paper-reading 安全扫描报告 — 低风险 | ClawSafe

15 /100

paper-reading

Read academic papers from local PDF files, arXiv URLs, or paper titles and generate structured reading notes in Chinese.

A legitimate academic paper reading skill with minor documentation flaws (hardcoded user paths, unversioned dependency) but no security violations.

技能名称paper-reading

分析耗时41.1s

引擎pi

✓

可以安装

Update hardcoded paths in SKILL.md to use relative paths or environment variables. Pin pdfplumber version to ensure reproducibility.

安全发现 3 项

严重性	安全发现	位置
低危	Hardcoded user-specific paths in SKILL.md 文档欺骗 SKILL.md contains hardcoded paths with /Users/yihe/ which will not work for other users. These should use relative paths or document that users need to customize. `python /Users/yihe/.config/opencode/skills/paper-reading/scripts/fetch_pdf.py` → Replace hardcoded paths with relative paths or document path customization.	`SKILL.md:17`
低危	Unversioned dependency 供应链 pip install pdfplumber has no version constraint. This could lead to unexpected behavior if a breaking change is released. `pip install pdfplumber # if not already installed` → Pin to a specific version: pip install pdfplumber==0.11.0	`SKILL.md:23`
提示	HTTP endpoint for arXiv API 文档欺骗 The script uses http://export.arxiv.org/api/query instead of HTTPS. While arXiv is a legitimate academic resource, HTTPS would be more secure. `ARXIV_API = "http://export.arxiv.org/api/query"` → Consider using HTTPS: https://export.arxiv.org/api/query	`scripts/fetch_pdf.py:25`

资源类型	声明权限	推断权限	状态	证据
文件系统	`READ`	`READ`	✓ 一致	SKILL.md:17 - Validates local PDF existence
网络访问	`READ`	`READ`	✓ 一致	scripts/fetch_pdf.py:25 - Downloads PDFs from arXiv.org
命令执行	`WRITE`	`WRITE`	✓ 一致	SKILL.md:23 - Executes python scripts for PDF extraction
环境变量	`NONE`	`NONE`	—	No environment variable access found

6 项发现

🔗

中危外部 URL 外部 URL

https://cdn.nlark.com/yuque/0/2026/png/57569306/1767535158747-34ab19e4-bec9-4f80-bb38-e36d876b77d9.png

references/mHC.md:30

🔗

中危外部 URL 外部 URL

https://cdn.nlark.com/yuque/0/2026/png/57569306/1767535744001-6619f6a6-10a5-4434-afe2-da36ff2d58af.png

references/mHC.md:304

🔗

中危外部 URL 外部 URL

http://export.arxiv.org/api/query

scripts/fetch_pdf.py:25

🔗

中危外部 URL 外部 URL

https://arxiv.org/pdf/

scripts/fetch_pdf.py:40

🔗

中危外部 URL 外部 URL

http://www.w3.org/2005/Atom

scripts/fetch_pdf.py:62

🔗

中危外部 URL 外部 URL

http://arxiv.org/schemas/atom

scripts/fetch_pdf.py:62

目录结构

5 文件 · 24.9 KB · 672 行

Markdown 3f · 455L Python 2f · 217L

├─ ▾ 📁 references

│ ├─ 📝 mHC.md Markdown 323L · 13.6 KB

│ └─ 📝 note-template.md Markdown 50L · 1.7 KB

├─ ▾ 📁 scripts

│ ├─ 🐍 fetch_pdf.py Python 141L · 4.2 KB

│ └─ 🐍 read_pdf.py Python 76L · 2.4 KB

└─ 📝 SKILL.md Markdown 82L · 3.0 KB

依赖分析 1 项

包名	版本	来源	已知漏洞	备注
`pdfplumber`	`*`	pip	否	Version not pinned

安全亮点

✓ No credential theft or harvesting detected

✓ No data exfiltration to external servers

✓ No obfuscation techniques (base64, eval, exec) found

✓ No subprocess abuse or shell command injection vectors

✓ Code is readable and straightforward with clear purpose

✓ Network access is limited to legitimate arXiv paper downloads

✓ No sensitive file access (no ~/.ssh, ~/.aws, .env access)

✓ No persistence mechanisms or backdoors

✓ No prompt injection attempts