zhua-distributed 安全扫描报告 — 低风险 | ClawSafe

20 /100

zhua-distributed

爪爪分布式部署系统 —— 实现多实例协同、负载均衡、故障转移

A minimal, benign distributed system init script that creates local config files with no network, credential, or execution risks; the only concern is incomplete allowed-tools declaration in SKILL.md.

技能名称zhua-distributed

分析耗时33.1s

引擎pi

✓

可以安装

Add explicit allowed-tools declaration and filesystem:WRITE permission to SKILL.md for the config directory write operation to fully document the skill's capabilities.

安全发现 2 项

严重性	安全发现	位置
低危	SKILL.md missing allowed-tools declaration 文档欺骗 SKILL.md has no allowed-tools section. The script init_master.py performs filesystem:WRITE to ~/.zhua/distributed/ which is not formally declared. `No allowed-tools declaration found in frontmatter` → Add an allowed-tools declaration to SKILL.md frontmatter, e.g. allowed-tools: Read, Write, Bash	`SKILL.md:1`
低危	Placeholder scripts with no real functionality 文档欺骗 example.py is a placeholder with TODO comments. While not malicious, it indicates incomplete implementation. `TODO: Add actual script logic here` → Either implement meaningful functionality or remove placeholder scripts	`scripts/example.py:1`

资源类型	声明权限	推断权限	状态	证据
文件系统	`NONE`	`WRITE`	✗ 越权	scripts/init_master.py:21 — Path.home() / ".zhua" / "distributed" with mkdir par…
网络访问	`NONE`	`NONE`	—	No network activity found in any script
命令执行	`NONE`	`NONE`	—	No subprocess or shell execution in any script
环境变量	`NONE`	`NONE`	—	No os.environ access found
数据库	`NONE`	`NONE`	—	No database access found
剪贴板	`NONE`	`NONE`	—	No clipboard access found
浏览器	`NONE`	`NONE`	—	No browser access found
技能调用	`NONE`	`NONE`	—	No skill invocation found

目录结构

5 文件 · 6.4 KB · 216 行

Markdown 2f · 122L Python 2f · 70L Text 1f · 24L

├─ ▾ 📁 assets

│ └─ 📄 example_asset.txt Text 24L · 865 B

├─ ▾ 📁 references

│ └─ 📝 api_reference.md Markdown 34L · 967 B

├─ ▾ 📁 scripts

│ ├─ 🐍 example.py Python 19L · 591 B

│ └─ 🐍 init_master.py Python 51L · 1.3 KB

└─ 📝 SKILL.md Markdown 88L · 2.8 KB

安全亮点

✓ No network requests, external IPs, or data exfiltration

✓ No credential harvesting or environment variable access

✓ No shell execution, subprocess, or curl/wget usage

✓ No base64, obfuscation, or anti-analysis patterns

✓ No sensitive paths (~/.ssh, ~/.aws, .env) accessed

✓ Simple, transparent Python code with no hidden functionality

✓ All reference docs and assets are clearly marked as placeholders

✓ No malicious indicators in pre-scan IOCs