低风险 — 风险评分 20/100
上次扫描:1 天前 重新扫描
20 /100
eightfold
Eightfold integration — manage data, records, and automate workflows via the Membrane CLI
This skill is documentation-only with no executable code files. It describes how to use the Membrane CLI for Eightfold integration, with no hidden functionality, credential harvesting, or malicious behavior detected.
技能名称eightfold
分析耗时33.2s
引擎pi
可以安装
Consider adding an explicit 'allowedTools' declaration in SKILL.md frontmatter to clarify shell:WRITE usage for CLI commands. Otherwise, the skill is safe to use.

安全发现 3 项

严重性 安全发现 位置
低危
Shell execution capabilities not declared in frontmatter 文档欺骗
The SKILL.md frontmatter declares no allowed-tools, yet the documentation instructs users to run shell commands (npm install -g, membrane login, membrane connect, membrane action run). While this is documented inline, it violates the principle of explicit capability declaration.
allowed-tools: (not declared)
→ Add 'allowed-tools: shell:WRITE' to the frontmatter if shell execution is expected. Alternatively, restructure docs to avoid CLI command examples.
SKILL.md:1
低危
External dependency on Membrane CLI not flagged as risky 文档欺骗
The skill relies on @membranehq/cli as an external dependency. The SKILL.md uses 'npm install -g @membranehq/cli' without version pinning (uses @latest). No verification that the CLI is trustworthy, and it is not a universally known tool.
npm install -g @membranehq/cli
→ Pin the CLI version (e.g., @membranehq/[email protected]) to prevent supply-chain attacks from a compromised package.
SKILL.md:34
低危
Feature enumeration lacks actionable documentation 文档欺骗
The skill lists ~50 capability areas (Security, Compliance, Budget, Forecasting, etc.) as a bullet list but provides no specific actions, parameters, or examples for most of them. This creates user confusion and suggests copy-pasted content.
- **Security**
- **Compliance**
- **Goal**
...
→ Remove non-functional capability entries or provide at least one actionable example per listed feature.
SKILL.md:22
资源类型声明权限推断权限状态证据
文件系统 NONE NONE No file operations found in documentation
网络访问 NONE READ ✓ 一致 SKILL.md line 7,19: External URLs to getmembrane.com and developer.eightfold.ai,…
命令执行 NONE WRITE ✓ 一致 SKILL.md lines 34-104: npm install -g, membrane login, membrane connect, membran…
环境变量 NONE NONE No direct environment variable access described; credentials managed by Membrane…
技能调用 NONE NONE No cross-skill invocation documented
剪贴板 NONE NONE No clipboard operations described
浏览器 NONE NONE Browser-based auth mentioned but handled by Membrane CLI, not by the skill itsel…
数据库 NONE NONE No direct database access described
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://developer.eightfold.ai/
SKILL.md:19

目录结构

1 文件 · 5.2 KB · 169 行
Markdown 1f · 169L
└─ 📝 SKILL.md Markdown 169L · 5.2 KB

依赖分析 1 项

包名版本来源已知漏洞备注
@membranehq/cli latest (implied) npm Version not pinned; external package from npm registry

安全亮点

✓ No executable code files present — skill is purely documentation
✓ No credential harvesting or exfiltration detected
✓ No base64, eval, or obfuscation patterns found
✓ No direct access to sensitive paths (~/.ssh, ~/.aws, .env)
✓ No reverse shell, C2 communication, or data theft indicators
✓ Credential management delegated to Membrane CLI rather than storing secrets locally
✓ Best practices section correctly advises against sharing API keys directly
✓ No supply-chain risk from internal dependencies (no package.json, requirements.txt, etc.)