ai-news-daily-brief 安全扫描报告 — 可信 | ClawSafe

0 /100

ai-news-daily-brief

从指定信源检索最近24小时AI新闻，完成去重/分类/摘要，生成结构化AI新闻日报

This skill is a pure-documentation news aggregation tool with no executable code, no scripts, no credential access, and no network exfiltration. All behavior is fully declared in documentation.

技能名称ai-news-daily-brief

分析耗时22.9s

引擎pi

✓

可以安装

No action needed. This skill is safe for use.

资源类型	声明权限	推断权限	状态	证据
文件系统	`NONE`	`NONE`	—	No file operations in any file
网络访问	`READ`	`READ`	✓ 一致	web_fetch and tavily for reading news sources (documented in SKILL.md)
命令执行	`NONE`	`NONE`	—	No bash/shell commands in any file
环境变量	`NONE`	`NONE`	—	No env var access
技能调用	`NONE`	`NONE`	—	No cross-skill invocations
剪贴板	`NONE`	`NONE`	—	No clipboard access
浏览器	`READ`	`READ`	✓ 一致	agent-browser for Chinese sites (documented in SKILL.md and sources.md)
数据库	`NONE`	`NONE`	—	No database access

17 项发现

🔗

中危外部 URL 外部 URL

https://news.aibase.com/zh/daily）

SKILL.md:38

🔗

中危外部 URL 外部 URL

https://docs.qq.com/aio/DUEtYTUNWaVNxeXN4

SKILL.md:372

🔗

中危外部 URL 外部 URL

https://docs.qq.com/aio/DUG9ocVBZU01sRmtG

SKILL.md:373

🔗

中危外部 URL 外部 URL

https://news.aibase.com/zh/daily

references/sources.md:31

🔗

中危外部 URL 外部 URL

https://36kr.com/

references/sources.md:40

🔗

中危外部 URL 外部 URL

https://www.jiqizhixin.com/

references/sources.md:48

🔗

中危外部 URL 外部 URL

https://www.qbitai.com/

references/sources.md:56

🔗

中危外部 URL 外部 URL

https://www.reuters.com/technology/artificial-intelligence/

references/sources.md:68

🔗

中危外部 URL 外部 URL

https://techcrunch.com/category/artificial-intelligence/

references/sources.md:77

🔗

中危外部 URL 外部 URL

https://www.theinformation.com/topics/artificial-intelligence

references/sources.md:86

🔗

中危外部 URL 外部 URL

https://www.theverge.com/ai-artificial-intelligence

references/sources.md:95

🔗

中危外部 URL 外部 URL

https://www.wired.com/tag/artificial-intelligence/

references/sources.md:103

🔗

中危外部 URL 外部 URL

https://venturebeat.com/category/ai/

references/sources.md:111

🔗

中危外部 URL 外部 URL

https://deepmind.google/discover/blog/

references/sources.md:123

🔗

中危外部 URL 外部 URL

https://www.anthropic.com/news

references/sources.md:137

🔗

中危外部 URL 外部 URL

https://www.anthropic.com/engineering

references/sources.md:144

🔗

中危外部 URL 外部 URL

https://blogs.nvidia.com/

references/sources.md:151

目录结构

7 文件 · 53.3 KB · 1526 行

Markdown 7f · 1526L

├─ ▾ 📁 references

│ ├─ 📝 category-taxonomy.md Markdown 280L · 9.0 KB

│ ├─ 📝 output-template.md Markdown 120L · 4.0 KB

│ ├─ 📝 search-rules.md Markdown 300L · 10.4 KB

│ ├─ 📝 sources.md Markdown 219L · 8.6 KB

│ └─ 📝 validation-checklist.md Markdown 53L · 1.6 KB

├─ 📝 README.md Markdown 172L · 3.9 KB

└─ 📝 SKILL.md Markdown 382L · 15.7 KB

安全亮点

✓ No executable code — only Markdown documentation files

✓ No scripts, binaries, or dependencies (no requirements.txt, package.json, etc.)

✓ No credential, API key, or token harvesting

✓ No shell or subprocess execution of any kind

✓ All network access is clearly declared (web_fetch, tavily, agent-browser) for news aggregation

✓ Strict source whitelist prevents abuse through undeclared sources

✓ No data exfiltration or outbound transmission of sensitive data

✓ Output target (Tencent Docs) is declared and documented

✓ No base64 encoding, obfuscation, or anti-analysis patterns

✓ No access to sensitive local paths (~/.ssh, ~/.aws, .env, etc.)