ZJU-REAL/Easel / skills/builtin/skill-wechat-publisher
skill-wechat-publisher
Verified微信公众号文章自动创作与发布工具。给定参考文章、文字或文档,自动搜索整理全网相关信息、生成图文并茂的公众号文章,并发布到微信公众号草稿箱。特别强调反 AI 检测写作。 触发场景(沾边就用):用户提到"公众号 / 微信文章 / 推文 / 公号 / 发文 / mp / 草稿箱 / 群发",或要求把话题/文档/笔记写成公众号文章并发布到微信。
scripts/html_converter.py
View on GitHub#!/usr/bin/env python3
"""
Markdown → 微信公众号 HTML 转换器
微信公众号编辑器的特殊限制:
- 不支持外部 CSS 和 JS 引用
- 不支持 class 和 id 属性(部分支持但不可靠)
- 所有样式必须使用内联 style
- 不支持 <style> 标签
- 不支持 position、float 等部分 CSS 属性
- 图片必须使用微信 CDN 的 URL
本转换器将 Markdown 转换为满足以上限制的 HTML,并应用精美的内联排版样式。
## 支持的 Markdown 扩展语法(排版增强标记)
除了标准 Markdown 之外,本转换器支持以下自定义行内标记,用来让段内
文字有更丰富的标色变化(字体颜色 + 背景高亮):
| 标记 | 含义 | 效果 |
|----------------|---------------|--------------------|
| `**文本**` | 加粗(主强调)| 深色 + 黄色下划线 |
| `==文本==` | 黄色高亮 | 荧光笔马克效果 |
| `++文本++` | 蓝色高亮 | 冷色马克效果 |
| `%%文本%%` | 粉色高亮 | 警示马克效果 |
| `&&文本&&` | 绿色高亮 | 正向马克效果 |
| `!!文本!!` | 红色强调 | 警告/反对/关键数字 |
| `@@文本@@` | 蓝色强调 | 链接/术语/名词 |
| `^^文本^^` | 橙色强调 | 温暖点缀 |
| `*文本*` | 斜体 | 保留 |
写作时在同一篇文章里混用多种标记,避免"整篇只有一种加粗色"的
AI 号指纹。建议密度:每 500 字出现 3-5 处行内标记,但不要集中堆在
同一段里。
## 主题
样式从 assets/themes/<theme>.json 加载,默认 refined-blue。
每个主题包含 styles(标签样式)+ highlights(行内标记样式)+
section_divider_text(分节符字符)。
"""
import re
import json
import sys
import html
from pathlib import Path
# ============================================================
# 样式加载 / 主题解析
# ============================================================
DEFAULT_STYLES = {
"body": "font-family: -apple-system, BlinkMacSystemFont, 'Helvetica Neue', 'PingFang SC', 'Hiragino Sans GB', 'Microsoft YaHei', sans-serif; font-size: 15.5px; color: #242832; line-height: 1.75; letter-spacing: 0.4px; word-spacing: 1.5px; padding: 12px 0;",
"h1": "font-size: 22px; font-weight: 800; color: #0b1530; text-align: left; margin: 32px 0 20px; padding: 0 0 12px 0; border-bottom: 2px solid #0b1530; letter-spacing: 0.5px; line-height: 1.4;",
"h2": "font-size: 18px; font-weight: 700; color: #0b1530; margin: 36px 0 16px; padding: 10px 0 10px 14px; border-left: 4px solid #2e5bff; background: linear-gradient(90deg, #eef2ff 0%, #ffffff 70%); border-radius: 0 6px 6px 0; line-height: 1.5; letter-spacing: 0.3px;",
"h3": "font-size: 16px; font-weight: 700; color: #1a2240; margin: 26px 0 12px; padding: 0 0 6px 0; border-bottom: 1px dashed #c7d2fe; letter-spacing: 0.3px;",
"p": "margin: 14px 0; text-indent: 0; text-align: justify; color: #242832; line-height: 1.85;",
"blockquote": "margin: 22px 0; padding: 18px 22px; background: #f4f6fc; border-left: 3px solid #2e5bff; color: #3a4463; font-size: 14.5px; border-radius: 0 10px 10px 0; line-height: 1.85; letter-spacing: 0.4px;",
"img": "max-width: 100%; height: auto; border-radius: 8px; margin: 22px auto; display: block; box-shadow: 0 6px 20px rgba(11,21,48,0.08); border: 1px solid #eef2ff;",
"strong": "color: #0b1530; font-weight: 800; background: linear-gradient(180deg, transparent 62%, #fff18a 62%, #fff18a 95%, transparent 95%); padding: 0 2px;",
"em": "font-style: italic; color: #5a6577;",
"code_inline": "background: #eef2ff; color: #2e3a8a; padding: 2px 8px; border-radius: 4px; font-size: 13px; font-family: 'JetBrains Mono', 'Menlo', 'Consolas', 'SF Mono', monospace; border: 1px solid #dbe4ff;",
"code_block": "background: #0f1729; color: #e1e7ff; padding: 20px 22px; border-radius: 10px; font-size: 13px; line-height: 1.7; font-family: 'JetBrains Mono', 'Menlo', 'Consolas', 'SF Mono', monospace; overflow-x: auto; white-space: pre-wrap; word-wrap: break-word; margin: 22px 0; border: 1px solid #1e2a4a;",
"ul": "margin: 16px 0; padding-left: 4px; list-style-type: none;",
"ol": "margin: 18px 0 18px 8px; padding-left: 28px; list-style-type: decimal; list-style-position: outside; color: #2e5bff; font-weight: 700;",
"li": "margin: 10px 0; line-height: 1.8; padding-left: 8px; color: #242832; font-weight: 400; min-height: 22px;",
"hr": "border: none; height: 1px; background: linear-gradient(to right, transparent, #c7d2fe, transparent); margin: 36px 0;",
"a": "color: #2e5bff; text-decoration: none; border-bottom: 1px solid rgba(46,91,255,0.35); padding-bottom: 1px;",
"table": "width: 100%; border-collapse: collapse; margin: 22px 0; font-size: 14px; border-radius: 8px; overflow: hidden; border: 1px solid #e8ecf7;",
"th": "background: #0b1530; color: #ffffff; padding: 12px 14px; text-align: left; font-weight: 600; font-size: 13px; letter-spacing: 0.3px;",
"td": "padding: 11px 14px; border-bottom: 1px solid #eef2ff; color: #242832;",
"caption": "font-size: 12.5px; color: #8a93a8; text-align: center; margin-top: 10px; letter-spacing: 0.5px; font-style: italic;",
"section_divider": "text-align: center; color: #c7d2fe; letter-spacing: 14px; font-size: 11px; margin: 36px 0 30px; user-select: none;",
}
DEFAULT_HIGHLIGHTS = {
"hl_yellow": "background: linear-gradient(180deg, transparent 55%, #fff18a 55%); color: #242832; padding: 0 3px; font-weight: 600;",
"hl_blue": "background: #dbe4ff; color: #1e2a8a; padding: 1px 5px; border-radius: 3px; font-weight: 500;",
"hl_pink": "background: #ffe0ec; color: #a6115a; padding: 1px 5px; border-radius: 3px; font-weight: 500;",
"hl_green": "background: #d7f5e2; color: #167a44; padding: 1px 5px; border-radius: 3px; font-weight: 500;",
"em_red": "color: #c43a30; font-weight: 700;",
"em_blue": "color: #2e5bff; font-weight: 700;",
"em_orange": "color: #ea7c1c; font-weight: 700;",
}
DEFAULT_SECTION_DIVIDER_TEXT = "● ● ●"
# 默认 list_style:让旧主题(没写 list_style 的)也能用统一的渲染逻辑。
# 渲染器会先把数字/项目符号包成 <span>,样式跟着主题走;旧主题的 ol/ul/li
# 还在,只是多出来一个 prefix span,视觉影响极小。
DEFAULT_LIST_STYLE = {
"num_container": "color: #4a6cf7; font-weight: 700; margin-right: 8px;",
"num_prefix": "",
"num_suffix": ".",
"num_formatter": "decimal",
"bullet_container": "color: #4a6cf7; margin-right: 8px;",
"bullet_char": "●",
}
def _themes_dir():
return Path(__file__).parent.parent / "assets" / "themes"
def list_themes():
"""列出可用主题名称"""
d = _themes_dir()
if not d.exists():
return []
return sorted(p.stem for p in d.glob("*.json"))
def load_theme(theme_name=None, style_path=None):
"""
加载一个完整主题(styles + highlights + section_divider_text + list_style)。
参数优先级:
1. style_path(显式指定 JSON 文件)
2. theme_name(查 assets/themes/<name>.json)
3. 内置 DEFAULT_STYLES / DEFAULT_HIGHLIGHTS / DEFAULT_LIST_STYLE
返回:
(styles: dict, highlights: dict, divider_text: str, list_style: dict)
注:为向后兼容,如果调用方按 3 元组解包(styles, highlights, divider),
多出来的 list_style 会被忽略。convert_markdown_to_wechat_html 自身
会从默认值兜底,没传 list_style 不会出问题。
"""
styles = DEFAULT_STYLES.copy()
highlights = DEFAULT_HIGHLIGHTS.copy()
divider_text = DEFAULT_SECTION_DIVIDER_TEXT
list_style = DEFAULT_LIST_STYLE.copy()
candidates = []
if style_path:
candidates.append(Path(style_path))
if theme_name:
candidates.append(_themes_dir() / f"{theme_name}.json")
for path in candidates:
if path and path.exists():
try:
with open(path, "r", encoding="utf-8") as f:
data = json.load(f)
styles.update(data.get("styles", {}))
highlights.update(data.get("highlights", {}))
if data.get("section_divider_text"):
divider_text = data["section_divider_text"]
if data.get("list_style"):
list_style.update(data["list_style"])
break
except (json.JSONDecodeError, KeyError):
continue
return styles, highlights, divider_text, list_style
def load_styles(style_path=None):
"""向后兼容:老接口只返回 styles。"""
styles, _, _, _ = load_theme(style_path=style_path)
return styles
# ---------- 数字格式化(给 list_style.num_formatter 用) ----------
_CN = ["", "一", "二", "三", "四", "五", "六", "七", "八", "九", "十"]
_ROMAN_U = ["", "Ⅰ", "Ⅱ", "Ⅲ", "Ⅳ", "Ⅴ", "Ⅵ", "Ⅶ", "Ⅷ", "Ⅸ", "Ⅹ"]
_ROMAN_L = ["", "i", "ii", "iii", "iv", "v", "vi", "vii", "viii", "ix", "x"]
_CIRCLED = ["", "①", "②", "③", "④", "⑤", "⑥", "⑦", "⑧", "⑨", "⑩"]
_CIRCLED_FILLED = ["", "❶", "❷", "❸", "❹", "❺", "❻", "❼", "❽", "❾", "❿"]
def _format_num(n: int, fmt: str) -> str:
if fmt == "padded": return f"{n:02d}"
if fmt == "decimal": return str(n)
if fmt == "chinese": return _CN[n] if 0 < n < len(_CN) else str(n)
if fmt == "roman_upper": return _ROMAN_U[n] if 0 < n < len(_ROMAN_U) else str(n)
if fmt == "roman_lower": return _ROMAN_L[n] if 0 < n < len(_ROMAN_L) else str(n)
if fmt == "circled": return _CIRCLED[n] if 0 < n < len(_CIRCLED) else str(n)
if fmt == "circled_filled": return _CIRCLED_FILLED[n] if 0 < n < len(_CIRCLED_FILLED) else str(n)
return str(n)
# ============================================================
# Markdown 解析与转换
# ============================================================
# 行内扩展标记(有序,先处理长模式)
INLINE_MARKS = [
# (正则, highlight_key)
(r'==([^=\n]+)==', "hl_yellow"),
(r'\+\+([^+\n]+)\+\+', "hl_blue"),
(r'%%([^%\n]+)%%', "hl_pink"),
(r'&&([^&\n]+)&&', "hl_green"),
(r'!!([^!\n]+)!!', "em_red"),
(r'@@([^@\n]+)@@', "em_blue"),
(r'\^\^([^\^\n]+)\^\^', "em_orange"),
]
def convert_markdown_to_wechat_html(
markdown_text, styles=None, highlights=None, divider_text=None, list_style=None
):
"""
将 Markdown 文本转换为微信公众号兼容的 HTML。
Args:
markdown_text: Markdown 格式的文章内容
styles: 标签样式字典(默认使用内置)
highlights: 行内标记样式字典(默认使用内置)
divider_text: 分节符文本
list_style: 主题的 list_style 块,控制 ol/ul 的序号 / 项目符号样式
Returns:
str: 微信兼容的 HTML 字符串
"""
if styles is None:
styles = DEFAULT_STYLES.copy()
if highlights is None:
highlights = DEFAULT_HIGHLIGHTS.copy()
if divider_text is None:
divider_text = DEFAULT_SECTION_DIVIDER_TEXT
if list_style is None:
list_style = DEFAULT_LIST_STYLE.copy()
lines = markdown_text.split("\n")
html_parts = []
in_code_block = False
code_block_content = []
in_list = False
list_type = None
list_items = []
in_blockquote = False
blockquote_lines = []
in_table = False
table_rows = []
h1_seen = False
prev_was_h2 = False
# ---- list_style 渲染辅助 ----
def _num_token(n: int) -> str:
prefix = list_style.get("num_prefix", "")
suffix = list_style.get("num_suffix", "")
fmt = list_style.get("num_formatter", "decimal")
container = list_style.get("num_container", "")
return f'<span style="{container}">{prefix}{_format_num(n, fmt)}{suffix}</span>'
def _bullet_token() -> str:
glyph = list_style.get("bullet_char", "•") or " "
container = list_style.get("bullet_container", "")
return f'<span style="{container}">{glyph}</span>'
def flush_list():
nonlocal in_list, list_items, list_type
if in_list and list_items:
tag = list_type or "ul"
if tag == "ul":
items_html = "\n".join(
f'<li style="{styles["li"]}">{_bullet_token()}{item}</li>'
for item in list_items
)
else:
items_html = "\n".join(
f'<li style="{styles["li"]}">{_num_token(n)}{item}</li>'
for n, item in enumerate(list_items, 1)
)
html_parts.append(f'<{tag} style="{styles[tag]}">{items_html}</{tag}>')
list_items = []
in_list = False
list_type = None
def flush_blockquote():
nonlocal in_blockquote, blockquote_lines
if in_blockquote and blockquote_lines:
content = "<br>".join(blockquote_lines)
html_parts.append(f'<blockquote style="{styles["blockquote"]}">{content}</blockquote>')
blockquote_lines = []
in_blockquote = False
def flush_table():
nonlocal in_table, table_rows
if in_table and table_rows:
rows_html = []
for i, row in enumerate(table_rows):
cells = [c.strip() for c in row.split("|")[1:-1]]
if i == 0:
cells_html = "".join(f'<th style="{styles["th"]}">{c}</th>' for c in cells)
rows_html.append(f"<tr>{cells_html}</tr>")
elif all(set(c.strip()) <= {"-", ":"} for c in cells):
continue
else:
cells_html = "".join(f'<td style="{styles["td"]}">{c}</td>' for c in cells)
rows_html.append(f"<tr>{cells_html}</tr>")
html_parts.append(f'<table style="{styles["table"]}">{"".join(rows_html)}</table>')
table_rows = []
in_table = False
def _escape_html(text):
return text.replace("&", "&").replace("<", "<").replace(">", ">")
def process_inline(text):
"""处理行内 Markdown 语法 + 扩展标记。
安全处理:
- 先转义 <,> 以阻止原始 HTML/JS (<script> 之类)透传
(& 不在这一步转义,因为自定义标记 &&...&& 会被破坏;
所有由我们自己生成的 style 属性都不含原始 &,所以无需
整体转义,只在最后把游离的 & 转成 &)
- 链接 [text](url) 的 url 若以 javascript: / data: / vbscript: 开头
(忽略大小写和前导空白),直接退化为纯文本
- 图片  的 alt/url 经 html.escape(quote=True) 处理
"""
# 1. 先把原始 <,> 转义掉 ── 禁止 <script> 透传
text = text.replace("<", "<").replace(">", ">")
# 2. 图片  —— alt/url 完整转义,防止属性注入
def _img_sub(m):
alt = html.escape(m.group(1), quote=True)
src = html.escape(m.group(2), quote=True)
return f'<img src="{src}" alt="{alt}" style="{styles["img"]}"/>'
text = re.sub(r'!\[([^\]]*)\]\(([^)]+)\)', _img_sub, text)
# 3. 链接 [text](url) —— 拒绝危险协议
_DANGEROUS_URL = re.compile(r'^\s*(javascript|data|vbscript)\s*:', re.IGNORECASE)
def _link_sub(m):
link_text = m.group(1)
url = m.group(2).strip()
if _DANGEROUS_URL.match(url):
# 不生成 <a>,仅保留纯文本
return link_text
safe_url = html.escape(url, quote=True)
return f'<a href="{safe_url}" style="{styles["a"]}">{link_text}</a>'
text = re.sub(r'\[([^\]]+)\]\(([^)]+)\)', _link_sub, text)
# 4. 行内代码 `code`
text = re.sub(
r'`([^`]+)`',
lambda m: f'<code style="{styles["code_inline"]}">{_escape_html(m.group(1))}</code>',
text,
)
# 5. 扩展行内标记(==/++/%%/&&/!!/@@/^^)—— 要在 ** 之前处理
for pattern, key in INLINE_MARKS:
style_str = highlights.get(key, "")
if not style_str:
continue
text = re.sub(
pattern,
lambda m, s=style_str: f'<span style="{s}">{m.group(1)}</span>',
text,
)
# 6. 粗体 **text**
text = re.sub(
r'\*\*([^*\n]+)\*\*',
lambda m: f'<strong style="{styles["strong"]}">{m.group(1)}</strong>',
text,
)
# 7. 斜体 *text*
text = re.sub(
r'(?<!\*)\*([^*\n]+)\*(?!\*)',
lambda m: f'<em style="{styles["em"]}">{m.group(1)}</em>',
text,
)
# 8. 收尾:把游离的 & 转成 &(不影响已生成的 </>/&/&#...)
text = re.sub(r'&(?!(?:[a-zA-Z][a-zA-Z0-9]{1,31}|#[0-9]{1,7}|#x[0-9a-fA-F]{1,6});)', '&', text)
return text
for line in lines:
stripped = line.strip()
# 代码块
if stripped.startswith("```"):
if in_code_block:
code_html = "\n".join(code_block_content)
html_parts.append(f'<pre style="{styles["code_block"]}">{code_html}</pre>')
code_block_content = []
in_code_block = False
else:
flush_list()
flush_blockquote()
flush_table()
in_code_block = True
prev_was_h2 = False
continue
if in_code_block:
code_block_content.append(_escape_html(line))
continue
# 表格
if "|" in stripped and stripped.startswith("|"):
flush_list()
flush_blockquote()
if not in_table:
in_table = True
table_rows.append(stripped)
prev_was_h2 = False
continue
else:
flush_table()
# 引用块
if stripped.startswith(">"):
flush_list()
if not in_blockquote:
in_blockquote = True
content = stripped.lstrip(">").strip()
blockquote_lines.append(process_inline(content))
prev_was_h2 = False
continue
else:
flush_blockquote()
# 自定义分节符: 单独一行的 `===` / `~~~` / `[SEC]`
# `[SEC]` 是新主题包(2026)推荐写法,等价于旧的 `===`。
if re.match(r'^(={3,}|~{3,})$', stripped) or stripped == "[SEC]":
flush_list()
html_parts.append(
f'<p style="{styles.get("section_divider", "")}">{divider_text}</p>'
)
prev_was_h2 = False
continue
# 标题
h_match = re.match(r'^(#{1,3})\s+(.+)$', stripped)
if h_match:
flush_list()
level = len(h_match.group(1))
text = process_inline(h_match.group(2))
tag = f"h{level}"
if level == 1 and not h1_seen:
h1_seen = True
prev_was_h2 = False
continue
# 如果两个 h2 之间已经有正文,在下一个 h2 前插一个分节符
if level == 2 and html_parts and not prev_was_h2:
html_parts.append(
f'<p style="{styles.get("section_divider", "")}">{divider_text}</p>'
)
html_parts.append(f'<{tag} style="{styles[tag]}">{text}</{tag}>')
prev_was_h2 = (level == 2)
continue
# 分割线
if re.match(r'^(-{3,}|\*{3,}|_{3,})$', stripped):
flush_list()
html_parts.append(f'<hr style="{styles["hr"]}"/>')
prev_was_h2 = False
continue
# 无序列表
ul_match = re.match(r'^[-*+]\s+(.+)$', stripped)
if ul_match:
if not in_list or list_type != "ul":
flush_list()
in_list = True
list_type = "ul"
list_items.append(process_inline(ul_match.group(1)))
prev_was_h2 = False
continue
# 有序列表
ol_match = re.match(r'^\d+\.\s+(.+)$', stripped)
if ol_match:
if not in_list or list_type != "ol":
flush_list()
in_list = True
list_type = "ol"
list_items.append(process_inline(ol_match.group(1)))
prev_was_h2 = False
continue
flush_list()
if not stripped:
continue
# 纯图片行 —— alt/src 必须转义,避免 "><script> 这种属性注入
img_match = re.match(r'^!\[([^\]]*)\]\(([^)]+)\)$', stripped)
if img_match:
alt_raw = img_match.group(1)
src_raw = img_match.group(2)
alt = html.escape(alt_raw, quote=True)
src = html.escape(src_raw, quote=True)
html_parts.append(f'<p style="text-align:center;margin:16px 0;"><img src="{src}" alt="{alt}" style="{styles["img"]}"/></p>')
if alt_raw:
html_parts.append(f'<p style="{styles["caption"]}">{alt}</p>')
prev_was_h2 = False
continue
# 普通段落
text = process_inline(stripped)
html_parts.append(f'<p style="{styles["p"]}">{text}</p>')
prev_was_h2 = False
flush_list()
flush_blockquote()
flush_table()
# #23: 未闭合的代码块 —— 尽力把已积累的内容按代码块渲染,避免默默丢失内容
if in_code_block and code_block_content:
print("警告: 未闭合的 code block,已按代码块渲染", file=sys.stderr)
code_html = "\n".join(code_block_content)
html_parts.append(f'<pre style="{styles["code_block"]}">{code_html}</pre>')
code_block_content = []
in_code_block = False
elif in_code_block:
# 代码块开了但一行内容都没有 —— 也提醒一下
print("警告: 未闭合的 code block,已按代码块渲染", file=sys.stderr)
in_code_block = False
body = "\n".join(html_parts)
return f'<section style="{styles["body"]}">\n{body}\n</section>'
# ============================================================
# 命令行入口
# ============================================================
if __name__ == "__main__":
import argparse
parser = argparse.ArgumentParser(description="Markdown → 微信公众号 HTML 转换器")
parser.add_argument("input", help="Markdown 文件路径")
parser.add_argument("-o", "--output", help="输出 HTML 文件路径(默认 stdout)")
parser.add_argument("--theme", help=f"主题名,可选: {', '.join(list_themes()) or '无'}")
parser.add_argument("--style", help="自定义样式 JSON 文件路径(覆盖 --theme)")
parser.add_argument("--list-themes", action="store_true", help="列出所有可用主题")
args = parser.parse_args()
if args.list_themes:
themes = list_themes()
if not themes:
print("未找到主题(assets/themes/ 为空)")
else:
for t in themes:
print(t)
sys.exit(0)
with open(args.input, "r", encoding="utf-8") as f:
md_text = f.read()
styles, highlights, divider_text, list_style = load_theme(
theme_name=args.theme, style_path=args.style
)
html = convert_markdown_to_wechat_html(
md_text, styles, highlights, divider_text, list_style
)
if args.output:
with open(args.output, "w", encoding="utf-8") as f:
f.write(html)
print(f"转换完成: {args.output}")
else:
print(html)
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |