中文 EN

扫描报告

扫描新文件

⚠ 低风险 — 风险评分 20/100
上次扫描：1 天前 重新扫描
20 /100
outgrow
Outgrow integration for managing data, records, and workflow automation via the Membrane CLI
A legitimate Outgrow integration skill using the Membrane CLI; no malicious behavior found, but documentation does not declare the required shell/Bash permissions.
技能名称outgrow
分析耗时44.3s
引擎pi
✓
可以安装
Add a declared permissions section to SKILL.md listing Bash→shell:WRITE and clarify that CLI tools are invoked through shell. Pin the npm package version with @membranehq/[email protected] instead of @latest.

安全发现 2 项

严重性 安全发现 位置
低危
Shell execution not declared in allowed-tools 文档欺骗
SKILL.md describes multiple Bash CLI invocations (npm install, membrane login, membrane action run, membrane request) but the declared allowed-tools mapping does not list Bash/shell:WRITE. This is a documentation gap that should be corrected for transparency.
npm install -g @membranehq/cli
→ Add 'Bash → shell:WRITE' to the skill's declared permissions
 SKILL.md:35
低危
Unpinned npm package version 供应链
The CLI installation uses @latest instead of a pinned version, allowing the dependency to change silently over time.
npm install -g @membranehq/cli
→ Pin to a specific version range, e.g., npm install -g @membranehq/cli@^1.0.0
 SKILL.md:35
资源类型声明权限推断权限状态证据
文件系统 NONE NONE — No filesystem access detected
网络访问 READ READ ✓ 一致 External URLs to getmembrane.com and developers.outgrow.co are declared in docs
命令执行 NONE WRITE ✗ 越权 SKILL.md instructs: npm install -g @membranehq/cli, membrane login --tenant, mem…
2 项发现
🔗
中危 外部 URL 外部 URL
https://getmembrane.com
SKILL.md:7
🔗
中危 外部 URL 外部 URL
https://developers.outgrow.co/
SKILL.md:19

目录结构

1 文件 · 4.3 KB · 126 行
Markdown 1f · 126L
└─ 📝 SKILL.md Markdown 126L · 4.3 KB

依赖分析 1 项

包名版本来源已知漏洞备注
@membranehq/cli latest (unpinned) npm 否 No version pinned — supply chain drift possible

安全亮点

✓ No credential harvesting or environment variable exfiltration observed
✓ No base64, eval, or obfuscated code patterns present
✓ No hidden HTML comments or shadow instructions
✓ No direct IP network requests — all network calls route through the Membrane CLI proxy
✓ No access to sensitive paths such as ~/.ssh, ~/.aws, or .env
✓ No reverse shell, C2, or persistence mechanisms
✓ Skill describes a legitimate, documented integration pattern
✓ Membrane's proxy request is scoped to authenticated connections, limiting abuse surface