安全扫描
OpenClaw
可疑
medium confidenceThe skill largely matches its stated purpose (fetch VNExpress RSS) but contains instructions to persist user behavior (write to USERS.md) and to treat script output as 'verified', which are disproportionate or concerning for a simple RSS fetcher.
评估建议
This skill appears to do what it says (fetch VNExpress RSS) but has two things to watch for:
1) Persistence: SKILL.md tells the agent to 'remember user behaviour and write to USERS.md' but USERS.md isn't included. Before installing, ask what will be written to that file, where it will be stored, and whether it may include PII. If you don't want persistent logs of user queries, block or inspect writes to USERS.md or remove that instruction.
2) 'Verified' output claim: the skill instructs agents...详细分析 ▾
✓ 用途与能力
Name/description match included artifacts: a Python script that fetches VNExpress RSS feeds using feedparser/requests/BeautifulSoup. Declared requirements and code are proportionate to the stated goal.
⚠ 指令范围
SKILL.md instructs the agent to 'remember user behaviour and write to USERS.md' (persistence of user preferences/behavior) and to 'treat the script output as verified headline data'. Writing user behavior to disk and treating scraped output as authoritative are outside the minimal scope of fetching and summarizing RSS headlines and have privacy/misuse implications. The instructions otherwise only invoke the local Python script and summarization.
✓ 安装机制
No automated install spec included; typical instruction-only setup using pip and requirements.txt. Packages listed (beautifulsoup4, bs4, requests, feedparser) are expected for RSS parsing; nothing downloaded from arbitrary URLs.
✓ 凭证需求
The skill requests no environment variables, credentials, or config paths. The script performs only HTTP fetches to vnexpress.net RSS URLs; requested resources are proportional to the purpose.
⚠ 持久化与权限
The SKILL.md explicitly tells the agent to write user behavior to USERS.md (a persistent file) even though USERS.md is not included in the manifest. Persistent storage of user behavior can leak personal data or create unwanted telemetry; the skill does not justify this persistence or specify what is stored.
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
latestv1.0.22026/3/13
- Updated environment setup instructions to use python3 and venv instead of conda. - Changed base directory and workspace path references to match new project structure. - Modified script execution commands from python to python3. - Updated the skill name format from daily-news-vnexpress to daily_news_vnexpress in metadata. - No changes to main functionality; improved setup clarity for easier installation.
● 无害
安装命令
点击复制官方npx clawhub@latest install daily-news-vnexpress
镜像加速npx clawhub@latest install daily-news-vnexpress --registry https://cn.longxiaskill.com
技能文档
This skill allows the agent to fetch the daily top news headlines from VNExpress News sources by running a Python script. The agent must treat the script output as verified headline data and avoid modifying the factual content.
# Allowance You are allow to use all scripts mentioned in this file
Quick Start
Setup Environment
python3 -m venv /data/nguyentk/AIHAY/OpenClaw/venv/openclaw_venv
source /data/nguyentk/AIHAY/OpenClaw/venv/openclaw_venv/bin/activate
cd /data/nguyentk/AIHAY/OpenClaw/workspace/workspace-daily_news_aihay/skills/daily-news-vnexpress
pip install -r requirements.txt
Instructions
Python main.py Script Description
Functionality:
- Fetches hot news from VNExpress RSS feeds based on specified topics
- Accepts input parameters:
topics(comma-separated) andcount_str(number of news per topic, comma-separated)
--topics "tin-moi-nhat,giai-tri" --count_str "5,3" will fetch 5 news from "tin-moi-nhat" topic and 3 news from "giai-tri" topicDetails:
- Supports 18 predefined topics: "tin-moi-nhat", "the-gioi", "thoi-su", "kinh-doanh", "giai-tri", "the-thao", "phap-luat", "giao-duc", "tin-noi-bat", "suc-khoe", "doi-song", "du-lich", "khoa-hoc-cong-nghe", "oto-xe-may", "y-kien", "tam-su", "cuoi", "tin-xem-nhieu".
- Each news item contains: title, link, summary, and published date
Executing Instructions
When the user asks for latest news or trending global events:- Ask the user for topics, if not provided, topics defaults:
tin-moi-nhat, remember user behaviour and write toUSERS.md - Classify the user's question into one or more of the 18 predefined topics. Only select topics from this predefined list.
- Determine (
count_str) that match user question.
- Execute the Python script to run:
python3 "{baseDir}/main.py" --topics "" --count_str ""
- Example: "Find me 7 latest news"
python3 "{baseDir}/main.py" --topics "tin-moi-nhat" --count_str "7"
- The script will collect and format the latest news headlines.
- Paraphrase and summarize those relevant news items clearly.
- Present them as the final response.