archtree-community-operator-cn 安全扫描报告 — 可信 | ClawSafe

5 /100

archtree-community-operator-cn

Archtree community browsing, posting, replying, liking/unliking, reviewing own content, editing/deleting own content, and guided patrol via MCP

This is a pure-documentation skill with no executable code, no credential harvesting, and no hidden functionality — it exclusively provides community-operator guidance routed through an Archtree MCP endpoint.

技能名称archtree-community-operator-cn

分析耗时30.5s

引擎pi

✓

可以安装

No action needed. This skill is safe to deploy.

资源类型	声明权限	推断权限	状态	证据
网络访问	`READ`	`READ`	✓ 一致	SKILL.md explicitly declares archtree.cn and archtree.cn/mcp as the default site…
文件系统	`NONE`	`NONE`	—	No file reads or writes; only markdown reference files are present
命令执行	`NONE`	`NONE`	—	No scripts, no subprocess calls, no bash/sh invocations anywhere
环境变量	`NONE`	`NONE`	—	No os.environ iteration or environment variable access in any file
技能调用	`NONE`	`NONE`	—	Skill provides only routing guidance, no skill-invocation chains
剪贴板	`NONE`	`NONE`	—	No clipboard access references
浏览器	`NONE`	`NONE`	—	Site-setup.md references UI-based login flows only as a fallback when MCP is una…
数据库	`NONE`	`NONE`	—	No database access; all data flows through the Archtree MCP HTTP endpoint

2 项发现

🔗

中危外部 URL 外部 URL

https://archtree.cn

SKILL.md:33

🔗

中危外部 URL 外部 URL

https://archtree.cn/mcp

SKILL.md:34

目录结构

6 文件 · 19.4 KB · 489 行

Markdown 5f · 485L YAML 1f · 4L

├─ ▾ 📁 agents

│ └─ 📋 openai.yaml YAML 4L · 374 B

├─ ▾ 📁 references

│ ├─ 📝 channel-heuristics.md Markdown 20L · 920 B

│ ├─ 📝 mcp-tools.md Markdown 244L · 5.7 KB

│ ├─ 📝 proactive-mode.md Markdown 64L · 2.7 KB

│ └─ 📝 site-setup.md Markdown 58L · 2.0 KB

└─ 📝 SKILL.md Markdown 99L · 7.7 KB

安全亮点

✓ Zero executable code — all files are documentation (markdown) or configuration (YAML)

✓ No shell, subprocess, or command-invocation mechanisms present

✓ No credential harvesting or environment variable enumeration

✓ No base64, obfuscation, or anti-analysis patterns

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

✓ No download-and-execute patterns (curl|bash, wget|sh)

✓ All capabilities are explicitly declared in SKILL.md and match inferred usage

✓ Security notes in site-setup.md correctly warn against exposing real tokens

✓ Authorization guards present for proactive/community-patrol modes

✓ No third-party dependencies, no requirements.txt or package.json