines, i):
                break
            para.append(lines[i].strip())
            i += 1
        blocks.append(("paragraph", " ".join(para)))
    return blocks


def split_title(text, max_len=12):
    clean = re.sub(r"\s+", "", text or "")
    if len(clean) <= max_len:
        return clean, ""
    target = min(max_len, max(6, len(clean) // 2))
    candidates = []
    for match in re.finditer(r"[,。!?;、:::??!!]", clean):
        pos = match.end()
        if 5 <= pos <= len(clean) - 5:
            candidates.append((abs(pos - target), pos))
    for match in re.finditer(r"(不是|而是|在|由|靠|和|与|但|所以|因为|答案)", clean):
        pos = match.start()
        if 5 <= pos <= len(clean) - 5:
            candidates.append((abs(pos - target) + 1, pos))
    if candidates:
        cut = min(candidates)[1]
    else:
        cut = target
        if len(clean) - cut < 5:
            cut = max(5, len(clean) - 5)
    return clean[:cut], clean[cut:]


def split_colored_title(text, max_len=13):
    """Split a two-tone title on a semantic boundary without hanging punctuation."""
    clean = re.sub(r"\s+", "", text or "")
    for marker in (":", ":"):
        if marker not in clean:
            continue
        left, right = clean.split(marker, 1)
        if len(left) >= 5 and len(right) >= 5:
            return left.rstrip(",、;:: "), right.lstrip(",、;:: ")

    line1, line2 = split_title(clean, max_len)
    if not line2:
        return line1, line2
    return line1.rstrip(",、;:: "), line2.lstrip(",、;:: ")


def short_text(text, limit=34):
    clean = re.sub(r"\s+", " ", text or "").strip()
    if len(clean) <= limit:
        return clean
    return clean[:limit].rstrip(",。;、 ") + "…"


def connected_title(text):
    raw = str(text or "").strip()
    parts = re.split(r"\s*[||]\s*", raw, maxsplit=1)
    if len(parts) != 2:
        return raw

    left, right = (part.strip() for part in parts)
    if len(right) > 8:
        right_line_1, right_line_2 = split_title(right, 6)
        if right_line_2:
            right = right_line_1 + "\u200b" + right_line_2
    return "%s · %s" % (left, right)


def emphasis_parts(text, limit=34):
    clean = short_text(text, limit)
    if not clean:
        return "", "", ""
    patterns = (
        r"真正[^,。!?;]{1,6}",
        r"更[^,。!?;]{1,6}",
        r"关键[^,。!?;]{0,5}",
        r"核心[^,。!?;]{0,5}",
        r"不是[^,。!?;]{0,5}",
        r"而是[^,。!?;]{0,5}",
    )
    match = next((found for pattern in patterns if (found := re.search(pattern, clean))), None)
    if match is None:
        keyword_len = min(6, len(clean))
        return "", clean[:keyword_len], clean[keyword_len:]
    return clean[:match.start()], match.group(0), clean[match.end():]


PUBLIC_FRONT_LABELS = {
    "实用指南",
    "行动清单",
    "信息指南",
    "时间提醒",
    "最新消息",
    "考试动态",
    "政策动态",
    "行业观察",
    "前沿观察",
    "深度解读",
    "判断参考",
    "选择参考",
    "避坑提醒",
    "案例复盘",
    "经验总结",
    "科技观察",
    "产品思考",
    "生活观察",
    "关系思考",
    "问题拆解",
    "方法参考",
}

FORBIDDEN_FRONT_LABELS = {
    "中立",
    "客观中立",
    "模型生成",
    "AI生成",
    "AI 生成",
    "提示词",
    "Prompt",
    "公众号排版",
    "深度文章",
    "内部标签",
    "信息祛魅",
    "信息去魅",
}


def safe_front_label(label, fallback="判断参考"):
    clean = re.sub(r"\s+", "", str(label or "")).strip("#::,,。 ")
    safe_fallback = fallback if fallback in PUBLIC_FRONT_LABELS else "判断参考"
    if not clean or clean not in PUBLIC_FRONT_LABELS:
        return safe_fallback
    if any(word.lower() in clean.lower() for word in FORBIDDEN_FRONT_LABELS):
        return safe_fallback
    return clean


def public_topic_tags(title, blocks):
    """Return reader-facing topic labels, never production metadata."""
    text = "%s\n%s" % (title, "\n".join(block[2] if block[0] == "heading" else str(block[1]) for block in blocks if len(block) > 1))
    rules = (
        (("最新", "刚刚", "公布", "发布", "消息", "动态"), ("最新消息", "信息指南")),
        (("含金量", "证书", "职称资格", "软考"), ("深度解读", "判断参考")),
        (("教程", "安装", "步骤", "指南", "怎么", "如何"), ("实用指南", "行动清单")),
        (("测评", "对比", "选择", "避坑"), ("选择参考", "避坑提醒")),
        (("复盘", "案例", "实战"), ("案例复盘", "经验总结")),
        (("科技", "AI", "产品", "设计"), ("科技观察", "产品思考")),
        (("报名", "考试", "时间"), ("考试动态", "时间提醒")),
        (("政策", "通知", "公告"), ("政策动态", "信息指南")),
        (("关系", "情感", "生活", "成长"), ("生活观察", "关系思考")),
    )
    for keywords, labels in rules:
        if any(keyword.lower() in text.lower() for keyword in keywords):
            return tuple(safe_front_label(label) for label in labels)
    return "问题拆解", "方法参考"


def split_display_lines(text, max_len=15, min_last=5, max_lines=3):
    clean = re.sub(r"\s+", " ", normalize_cn_punctuation(str(text or ""))).strip(" 「」")
    if len(clean) <= max_len:
        return [clean] if clean else []

    lines = []
    rest = clean
    break_chars = ",。!?;、:: =×+-//·"
    while len(rest) > max_len and len(lines) < max_lines - 1:
        window = rest[: max_len + 1]
        candidates = []
        for idx, char in enumerate(window):
            pos = idx + 1
            if char in break_chars and 5 <= pos <= len(rest) - min_last:
                candidates.append((abs(pos - max_len), pos))
        for match in re.finditer(r"(不是|而是|在|由|靠|和|与|但|所以|因为|答案|制度|标准)", window):
            pos = match.start()
            if 5 <= pos <= len(rest) - min_last:
                candidates.append((abs(pos - max_len) + 1, pos))
        if candidates:
            cut = min(candidates)[1]
        else:
            cut = max_len
            if len(rest) - cut < min_last:
                cut = max(5, len(rest) - min_last)
        lines.append(rest[:cut].strip())
        rest = rest[cut:].strip()

    if rest:
        lines.append(rest)

    if len(lines) >= 2 and len(lines[-1]) < min_last:
        tail = lines.pop()
        prev = lines.pop()
        merged = prev + tail
        if len(merged) <= max_len + 5:
            lines.append(merged)
        else:
            split_at = max(5, len(merged) - min_last)
            lines.extend([merged[:split_at], merged[split_at:]])
    return [line for line in lines if line]


def balanced_leaf_lines(text, max_len=15):
    return "<br>".join('<span leaf="">%s</span>' % html_text(line) for line in split_display_lines(text, max_len=max_len))


def visible_text(fragment):
    text = re.sub(r"<br\s*/?>", "\n", str(fragment or ""), flags=re.I)
    text = re.sub(r"<[^>]+>", "", text)
    return html.unescape(re.sub(r"\s+", "", text)).strip()


def repair_centered_text_orphans(section_html):
    """Avoid long centered emphasis blocks leaving 1-2 CJK chars on mobile."""
    tag_re = re.compile(
        r"(<p\b(?=[^>]*text-align\s*:\s*center)[^>]*>)(.*?)(</p>)",
        re.I | re.S,
    )

    def repl(match):
        open_tag, inner, close_tag = match.groups()
        clean = visible_text(inner)
        if len(clean) < 18:
            return match.group(0)
        if re.search(r"<(img|table|tr|td|th)\b", inner, re.I):
            return match.group(0)
        fixed = re.sub(r"text-align\s*:\s*center", "text-align:left", open_tag, flags=re.I)
        if "data-balanced=" not in fixed:
            fixed = fixed.replace("<p", '<p data-balanced="center-text"', 1)
        if "word-break" not in fixed:
            fixed = re.sub(r'style="([^"]*)"', r'style="\1;word-break:break-word;overflow-wrap:anywhere;"', fixed, count=1)
        return fixed + inner + close_tag

    return tag_re.sub(repl, section_html)


def is_prose_quote(text):
    clean = re.sub(r"\s+", "", str(text or "")).strip()
    if len(clean) >= 42:
        return True
    sentence_marks = len(re.findall(r"[。!?;]", clean))
    return sentence_marks >= 2 and len(clean) >= 28


def pick_first_paragraph(blocks):
    for block in blocks:
        if block[0] == "paragraph":
            return block[1]
        if block[0] == "quote":
            return block[1]
    return "把一个想法整理成更清楚、更适合发布的文章。"


def pick_last_paragraph(blocks):
    for block in reversed(blocks):
        if block[0] == "paragraph":
            return block[1]
    return "真正重要的不是多写一点,而是把值得说的话说清楚。"


def html_text(text):
    text = str(text or "")
    text = re.sub(r"\*\*(.+?)\*\*", r"\1", text)
    text = re.sub(r"==(.+?)==", r"\1", text)
    text = re.sub(r"\+\+(.+?)\+\+", r"\1", text)
    text = re.sub(r"<u>(.+?)</u>", r"\1", text)
    text = re.sub(r"~~(.+?)~~", r"\1", text)
    text = re.sub(r"`(.+?)`", r"\1", text)
    return html.escape(normalize_cn_punctuation(text))


def load_theme_components(theme_key):
    path = os.path.join(GZH_REF_DIR, "theme-%s.md" % theme_key)
    if not os.path.isfile(path):
        return {}
    with open(path, encoding="utf-8", errors="replace") as source:
        text = source.read()
    components = {}
    for match in COMPONENT_BLOCK_RE.finditer(text):
        number = int(match.group(1))
        title = match.group(2).strip()
        codes = [code.strip() for code in HTML_CODE_RE.findall(match.group(0))]
        if codes:
            components[number] = {"title": title, "codes": codes}
    return components


def select_code(components, number, variant=0):
    item = components.get(number)
    if not item or not item["codes"]:
        return ""
    codes = item["codes"]
    if variant < 0:
        variant = 0
    return codes[min(variant, len(codes) - 1)]


def strip_optional_old_title(snippet, replacements):
    value = replacements.get("旧标题占位") or replacements.get("划线旧认知")
    if value:
        return snippet
    return re.sub(r'\n\s*<p[^>]*>\s*<span leaf="">\{\{(?:旧标题占位|划线旧认知)\}\}</span>\s*</p>', "", snippet)


def apply_component(snippet, replacements):
    snippet = strip_optional_old_title(snippet, replacements)

    def repl(match):
        key = match.group(1).strip()
        before = snippet[:match.start()]
        in_tag = before.rfind("<") > before.rfind(">")
        in_leaf = "leaf" in before[max(0, len(before) - 48):]
        if key in replacements:
            value = html_text(replacements[key])
            if value and not in_tag and not in_leaf:
                return '<span leaf="">%s</span>' % value
            return value
        if "作者名" in key:
            return "Mr.Li Writer"
        if "一句话简介" in key:
            return "让想法变成文章"
        if "图片URL" in key or "个人名片" in key:
            return ""
        if "标签" in key:
            return "写作"
        if "标题" in key:
            return html_text(replacements.get("标题", replacements.get("主标题", "文章标题")))
        if "正文" in key or "内容" in key or "段落" in key or "说明" in key:
            return html_text(replacements.get("正文内容", replacements.get("内容", "")))
        if "编号" in key or key == "01":
            return html_text(replacements.get("编号", "01"))
        return html_text(replacements.get(key, ""))

    rendered = PLACEHOLDER_RE.sub(repl, snippet)
    rendered = re.sub(r"\s*<span\b[^>]*>\s*<span leaf=\"\"></span>\s*</span>", "", rendered)
    rendered = re.sub(r"\s*<span leaf=\"\"></span>", "", rendered)
    rendered = re.sub(r"\n\s*<p\b[^>]*>\s*</p>", "", rendered, flags=re.S)
    return re.sub(r"\n?\s*<!--.*?-->\s*", "\n", rendered, flags=re.S)


def zen_hero(title, blocks):
    intro = pick_first_paragraph(blocks)
    title_alignment = (
        "text-align:left;text-align-last:left"
        if len(re.sub(r"\s+", "", title or "")) >= 18
        else "text-align:center;text-align-last:center"
    )
    return """
<section style="margin:32px 16px 48px;padding:40px 24px;border-top:1px solid #E8E8E8;border-bottom:1px solid #E8E8E8;text-align:center;">
  <p style="font-family:'Noto Serif SC',Georgia,'Times New Roman',serif;font-size:19px;font-weight:600;color:#2B2B2B;margin:0 0 24px;line-height:1.85;letter-spacing:0.8px;%s;word-spacing:normal;white-space:normal;word-break:normal;overflow-wrap:anywhere;">
    <span leaf="">%s</span>
  </p>
  <p style="font-size:12px;color:#A3A3A3;margin:0;letter-spacing:1.5px;">
    <span leaf="">%s</span>
  </p>
</section>""" % (title_alignment, html_text(title), html_text("— " + short_text(intro, 22)))


def zen_section_title(text, number):
    label = "∞ · POSTSCRIPT" if re.search(r"(结语|总结|最后|尾声|写在最后)", text) else "%02d · CHAPTER" % number
    return """
<section style="margin-top:64px;margin-bottom:32px;padding:0 16px;">
  <p style="font-size:10px;color:#4A5D52;font-weight:600;letter-spacing:4px;margin:0 0 10px;text-transform:uppercase;">
    <span leaf="">%s</span>
  </p>
  <h3 style="font-family:'Noto Serif SC',Georgia,'Times New Roman',serif;font-size:22px;font-weight:700;color:#2B2B2B;margin:0 0 16px;letter-spacing:0.5px;line-height:1.4;">
    <span leaf="">%s</span>
  </h3>
  <section style="width:40px;height:2px;background:#4A5D52;"><span leaf=""><br></span></section>
</section>""" % (html_text(label), html_text(text))


def zen_paragraph(text):
    return """
<p style="margin-bottom:26px;font-size:15px;line-height:1.9;text-align:justify;color:#525252;padding:0 16px;">
  <span leaf="">%s</span>
</p>""" % html_text(text)


def zen_quote(text):
    return """
<section style="margin:40px 16px;padding:36px 20px;border-top:1px solid #E8E8E8;border-bottom:1px solid #E8E8E8;text-align:center;">
  <p style="font-family:'Noto Serif SC',Georgia,'Times New Roman',serif;font-size:17px;font-weight:600;color:#2B2B2B;margin:0;line-height:1.9;letter-spacing:0.8px;">
    <span leaf="">「%s」</span>
  </p>
</section>""" % html_text(text.strip("「」"))


def zen_list(items, ordered=False):
    rows = []
    for idx, item in enumerate(items, 1):
        label = "%02d" % idx if ordered else "•"
        rows.append(
            '<section style="display:flex;align-items:baseline;padding:12px 0;border-bottom:1px solid #E8E8E8;">'
            '<p style="font-size:11px;color:#4A5D52;font-weight:600;letter-spacing:1px;margin:0;min-width:28px;"><span leaf="">%s</span></p>'
            '<p style="font-size:14px;color:#2B2B2B;margin:0;line-height:1.7;padding-left:12px;"><span leaf="">%s</span></p>'
            '</section>' % (html_text(label), html_text(item))
        )
    return '<section style="margin:0 16px 32px;border-top:1px solid #E8E8E8;">%s</section>' % "".join(rows)


def component_container(components, theme):
    snippet = select_code(components, 1)
    m = re.search(r"(<section\b[^>]*>)", snippet)
    if not m:
        open_tag = container_start(theme)
    else:
        open_tag = m.group(1)
    theme_key = next((key for key, value in THEMES.items() if value is theme), "unknown")
    marker = (
        ' data-mr-li-writer-renderer="component-library-v1"'
        ' data-mr-li-writer-theme="%s"' % html.escape(theme_key, quote=True)
    )
    return open_tag[:-1] + marker + ">", "</section>"


def component_hero(theme_key, components, title, blocks):
    display_title = connected_title(title)
    if theme_key == "zen-whitespace":
        return zen_hero(display_title, blocks)
    intro = pick_first_paragraph(blocks)
    quote_before, quote_keyword, quote_after = emphasis_parts(intro)
    if theme_key == "moyu-green" and not re.search(r"[||]", title or ""):
        line1, line2 = split_colored_title(display_title, 13)
    else:
        line1, line2 = split_title(display_title, 13)
    highlight_len = min(4, max(2, len(line1) // 3)) if len(line1) > 2 else 0
    line1_text = line1[:-highlight_len] if highlight_len else line1
    line1_highlight = line1[-highlight_len:] if highlight_len else ""
    today = __import__("datetime").date.today().strftime("%Y.%m.%d")
    topic_tag_1, topic_tag_2 = public_topic_tags(title, blocks)
    if theme_key == "moyu-green":
        if re.search(r"[||]", title or ""):
            line1_text = display_title
            line2 = ""
        else:
            line1_text = line1 or title
        line1_highlight = ""
    base = {
        "标题": display_title,
        "主标题": display_title,
        "主标题行1": line1_text or display_title,
        "主标题行2": line2,
        "绿色高亮词": line1_highlight,
        "强调词": "判断",
        "顶部标签": "MR.LI WRITER",
        "内刊标签": "MR.LI WRITER",
        "头部标签": "MR.LI WRITER",
        "日期": today,
        "副标题关键词": short_text(intro, 28),
        "副标题说明": short_text(intro, 34),
        "金句前段": quote_before,
        "高亮关键词": quote_keyword,
        "金句中段": quote_after,
        "高亮关键词2": "",
        "石墨下划线关键词": quote_keyword,
        "金句收尾": "" if theme_key == "red-white" else quote_after,
        "底部左侧文字": short_text(pick_last_paragraph(blocks), 30),
        "底部摘要": short_text(pick_last_paragraph(blocks), 30),
        "标签1": topic_tag_1,
        "标签2": topic_tag_2,
        "#标签1": "#%s" % topic_tag_1,
        "#标签2": "#%s" % topic_tag_2,
        "#标签3": "#阅读笔记",
        "大标题": display_title,
        "副标题": short_text(intro, 28),
        "简介段落": short_text(intro, 38),
        "作者名": "Mr.Li Writer",
        "作者身份": "写作与内容观察",
        "等级": "精选",
        "编号": "001",
        "竖排文字": "WRITER",
    }
    if theme_key == "moyu-green":
        # 无右侧图片版更接近常规文章排版。
        return apply_component(select_code(components, 2, 1), base)
    return apply_component(select_code(components, 2, 0), base)


def component_toc(theme_key, components, headings):
    selected = [h[2] for h in headings if h[1] == 2][:3]
    if len(selected) < 2:
        return ""
    if theme_key == "moyu-green":
        cards = []
        for idx, text in enumerate(selected, 1):
            active = idx == 1
            background = "linear-gradient(135deg,#059669,#10B981)" if active else "#fff"
            color = "#fff" if active else "#111827"
            muted = "rgba(255,255,255,0.72)" if active else "#9CA3AF"
            border = "border:1px solid #E5E7EB;" if not active else ""
            shadow = "" if active else "box-shadow:0 2px 6px rgba(0,0,0,0.04);"
            margin = "8px" if idx < len(selected) else "0"
            cards.append(
                '<section style="display:block;flex:1;min-width:0;box-sizing:border-box;'
                'background:%s;%sborder-radius:12px;padding:12px 8px;margin-right:%s;%s">'
                '<p style="font-size:9px;font-weight:700;color:%s;letter-spacing:1px;margin:0 0 5px;"><span leaf="">PART %02d</span></p>'
                '<p style="font-size:13px;font-weight:800;color:%s;margin:0 0 3px;line-height:1.5;min-height:39px;white-space:normal;word-break:break-word;overflow-wrap:anywhere;">'
                '<span leaf="">%s</span></p>'
                '<p style="font-size:10px;color:%s;margin:0;"><span leaf="">CHAPTER %02d</span></p>'
                '</section>'
                % (background, border, margin, shadow, muted, idx, color, html_text(short_text(text, 24)), muted, idx)
            )
        return (
            '<section style="margin:0 20px 32px;">'
            '<section style="display:flex;align-items:center;margin-bottom:10px;">'
            '<p style="font-size:10px;color:#9CA3AF;margin:0;letter-spacing:2px;font-weight:600;"><span leaf="">%02d PARTS</span></p>'
            '</section>'
            '<section style="display:flex;width:100%%;max-width:100%%;box-sizing:border-box;align-items:stretch;">%s</section>'
            '</section>' % (len(selected), "".join(cards))
        )
    if theme_key in {"red-white", "graphite-minimal"}:
        repl = {}
        for idx, text in enumerate(selected, 1):
            repl["看点%s" % "一二三"[idx - 1]] = text
        return apply_component(select_code(components, 3, 0), repl)
    return ""


SECTION_COMPONENT = {
    "moyu-green": 4,
    "red-white": 5,
    "graphite-minimal": 5,
    "zen-whitespace": 5,
    "moyu-ticket": 3,
    "olive-journal": 3,
}

PARAGRAPH_COMPONENT = {
    "moyu-green": 5,
    "red-white": 6,
    "graphite-minimal": 6,
    "zen-whitespace": 6,
    "moyu-ticket": 5,
    "olive-journal": 10,
}

QUOTE_COMPONENT = {
    "moyu-green": 9,
    "red-white": 8,
    "graphite-minimal": 8,
    "zen-whitespace": 8,
    "moyu-ticket": 11,
    "olive-journal": 15,
}

FOOTER_COMPONENT = {
    "moyu-green": 13,
    "red-white": 16,
    "graphite-minimal": 16,
    "zen-whitespace": 15,
    "moyu-ticket": 13,
    "olive-journal": 28,
}


def component_section_title(theme_key, components, title, number):
    if theme_key == "zen-whitespace":
        return zen_section_title(title, number)
    snippet = select_code(components, SECTION_COMPONENT.get(theme_key, 3), 0)
    if not snippet:
        return ""
    return apply_component(snippet, {
        "编号": "%02d" % number,
        "01": "%02d" % number,
        "中文标题": title,
        "中文章节标题": title,
        "标题": title,
        "副标题": "CHAPTER %02d" % number,
        "ENGLISH TAG": "CHAPTER",
        "ENGLISH · 英文副标题": "CHAPTER %02d" % number,
    })


def component_paragraph(theme_key, components, text):
    if theme_key == "zen-whitespace":
        return zen_paragraph(text)
    snippet = select_code(components, PARAGRAPH_COMPONENT.get(theme_key, 5), 0)
    if not snippet:
        return paragraph(text, THEMES[theme_key])
    return apply_component(snippet, {
        "正文内容": text,
        "文字": text,
        "内容": text,
        "前半句": text[: max(1, len(text) // 2)],
        "后半句": text[max(1, len(text) // 2):],
        "需要强调的关键短语": short_text(text, 8),
    })


def component_subheading(theme_key, components, text, number):
    if theme_key == "zen-whitespace":
        return zen_paragraph("【%s】" % text)
    if theme_key == "olive-journal":
        snippet = select_code(components, 9, 0)
        return apply_component(snippet, {"前导词": "POINT %02d" % number, "标题": text, "说明": ""})
    if theme_key == "moyu-ticket":
        snippet = select_code(components, 4, 0)
        return apply_component(snippet, {"小节标题": text})
    if theme_key == "moyu-green":
        snippet = select_code(components, 9, 2)
        return apply_component(snippet, {"小标题": text, "副标题": ""})
    return component_paragraph(theme_key, components, "【%s】" % text)


def component_quote(theme_key, components, text):
    if is_prose_quote(text):
        return prose_quote(theme_key, text)
    if len(re.sub(r"\s+", "", text or "")) >= 16:
        return balanced_quote(theme_key, text)
    if theme_key == "zen-whitespace":
        return zen_quote(text)
    snippet = select_code(components, QUOTE_COMPONENT.get(theme_key, 9), 0)
    if not snippet:
        return quote(text, THEMES[theme_key])
    return apply_component(snippet, {
        "引用内容,可嵌入绿色加粗等内联样式": text,
        "重点观点": text,
        "补充说明": "",
        "金句前段": text,
        "金句中段": text,
        "金句收尾": "",
        "核心观点或关键金句": text,
        "居中金句": text,
        "亮点内容": text,
        "高亮内容": text,
        "结尾金句": text,
        "高亮关键词": short_text(text, 8),
        "正文内容": text,
        "内容": text,
    })


def prose_quote(theme_key, text):
    theme = THEMES[theme_key]
    content = inline_html(text, theme)
    if theme_key == "moyu-green":
        return (
            '<section data-balanced="prose-quote" style="background:#FFF;border:1px dashed #BBF7D0;'
            'border-radius:8px;padding:14px 16px;margin-bottom:24px;overflow:hidden;">'
            '<p style="margin:0;font-size:15px;line-height:1.85;text-align:left;color:#059669;font-weight:700;'
            'word-break:break-word;overflow-wrap:anywhere;">%s</p></section>' % content
        )
    if theme_key == "red-white":
        return (
            '<section data-balanced="prose-quote" style="background:#FEF2F2;border-radius:0 10px 10px 0;'
            'border-left:4px solid #DC2626;padding:18px 22px;margin-bottom:24px;overflow:hidden;">'
            '<p style="font-size:15px;font-weight:700;color:#991B1B;margin:0;line-height:1.85;text-align:left;'
            'word-break:break-word;overflow-wrap:anywhere;">%s</p></section>' % content
        )
    if theme_key == "graphite-minimal":
        return (
            '<section data-balanced="prose-quote" style="border-left:3px solid #52525B;padding:16px 0 16px 24px;'
            'margin:0 10px 28px;overflow:hidden;">'
            '<p style="font-size:15px;font-weight:600;color:#27272A;margin:0;line-height:1.85;letter-spacing:0.3px;'
            'text-align:left;word-break:break-word;overflow-wrap:anywhere;">%s</p></section>' % content
        )
    if theme_key == "zen-whitespace":
        return (
            '<section data-balanced="prose-quote" style="margin:40px 16px;padding:28px 20px;border-top:1px solid #E8E8E8;'
            'border-bottom:1px solid #E8E8E8;overflow:hidden;">'
            '<p style="font-family:\'Noto Serif SC\',Georgia,\'Times New Roman\',serif;font-size:15px;font-weight:600;'
            'color:#2B2B2B;margin:0;line-height:1.95;letter-spacing:0.5px;text-align:left;word-break:break-word;'
            'overflow-wrap:anywhere;">%s</p></section>' % content
        )
    if theme_key == "moyu-ticket":
        return (
            '<section data-balanced="prose-quote" style="margin:0 10px 26px;background:#fffef8;border:2px solid #1a1a1a;'
            'box-shadow:4px 4px 0 #1a1a1a;padding:18px;overflow:hidden;">'
            '<p style="font-size:15px;font-weight:800;color:#064E3B;line-height:1.85;margin:0;text-align:left;'
            'word-break:break-word;overflow-wrap:anywhere;">%s</p></section>' % content
        )
    if theme_key == "olive-journal":
        return (
            '<section data-balanced="prose-quote" style="margin:0 0 24px;background:#fffaf0;border:1px solid #d8d1c4;'
            'border-left:4px solid #ed7b2f;border-radius:8px;padding:18px 20px;overflow:hidden;">'
            '<p style="font-size:15px;line-height:1.85;color:#1E1F23;font-weight:700;margin:0;text-align:left;'
            'word-break:break-word;overflow-wrap:anywhere;">%s</p></section>' % content
        )
    return (
        '<section data-balanced="prose-quote" style="margin:0 10px 24px;background:%s;border-left:4px solid %s;'
        'border-radius:0 10px 10px 0;padding:16px 18px;overflow:hidden;">'
        '<p style="font-size:15px;font-weight:700;color:%s;margin:0;line-height:1.85;text-align:left;'
        'word-break:break-word;overflow-wrap:anywhere;">%s</p></section>'
        % (theme["light"], theme["main"], theme["deep"], content)
    )


def balanced_quote(theme_key, text):
    theme = THEMES[theme_key]
    content = balanced_leaf_lines(text, max_len=15)
    if theme_key == "moyu-green":
        return (
            '<section data-balanced="quote" style="background:#FFF;border:1px dashed #BBF7D0;'
            'border-radius:8px;padding:14px 16px;margin-bottom:24px;text-align:center;overflow:hidden;">'
            '<p style="margin:0;line-height:1.75;">'
            '<span style="font-size:15px;color:#059669;font-weight:bold;border-bottom:3px solid #FDE68A;padding-bottom:2px;">%s</span>'
            '</p></section>' % content
        )
    if theme_key == "red-white":
        return (
            '<section data-balanced="quote" style="background:#FEF2F2;border-radius:0 10px 10px 0;'
            'border-left:4px solid #DC2626;padding:18px 22px;margin-bottom:24px;overflow:hidden;">'
            '<p style="font-size:16px;font-weight:800;color:#991B1B;margin:0;line-height:1.85;text-align:left;">「%s」</p>'
            '</section>' % content
        )
    if theme_key == "graphite-minimal":
        return (
            '<section data-balanced="quote" style="border-left:3px solid #52525B;padding:16px 0 16px 24px;'
            'margin:0 10px 28px;overflow:hidden;">'
            '<p style="font-size:16px;font-weight:700;color:#27272A;margin:0;line-height:1.85;letter-spacing:0.5px;">「%s」</p>'
            '</section>' % content
        )
    if theme_key == "zen-whitespace":
        return (
            '<section data-balanced="quote" style="margin:40px 16px;padding:36px 20px;border-top:1px solid #E8E8E8;'
            'border-bottom:1px solid #E8E8E8;text-align:center;overflow:hidden;">'
            '<p style="font-family:\'Noto Serif SC\',Georgia,\'Times New Roman\',serif;font-size:17px;font-weight:600;'
            'color:#2B2B2B;margin:0;line-height:1.95;letter-spacing:0.8px;">「%s」</p></section>' % content
        )
    if theme_key == "moyu-ticket":
        return (
            '<section data-balanced="quote" style="margin:0 10px 26px;background:#fffef8;border:2px solid #1a1a1a;'
            'box-shadow:4px 4px 0 #1a1a1a;padding:18px 18px;overflow:hidden;">'
            '<p style="font-size:16px;font-weight:900;color:#064E3B;line-height:1.85;margin:0;">%s</p></section>' % content
        )
    if theme_key == "olive-journal":
        return (
            '<section data-balanced="quote" style="margin:0 0 24px;background:#fffaf0;border:1px solid #d8d1c4;'
            'border-left:4px solid #ed7b2f;border-radius:8px;padding:18px 20px;overflow:hidden;">'
            '<p style="font-size:16px;line-height:1.85;color:#1E1F23;font-weight:800;margin:0;">%s</p></section>' % content
        )
    return (
        '<section data-balanced="quote" style="margin:0 10px 24px;background:%s;border-left:4px solid %s;'
        'border-radius:0 10px 10px 0;padding:16px 18px;overflow:hidden;">'
        '<p style="font-size:16px;font-weight:800;color:%s;margin:0;line-height:1.85;">%s</p></section>'
        % (theme["light"], theme["main"], theme["deep"], content)
    )


def component_list(theme_key, components, items, ordered=False):
    if theme_key == "zen-whitespace":
        return zen_list(items, ordered=ordered)
    if theme_key == "olive-journal":
        rows = "".join(
            '<li style="margin:0 0 8px;font-size:14px;line-height:1.8;color:#4d4f46;"><span leaf="">%s</span></li>' % html_text(item)
            for item in items
        )
        return '<section style="margin-top:24px;background:#eeefe9;border:1px solid #bfc1b7;border-radius:6px;padding:16px 20px;"><ul style="margin:0;padding-left:18px;">%s</ul></section>' % rows
    return list_block(items, THEMES[theme_key], ordered=ordered)


def component_table(theme_key, headers, rows):
    """Render mobile-first semantic tables with bounded scrolling only when needed."""
    theme = THEMES[theme_key]
    column_count = max(1, len(headers))
    compact = column_count >= 4
    scrollable = column_count >= 5
    font_size = "12px" if compact else "13px"
    padding = "8px 5px" if compact else "9px 7px"
    cell_style = (
        "box-sizing:border-box;padding:%s;border:1px solid %s;vertical-align:top;"
        "font-size:%s;line-height:1.55;word-break:normal;overflow-wrap:anywhere;white-space:normal;"
        % (padding, theme["line"], font_size)
    )
    head_cells = "".join(
        '<th style="%sbackground:%s;color:%s;font-weight:800;text-align:left;">%s</th>'
        % (cell_style, theme["light"], theme["deep"], inline_html(value, theme))
        for value in headers
    )
    body_rows = []
    for row_index, row in enumerate(rows):
        background = theme["surface"] if row_index % 2 == 0 else theme["lighter"]
        cells = "".join(
            '<td style="%sbackground:%s;color:%s;">%s</td>'
            % (cell_style, background, theme["text"], inline_html(value, theme))
            for value in row
        )
        body_rows.append("<tr>%s</tr>" % cells)
    wrapper_style = "width:100%;max-width:100%;box-sizing:border-box;margin:0 0 24px;"
    table_width_style = "max-width:100%;"
    if scrollable:
        min_width = min(680, max(420, column_count * 92))
        wrapper_style += "overflow-x:auto;overflow-y:hidden;-webkit-overflow-scrolling:touch;"
        table_width_style = "min-width:%dpx;max-width:680px;" % min_width
    table_style = (
        "width:100%;"
        + table_width_style
        + "table-layout:fixed;border-collapse:collapse;border-spacing:0;box-sizing:border-box;"
    )
    return (
        '<section style="%s"><table style="%s"><thead><tr>%s</tr></thead><tbody>%s</tbody></table></section>'
        % (wrapper_style, table_style, head_cells, "".join(body_rows))
    )


def component_footer(theme_key, components, blocks):
    snippet = select_code(components, FOOTER_COMPONENT.get(theme_key, 13), 0)
    if not snippet:
        return signature(THEMES[theme_key])
    return apply_component(snippet, {
        "作者名": "Mr.Li Writer",
        "一句话简介,如:热衷于分享 AI 观察与干货": "让写作如此简单",
        "文末互动引导": "如果你觉得今天这篇有收获,欢迎点赞、在看、转发,我们下篇见",
        "互动文案": "如果这篇文章对你有帮助,欢迎点赞、在看、转发。",
    })


def build_component_section(md_text, title, theme_key):
    components = load_theme_components(theme_key)
    if len(components) < 5:
        raise RuntimeError("完整组件库不可用: %s" % theme_key)
    theme = THEMES[theme_key]
    blocks = parse_blocks(md_text)
    if not title:
        for block in blocks:
            if block[0] == "heading" and block[1] == 1:
                title = block[2]
                break
    title = title or "公众号文章"
    headings = [b for b in blocks if b[0] == "heading"]

    open_tag, close_tag = component_container(components, theme)
    parts = [open_tag, component_hero(theme_key, components, title, blocks), component_toc(theme_key, components, headings)]
    chapter = 0
    sub_count = 0
    for block in blocks:
        kind = block[0]
        if kind == "heading":
            level, text = block[1], block[2]
            if level == 1:
                continue
            if level == 2:
                chapter += 1
                sub_count = 0
                parts.append(component_section_title(theme_key, components, text, chapter))
            else:
                sub_count += 1
                parts.append(component_subheading(theme_key, components, text, sub_count))
        elif kind == "paragraph":
            parts.append(component_paragraph(theme_key, components, block[1]))
        elif kind == "quote":
            parts.append(component_quote(theme_key, components, block[1]))
        elif kind == "list":
            parts.append(component_list(theme_key, components, block[1], ordered=False))
        elif kind == "olist":
            parts.append(component_list(theme_key, components, block[1], ordered=True))
        elif kind == "table":
            parts.append(component_table(theme_key, block[1], block[2]))
        elif kind == "code":
            parts.append(code_block(block[2], block[1], theme))
        elif kind == "image":
            parts.append(image_block(block[1], block[2], theme))
        elif kind == "hr":
            parts.append(hr(theme))
    parts.append(component_footer(theme_key, components, blocks))
    parts.append(close_tag)
    parts.append('<p style="display:none;"><mp-style-type data-value="3"></mp-style-type></p>')
    return repair_centered_text_orphans("\n".join(p for p in parts if p))


def infer_theme(md_text, title=""):
    text = (title + "\n" + md_text[:5000]).lower()
    scores = {k: 0 for k in THEMES}
    rules = [
        ("moyu-green", ["教程", "清单", "工具", "盘点", "步骤", "方法", "指南", "避坑"]),
        ("red-white", ["观点", "深度", "分析", "争议", "为什么", "评论", "真相"]),
        ("graphite-minimal", ["科技", "ai", "设计", "专业", "品牌", "产品"]),
        ("zen-whitespace", ["随笔", "生活", "关系", "情感", "冥想", "极简", "思考"]),
        ("moyu-ticket", ["测评", "对比", "评测", "选择", "评分"]),
        ("olive-journal", ["复盘", "案例", "内刊", "系统", "报告", "实战"]),
    ]
    for theme, keywords in rules:
        if any(k.lower() in text for k in keywords):
            scores[theme] += 2
    return max(scores.items(), key=lambda item: item[1])[0] if max(scores.values()) else "moyu-green"


def resolve_theme(theme, md_text, title=""):
    if theme == "random":
        return random.choice(list(THEMES)), "random"
    if theme == "auto":
        selected = infer_theme(md_text, title)
        return selected, "auto"
    if theme not in THEMES:
        raise ValueError("未知公众号主题: %s" % theme)
    return theme, "manual"


def style_attr(value):
    return value.replace("\n", "").strip()


def container_start(theme):
    bg = theme["surface"]
    return (
        '<section style="%s">'
        % style_attr(
            "max-width:677px;margin:0 auto;background:%s;font-family:-apple-system,BlinkMacSystemFont,'PingFang SC','Hiragino Sans GB','Microsoft YaHei',sans-serif;color:%s;line-height:1.8;letter-spacing:0.4px;overflow-x:hidden;padding:10px 0 22px;"
            % (bg, theme["text"])
        )
    )


def hero(title, theme):
    if theme["mode"] == "zen":
        return """
<section style="margin:32px 16px 48px;padding:40px 24px;border-top:1px solid {line};border-bottom:1px solid {line};text-align:center;">
  <p style="font-family:Georgia,'Times New Roman',serif;font-size:23px;font-weight:700;color:{title_color};margin:0;line-height:1.65;letter-spacing:0.8px;">{title}</p>
</section>""".format(line=theme["line"], title_color=theme["title"], title=leaf(title))
    if theme["mode"] == "ticket":
        return """
<section style="margin:10px 10px 34px;background:{surface};border:2px solid {main};border-radius:3px;padding:26px 22px;box-shadow:{shadow};">
  <p style="font-size:12px;color:{deep};font-weight:800;margin:0 0 10px;letter-spacing:2px;">{label}</p>
  <p style="font-size:22px;font-weight:900;color:{title_color};margin:0;line-height:1.55;">{title}</p>
</section>""".format(surface=theme["surface"], main=theme["main"], deep=theme["deep"], shadow=theme["shadow"], title_color=theme["title"], label=leaf("FEATURE TICKET"), title=leaf(title))
    return """
<section style="margin:10px 10px 32px;background:{surface};border-radius:{radius};box-shadow:{shadow};padding:28px 24px 24px;overflow:hidden;border:1px solid {line};">
  <p style="font-size:12px;color:{main};font-weight:800;margin:0 0 10px;letter-spacing:2px;">{label}</p>
  <p style="font-size:23px;font-weight:900;color:{title_color};margin:0;line-height:1.55;">{title}</p>
</section>""".format(surface=theme["surface"], radius=theme["radius"], shadow=theme["shadow"], line=theme["line"], main=theme["main"], title_color=theme["title"], label=leaf("MR.LI WRITER"), title=leaf(title))


def toc(headings, theme):
    selected = [h for h in headings if h[0] == 2][:3]
    if len(selected) < 2:
        return ""
    items = []
    for idx, (_, text) in enumerate(selected, 1):
        items.append(
            '<section style="flex:1;background:%s;border-radius:%s;padding:15px 10px;margin-right:%s;text-align:center;border:1px solid %s;">'
            '<p style="display:inline-block;background:%s;color:#FFFFFF;font-size:12px;font-weight:800;padding:2px 9px;border-radius:4px;margin:0 0 8px;">%s</p>'
            '<p style="font-size:13px;font-weight:700;color:%s;margin:0;line-height:1.5;">%s</p></section>'
            % (
                theme["light"],
                theme["radius"],
                "8px" if idx < len(selected) else "0",
                theme["lighter"],
                theme["main"],
                leaf("%02d" % idx),
                theme["title"],
                leaf(text),
            )
        )
    return '<section style="padding:0 10px 30px;"><p style="font-size:14px;color:%s;margin:0 0 14px;letter-spacing:1px;">%s</p><section style="display:flex;justify-content:space-between;">%s</section></section>' % (theme["muted"], leaf("本文看点"), "".join(items))


def section_title(text, number, theme, ending=False):
    label = "∞" if ending else "%02d" % number
    if theme["mode"] == "zen":
        return """
<section style="margin-top:64px;margin-bottom:32px;padding:0 16px;">
  <p style="font-size:10px;color:{main};font-weight:600;letter-spacing:4px;margin:0 0 10px;text-transform:uppercase;">{label}</p>
  <h3 style="font-family:Georgia,'Times New Roman',serif;font-size:22px;font-weight:700;color:{title_color};margin:0 0 16px;letter-spacing:0.5px;line-height:1.4;">{text}</h3>
  <section style="width:40px;height:2px;background:{main};"><span leaf=""><br></span></section>
</section>""".format(main=theme["main"], title_color=theme["title"], label=leaf("%s · CHAPTER" % label), text=leaf(text))
    return """
<section style="margin-top:46px;margin-bottom:26px;padding:0 10px;">
  <section style="display:flex;align-items:center;margin-bottom:18px;padding-bottom:14px;border-bottom:3px solid {main};">
    <span style="display:inline-block;background:{main};color:#FFFFFF;font-size:17px;font-weight:900;padding:4px 13px;border-radius:6px;margin-right:14px;line-height:1.3;">{label}</span>
    <section>
      <p style="font-size:10px;color:{main};font-weight:700;letter-spacing:3px;margin:0 0 2px;text-transform:uppercase;">{en}</p>
      <h3 style="font-size:18px;font-weight:800;color:{title_color};margin:0;letter-spacing:0.5px;">{text}</h3>
    </section>
  </section>
</section>""".format(main=theme["main"], title_color=theme["title"], label=leaf(label), en=leaf("CHAPTER"), text=leaf(text))


def paragraph(text, theme):
    return '<p style="margin:0 10px 20px;font-size:15px;line-height:1.8;text-align:justify;color:%s;">%s</p>' % (theme["text"], inline_html(text, theme))


def subheading(text, theme):
    return '<p style="font-size:15px;font-weight:800;color:%s;margin:28px 10px 14px;padding-left:10px;border-left:3px solid %s;line-height:1.4;">%s</p>' % (theme["title"], theme["main"], leaf(text))


def quote(text, theme):
    return '<section style="margin:0 10px 24px;background:%s;border-radius:0 10px 10px 0;border-left:4px solid %s;padding:16px 18px;"><p style="font-size:16px;font-weight:800;color:%s;margin:0;line-height:1.8;">%s</p></section>' % (theme["light"], theme["main"], theme["deep"], leaf("「%s」" % text.strip("「」")))


def list_block(items, theme, ordered=False):
    rows = []
    for idx, item in enumerate(items, 1):
        badge = "%02d" % idx if ordered else "•"
        rows.append('<p style="margin:0 0 12px;font-size:15px;line-height:1.75;color:%s;"><span style="display:inline-block;background:%s;color:%s;border-radius:5px;padding:1px 8px;margin-right:8px;font-weight:900;">%s</span>%s</p>' % (theme["text"], theme["lighter"], theme["deep"], leaf(badge), inline_html(item, theme)))
    return '<section style="margin:0 10px 22px;padding:16px 16px 4px;background:%s;border-radius:%s;border:1px solid %s;">%s</section>' % (theme["light"], theme["radius"], theme["lighter"], "".join(rows))


def table_block(headers, rows, theme):
    theme_key = next((key for key, value in THEMES.items() if value is theme), "moyu-green")
    return component_table(theme_key, headers, rows)


def code_block(code, lang, theme):
    rows = []
    for line in code.split("\n") or [""]:
        rows.append(
            '<p style="margin:0;font-family:Consolas,Monaco,monospace;font-size:13px;line-height:1.6;'
            'letter-spacing:0;color:#E2E8F0;white-space:nowrap;word-break:normal;overflow-wrap:normal;">%s</p>'
            % code_leaf(line)
        )
    return (
        '<section style="margin:0 10px 22px;max-width:100%%;border-radius:8px;overflow:hidden;background:#1E293B;'
        'box-shadow:0 4px 16px -8px rgba(15,23,42,0.4);">'
        '<section style="padding:8px 14px;background:#0F172A;"><span style="font-size:12px;color:#94A3B8;'
        'font-family:Consolas,Monaco,monospace;letter-spacing:1px;">%s</span></section>'
        '<section style="width:100%%;max-width:100%%;box-sizing:border-box;padding:11px 14px;overflow-x:auto;'
        'overflow-y:hidden;-webkit-overflow-scrolling:touch;">%s</section></section>'
        % (leaf(lang), "".join(rows))
    )


def image_block(alt, src, theme):
    caption = ""
    if alt:
        caption = '<p style="font-size:12px;color:%s;text-align:center;margin:0 10px 24px;">%s</p>' % (theme["muted"], leaf("— %s" % alt))
    return '<section style="background:#FFF;border-radius:12px;padding:6px;border:1px solid %s;box-shadow:0 4px 12px -2px rgba(0,0,0,0.08);margin:0 10px 8px;"><section style="margin:0;border-radius:8px;overflow:hidden;"><span leaf=""><img src="%s" style="max-width:100%%;height:auto;display:block;margin:0 auto;"></span></section></section>%s' % (theme["line"], html.escape(src), caption)


def hr(theme):
    return '<section style="padding:0 10px;"><section style="height:1px;background:linear-gradient(to right,transparent,%s,%s,%s,transparent);margin:36px 0;"><span leaf=""><br></span></section></section>' % (theme["mark"], theme["main"], theme["mark"])


def signature(theme):
    return '<section style="margin:44px 10px 10px;padding:22px 20px;background:%s;border-radius:%s;border:1px solid %s;text-align:center;"><p style="font-size:14px;color:%s;margin:0 0 8px;line-height:1.8;">%s</p><p style="font-size:12px;color:%s;margin:0;">%s</p></section>' % (theme["light"], theme["radius"], theme["lighter"], theme["text"], leaf("我是 Mr.Li Writer,让想法变成文章"), theme["muted"], leaf("如果你觉得今天这篇有收获,欢迎点赞、在看、转发,我们下篇见"))


def build_section(md_text, title, theme):
    blocks = parse_blocks(md_text)
    if not title:
        for block in blocks:
            if block[0] == "heading" and block[1] == 1:
                title = block[2]
                break
    title = title or "公众号文章"
    headings = [(b[1], b[2]) for b in blocks if b[0] == "heading"]
    parts = [container_start(theme), hero(title, theme), toc(headings, theme)]
    chapter = 0
    for block in blocks:
        kind = block[0]
        if kind == "heading":
            level, text = block[1], block[2]
            if level == 1:
                continue
            if level == 2:
                chapter += 1
                ending = bool(re.search(r"(结语|总结|最后|尾声|写在最后)", text))
                parts.append(section_title(text, chapter, theme, ending=ending))
            else:
                parts.append(subheading(text, theme))
        elif kind == "paragraph":
            parts.append(paragraph(block[1], theme))
        elif kind == "quote":
            parts.append(quote(block[1], theme))
        elif kind == "list":
            parts.append(list_block(block[1], theme, ordered=False))
        elif kind == "olist":
            parts.append(list_block(block[1], theme, ordered=True))
        elif kind == "table":
            parts.append(table_block(block[1], block[2], theme))
        elif kind == "code":
            parts.append(code_block(block[2], block[1], theme))
        elif kind == "image":
            parts.append(image_block(block[1], block[2], theme))
        elif kind == "hr":
            parts.append(hr(theme))
    parts.append(signature(theme))
    parts.append("</section>")
    return repair_centered_text_orphans("\n".join(p for p in parts if p))


PREVIEW_TEMPLATE = """<!DOCTYPE html>
<html lang="zh-CN">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1">
<meta name="mr-li-writer-template" content="gzh-preview-v2">
<meta name="mr-li-writer-theme" content="__THEME__">
<meta name="mr-li-writer-content-sha256" content="__CONTENT_SHA256__">
<title>__TITLE__ · 公众号排版预览</title>
<style>
body{margin:0;background:#eef0f2;font-family:-apple-system,BlinkMacSystemFont,'PingFang SC','Hiragino Sans GB','Microsoft YaHei',sans-serif;-webkit-text-size-adjust:100%;}
.gzh-toolbar{position:fixed;top:0;left:0;right:0;height:54px;background:#ffffff;box-shadow:0 1px 10px rgba(0,0,0,.08);display:flex;align-items:center;justify-content:space-between;padding:0 16px;z-index:99;}
.gzh-hint{font-size:13px;color:#6b7280;line-height:1.4;}
.gzh-hint b{color:#111827;}
.gzh-copy{background:__MAIN__;color:#fff;border:0;border-radius:9px;padding:10px 20px;font-size:14px;font-weight:700;cursor:pointer;box-shadow:0 3px 10px rgba(0,0,0,.16);white-space:nowrap;}
.gzh-toast{position:fixed;top:66px;left:50%;transform:translateX(-50%);background:#111827;color:#fff;padding:11px 20px;border-radius:10px;font-size:14px;font-weight:600;opacity:0;pointer-events:none;transition:opacity .25s;z-index:100;box-shadow:0 6px 20px rgba(0,0,0,.25);max-width:calc(100% - 32px);box-sizing:border-box;text-align:center;}
.gzh-toast.show{opacity:1;}
.gzh-stage{max-width:700px;margin:78px auto 64px;padding:0 8px;}
.gzh-sr-only{position:absolute;width:1px;height:1px;padding:0;margin:-1px;overflow:hidden;clip:rect(0,0,0,0);white-space:nowrap;border:0;}
</style>
</head>
<body>
<header class="gzh-toolbar">
  <span class="gzh-hint">下方是排版效果 · 点右侧 <b>复制</b> 直接粘到公众号</span>
  <button type="button" class="gzh-copy" id="gzhCopyBtn" onclick="gzhCopy()">复制到公众号</button>
</header>
<div class="gzh-toast" id="gzhToast" role="status" aria-live="polite"></div>
<main class="gzh-stage">
<h1 class="gzh-sr-only">__TITLE__</h1>
<article id="gzh-content">
__CONTENT__
</article>
</main>
<script>
function gzhShowToast(msg){var t=document.getElementById('gzhToast');t.textContent=msg;t.classList.add('show');clearTimeout(t._timer);t._timer=setTimeout(function(){t.classList.remove('show');},2800);}
function gzhClipboardHtml(el){
  return '<meta charset="utf-8">'+el.innerHTML;
}
function gzhFallbackCopy(el){
  var range=document.createRange();
  range.selectNodeContents(el);
  var sel=window.getSelection();
  sel.removeAllRanges();
  sel.addRange(range);
  var ok=false;
  try{ok=document.execCommand('copy');}catch(e){ok=false;}
  sel.removeAllRanges();
  return ok;
}
function gzhCopy(){
  var el=document.getElementById('gzh-content');
  var html=gzhClipboardHtml(el);
  var text=el.innerText||el.textContent||'';
  try{
    if(navigator.clipboard&&window.ClipboardItem){
      var item=new ClipboardItem({
        'text/html':new Blob([html],{type:'text/html'}),
        'text/plain':new Blob([text],{type:'text/plain'})
      });
      navigator.clipboard.write([item]).then(function(){
        gzhShowToast('已复制富文本,去公众号编辑器粘贴即可');
      },function(){
        var ok=gzhFallbackCopy(el);
        gzhShowToast(ok?'已复制,去公众号编辑器粘贴即可':'自动复制失败,请手动全选再复制');
      });
      return;
    }
  }catch(e){}
  var ok=gzhFallbackCopy(el);
  gzhShowToast(ok?'已复制,去公众号编辑器粘贴即可':'自动复制失败,请手动全选再复制');
}
</script>
</body>
</html>
"""


def normalized_section_html(section_html):
    return section_html.replace("\r\n", "\n").replace("\r", "\n").strip()


def section_digest(section_html):
    return hashlib.sha256(normalized_section_html(section_html).encode("utf-8")).hexdigest()


def write_preview(section_html, out_path, title, theme):
    preview_path = os.path.splitext(out_path)[0] + "-preview.html"
    theme_key = next((key for key, value in THEMES.items() if value is theme), "unknown")
    page = (
        PREVIEW_TEMPLATE.replace("__TITLE__", html.escape(title))
        .replace("__MAIN__", theme["main"])
        .replace("__THEME__", html.escape(theme_key, quote=True))
        .replace("__CONTENT_SHA256__", section_digest(section_html))
        .replace("__CONTENT__", section_html)
    )
    with open(preview_path, "w", encoding="utf-8") as f:
        f.write(page)
    return preview_path


def validate_file(path):
    validator = os.path.join(SCRIPT_DIR, "validate_gzh_html.py")
    if not os.path.isfile(validator):
        return 0
    return subprocess.call([sys.executable, validator, path])


def require_task_state(task_state, expected_style):
    if not task_state:
        print("[错误] 公众号正式排版前必须传入 --task-state,并通过 scripts/validate_task_intake.py 确认必问项。", file=sys.stderr)
        return 2
    checker = os.path.join(SCRIPT_DIR, "validate_task_intake.py")
    return subprocess.call([
        sys.executable,
        checker,
        task_state,
        "--phase",
        "layout",
        "--platform",
        "公众号",
        "--expected-style",
        expected_style,
    ])


def list_themes():
    print("可用公众号主题:")
    print("  %-18s %s" % ("auto", "按题材自动匹配;最终排版需加 --auto-theme-ok 表示用户已授权"))
    print("  %-18s %s" % ("random", "从 6 套公众号主题中随机选择;最终排版需加 --auto-theme-ok 表示用户已授权"))
    for key, item in THEMES.items():
        print("  %-18s %s - %s" % (key, item["name"], item["desc"]))
    if os.path.isdir(GZH_REF_DIR):
        print("\n完整组件库: %s" % GZH_REF_DIR)
        print("查看组件清单: python3 scripts/gzh_component_inventory.py .")
    else:
        print("\n[提示] 未找到完整组件库目录: %s" % GZH_REF_DIR)


def list_components():
    if not os.path.isdir(GZH_REF_DIR):
        print("[错误] 未找到完整组件库目录: %s" % GZH_REF_DIR, file=sys.stderr)
        return 1
    total = 0
    for key in THEMES:
        path = os.path.join(GZH_REF_DIR, "theme-%s.md" % key)
        if not os.path.isfile(path):
            print("[缺失] %s -> %s" % (key, path))
            continue
        with open(path, encoding="utf-8", errors="replace") as f:
            components = COMPONENT_HEADING_RE.findall(f.read())
        total += len(components)
        print("%s/%s: %d 个组件" % (THEMES[key]["name"], key, len(components)))
        for item in components:
            print("  - %s" % item.strip())
    print("总计: %d 个主题组件" % total)
    return 0


def main():
    parser = argparse.ArgumentParser(description="Markdown -> 公众号内联 HTML")
    parser.add_argument("md", nargs="?", help="Markdown 正文文件路径")
    parser.add_argument("-o", "--output", help="输出干净公众号 HTML 片段")
    parser.add_argument("-t", "--theme", default="auto", help="公众号主题: auto/random/%s" % "/".join(THEMES))
    parser.add_argument(
        "--auto-theme-ok",
        action="store_true",
        help="确认用户已授权系统自动/随机选择公众号主题;未授权时请使用 -t 指定具体主题",
    )
    parser.add_argument(
        "--theme-confirmed",
        action="store_true",
        help="确认用户已选择当前公众号主题;具体主题最终排版必须带此参数",
    )
    parser.add_argument("--title", default="", help="文章标题")
    parser.add_argument("--task-state", default="", help="任务状态 JSON;正式排版前必须通过必问项/恢复任务门禁")
    parser.add_argument("--no-preview", action="store_true", help="只生成干净 HTML,不生成复制预览页")
    parser.add_argument(
        "--allow-fallback-preview",
        action="store_true",
        help="只允许临时预览在完整组件库失败时回退到快速预览器;最终交付不要使用",
    )
    parser.add_argument("--list-themes", action="store_true", help="列出公众号主题")
    parser.add_argument("--list-components", action="store_true", help="列出已接入的 gzh-design 完整组件库")
    args = parser.parse_args()

    if args.list_themes:
        list_themes()
        return 0
    if args.list_components:
        return list_components()
    if not args.md:
        parser.error("缺少 Markdown 文件路径")
    if args.theme not in THEMES and args.theme not in SPECIAL_THEMES:
        list_themes()
        parser.error("未知公众号主题: %s" % args.theme)
    if args.theme in SPECIAL_THEMES and not args.auto_theme_ok:
        parser.error(
            "公众号最终排版不能静默使用主题 %s;请先确认公众号主题,使用 -t <主题名>,或在用户明确授权自动匹配后添加 --auto-theme-ok"
            % args.theme
        )
    if args.theme in THEMES and not args.theme_confirmed:
        parser.error(
            "公众号最终排版不能由智能体静默指定主题 %s;请先让用户确认主题,再添加 --theme-confirmed"
            % args.theme
        )
    intake_rc = require_task_state(args.task_state, args.theme)
    if intake_rc != 0:
        return intake_rc

    md_path = os.path.abspath(args.md)
    if not os.path.isfile(md_path):
        print("[错误] 文件不存在: %s" % md_path, file=sys.stderr)
        return 1
    with open(md_path, encoding="utf-8") as f:
        md_text = f.read()
    warn_process_leaks(md_text)
    warn_commercial_source_exposure(md_text, task_state=args.task_state)

    title = args.title or os.path.splitext(os.path.basename(md_path))[0]
    theme_key, reason = resolve_theme(args.theme, md_text, title)
    theme = THEMES[theme_key]
    try:
        section = build_component_section(md_text, title, theme_key)
        render_mode = "component-library"
    except Exception as exc:
        if not args.allow_fallback_preview:
            print("[错误] 完整组件库渲染失败,已阻断正式交付: %s" % exc, file=sys.stderr)
            print("[提示] 只做临时预览时可添加 --allow-fallback-preview;最终交付必须修复组件库渲染问题。", file=sys.stderr)
            return 1
        print("[警告] 完整组件库渲染失败,仅回退到快速预览器生成临时预览: %s" % exc, file=sys.stderr)
        section = build_section(md_text, title, theme)
        render_mode = "fallback-preview"

    out_path = os.path.abspath(args.output or os.path.splitext(md_path)[0] + "_gzh.html")
    with open(out_path, "w", encoding="utf-8") as f:
        f.write(section)

    rc = validate_file(out_path)
    if rc != 0:
        return rc

    preview = ""
    if not args.no_preview:
        preview = write_preview(section, out_path, title, theme)

    print("[完成] 公众号主题=%s/%s (%s), 渲染模式=%s" % (theme["name"], theme_key, reason, render_mode))
    print("[输出] 干净正文片段: %s" % out_path)
    if preview:
        print("[输出] 复制预览页: %s" % preview)
    return 0


if __name__ == "__main__":
    sys.exit(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