by 安全扫描
OpenClaw
安全
high confidenceThe skill's instructions, required actions, and permissions are consistent with its stated purpose of fetching, summarizing, tagging, and storing WeChat public-account articles, with only minor implementation mismatches to note.
评估建议
This skill appears to do what it says: scrape mp.weixin.qq.com article pages, extract title/author/body, generate a 1–2 sentence summary and 1–2 tags, and append the record to memory/YYYY-MM-DD.md for daily reports. Before installing, consider: (1) Ensure the agent environment has curl and python3 available (the SKILL.md uses them but metadata doesn't list required binaries). (2) Be aware fetched article contents are stored in memory files — if you forward private or sensitive links, their conte...详细分析 ▾
ℹ 用途与能力
The SKILL.md clearly implements WeChat article fetching, tagging, summarization, and appending entries to memory files — which matches the description. One inconsistency: the runtime instructions call curl and python3 (via exec), but the skill metadata lists no required binaries. Either the environment must already provide curl/python3 or the metadata should declare them.
ℹ 指令范围
Instructions are narrowly scoped to fetching mp.weixin.qq.com article pages, extracting title/author/body, generating summaries and tags, and writing per-day markdown files in memory/. This matches the purpose. Items to be aware of: the skill executes shell commands (curl) and a Python extraction snippet, and it writes full article text into memory/YYYY-MM-DD.md — that is expected but means fetched content will persist in agent memory.
✓ 安装机制
There is no install spec or downloaded code; the skill is instruction-only, so nothing is written to disk by an installer. This is the lowest-risk install model.
✓ 凭证需求
The skill requests no environment variables, credentials, or external configuration paths. All required actions (HTTP fetch, local parsing, memory write) are proportional to its purpose.
✓ 持久化与权限
always is false and the skill writes only to its own memory/YYYY-MM-DD.md files as described. It does not request elevated or system-wide privileges. Note: allowed-tools includes exec, read, write, edit which permits running commands and modifying agent memory — appropriate for this skill but worth confirming you are comfortable with those tool permissions.
安全有层次，运行前请审查代码。
运行时依赖
无特殊依赖
版本
latestv1.0.02026/4/2
Initial release: fetch WeChat articles via curl+regex, auto summarize and categorize, support daily digest
● 可疑
安装命令
点击复制官方npx clawhub@latest install wechat-article-digest
镜像加速npx clawhub@latest install wechat-article-digest --registry https://cn.longxiaskill.com
技能文档
核心流程
1. 抓取文章内容
用户发送 mp.weixin.qq.com/s/ 链接时：
curl -s -L \
-H "User-Agent: Mozilla/5.0 (iPhone; CPU iPhone OS 17_0 like Mac OS X) AppleWebKit/605.1.15" \
"链接" \
| python3 -c "
import sys, re
html = sys.stdin.read()
# 提取标题
t = re.search(r'var msg_title = \"(.?)\";', html)
title = t.group(1) if t else ''
# 提取公号名
n = re.search(r'var nickname = \"(.?)\";', html)
name = n.group(1) if n else ''
# 提取正文
m = re.search(r'id=\"js_content\"[^>]>(.?)', html, re.DOTALL)
body = re.sub(r'<[^>]+>', '', m.group(1)).strip() if m else ''
print(f'TITLE:{title}')
print(f'NAME:{name}')
print(f'BODY:{body}')
"
如果标题和正文都为空，用 web_fetch 作为备选。
2. 分类标签体系
每篇文章分配 1-2 个标签：
- 🤖 AI/科技 — AI、大模型、技术产品
- 💰 商业/创业 — 创业、商业模式、融资
- 📈 职场/管理 — 职场经验、团队管理、效率
- 📊 行业分析 — 行业趋势、市场分析、竞品
- 🧠 个人成长 — 学习方法、思维方式、认知
- 💰 投资/理财 — 股市、基金、财务规划
- 🌍 宏观/政策 — 经济政策、国际局势
- 🎯 产品/设计 — 产品思维、用户体验
- 📖 读书/文化 — 书评、文化、人文
- 🔧 工具/效率 — 工具推荐、自动化、工作流
- ❤️ 健康/生活 — 健康、生活、情感
3. 单篇响应格式
用户发单篇链接时，立即响应：
📌 [标题]
📁 公号：[名称] | 标签：[标签]
📝 摘要：[1-2句话提炼核心观点]
同时将文章信息追加到 memory/YYYY-MM-DD.md。
4. 日报生成格式
当触发日报汇总时（定时任务或用户说"文章日报"），读取当天 memory 文件，生成：
📰 今日阅读日报 — YYYY-MM-DD共 N 篇文章
[标签emoji] [标签名]（X篇）
├ 📌 标题1 — 公号名 | 一句话摘要
├ 📌 标题2 — 公号名 | 一句话摘要
└ ...
[标签emoji] [标签名]（X篇）
└ ...
💡 今日要点：
- [从所有文章中提炼 2-3 个值得关注的观点]
如果当天没有文章，回复 NO_REPLY。
5. 记录格式
在 memory/YYYY-MM-DD.md 中以如下格式记录每篇文章：
## 公众号文章- [标签emoji] 标题 — 公号名
> 摘要内容
> 链接：https://mp.weixin.qq.com/s/xxx
注意事项
- 正文超过 5000 字时，摘要控制在 3 句话以内，突出核心观点
- 如果抓取失败，告知用户并建议复制文字或截图
- 分类标签可扩展，遇到无法归类的文章放在「📌 其他」下