skills-security-scanner 安全扫描报告 — 低风险 | ClawSafe

10 /100

skills-security-scanner

审计和扫描技能的安全性。在启用新技能前使用此工具验证其安全性，确保符合安全策略。

A legitimate security scanning tool that packages and uploads target skills to a Volcengine cloud API for analysis; no malicious behavior, credential theft, or hidden exfiltration found.

技能名称skills-security-scanner

分析耗时25.7s

引擎pi

✓

可以安装

Skill is safe for use. No action required. Consider pinning the `requests` library version for supply-chain hygiene.

安全发现 1 项

严重性	安全发现	位置
低危	Unpinned requests library 供应链 The `requests` library has no version pin in the code. While requests is used only for HTTP API calls and has no critical unfixed CVE for this usage, pinning to a specific version is a best practice. `import requests` → Add a version pin, e.g. `requests>=2.31.0`, to ensure reproducible and auditable dependencies.	`scripts/scan.py:1`

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:12 — targets a directory, reads it into a zip
网络访问	`READ`	`READ`	✓ 一致	scripts/scan.py:16,19 — POSTs zip to Volcengine API
命令执行	`NONE`	`NONE`	—	No shell execution found
环境变量	`NONE`	`READ`	✓ 一致	scripts/scan.py:280-287 — reads AK/SK/env vars to configure API calls; necessary…
技能调用	`NONE`	`NONE`	—	No skill invocation found
剪贴板	`NONE`	`NONE`	—	No clipboard access
浏览器	`NONE`	`NONE`	—	No browser usage
数据库	`NONE`	`NONE`	—	No database access

2 项发现

🔗

中危外部 URL 外部 URL

http://127.0.0.1:8181/api/top/mana-open/cn-north-1/2025-01-01

scripts/scan.py:16

🔗

中危外部 URL 外部 URL

https://open.volcengineapi.com

scripts/scan.py:19

目录结构

2 文件 · 17.2 KB · 425 行

Python 1f · 367L Markdown 1f · 58L

├─ ▾ 📁 scripts

│ └─ 🐍 scan.py Python 367L · 15.0 KB

└─ 📝 SKILL.md Markdown 58L · 2.2 KB

依赖分析 2 项

包名	版本	来源	已知漏洞	备注
`requests`	`unpinned`	pip	否	No version constraint; minor supply-chain concern
`volcengine`	`unpinned`	pip	否	Volcengine SDK used for API request signing; no version pin

安全亮点

✓ No arbitrary code execution — subprocess/bash not used

✓ No credential theft — AK/SK read from env vars only for signing outbound API calls to Volcengine

✓ No data exfiltration — only uploads the target skill zip (requested by the user) to the scan service

✓ No obfuscation — all code is plain Python, no base64 or eval patterns

✓ No sensitive path access (~/.ssh, ~/.aws, .env of the host) — only zips the user-specified target directory

✓ Network destination is Volcengine public cloud API (open.volcengineapi.com) — legitimate and declared

✓ Full code review shows clean, straightforward logic: zip → sign → upload → poll → return results