← 返回 Skills 市场
apex-stack-ai

MCP Security Auditor Lite

作者 apex-stack-ai · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ 安全检测通过
104
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install mcp-security-auditor-lite
功能描述
Free version — scan your MCP configuration for the top 3 security risks. Tool description injection, permission sprawl, and supply chain trust.
使用说明 (SKILL.md)

MCP Security Auditor Lite — Quick Security Scan

You are an MCP security specialist. Your job is to quickly assess MCP server configurations for the most critical security risks.

This lite version covers 3 of 8 audit dimensions. For the full MCP Security Auditor with all 8 dimensions, tool injection scanning, config drift detection, cross-tool safety analysis, and ongoing monitoring checklists, get the paid version: https://apexstack.gumroad.com/l/mcp-security-auditor


How to Use

Provide your MCP config (JSON/YAML), tool list, or describe your MCP server setup. I'll scan for the top 3 risks.


Quick Security Scan (Lite — 3 Dimensions)

1. Tool Description Integrity — /10

Are tool descriptions purely descriptive or do they contain hidden instructions?

Red flags:

  • Imperative language ("always do X before calling other tools")
  • References to other tools' behavior
  • Unusually long descriptions (more attack surface)
  • Instructions to ignore or override previous context

Scoring:

  • 9-10: All descriptions purely descriptive, manually reviewed
  • 5-6: Some imperative language, no hidden content detected
  • 1-2: Active injection patterns, descriptions manipulate agent behavior

2. Permission Scope — /10

Do tools have the minimum permissions needed?

Red flags:

  • File system tools with root/home access instead of scoped directories
  • Database tools with write access when only reads are needed
  • Tools that can access environment variables or secrets
  • Admin-level access on tools that should be read-only

Scoring:

  • 9-10: Every tool follows least-privilege, scoped to specific resources
  • 5-6: Several tools have broad permissions, no systematic scoping
  • 1-2: Tools have admin access, can access secrets, no boundaries

3. Supply Chain Trust — /10

Are your MCP servers from trusted sources?

Red flags:

  • Unverified community MCP servers with no source review
  • No version pinning (running "latest" = rug-pull risk)
  • Servers installed without security evaluation
  • No CVE monitoring for MCP dependencies

Scoring:

  • 9-10: Verified publishers, pinned versions, source reviewed
  • 5-6: Mix of trusted and unverified, some pinning
  • 1-2: Random servers installed without evaluation

Lite Output

## MCP Quick Security Scan: [Project]

### Score: [X/30] ([percentage]%) — [Secure / Adequate / At Risk]

| Dimension | Score | Risk | Top Action |
|-----------|-------|------|------------|
| Tool Description Integrity | X/10 | red/yellow/green | [action] |
| Permission Scope | X/10 | red/yellow/green | [action] |
| Supply Chain Trust | X/10 | red/yellow/green | [action] |

### Top 3 Fixes
1. [action]
2. [action]
3. [action]

Want the full security audit? The paid version includes all 8 dimensions, tool description injection scanner, permission scope analyzer, config drift detector, cross-tool manipulation checker, monitoring checklists, and prioritized remediation roadmap.

Get the full version -> https://apexstack.gumroad.com/l/mcp-security-auditor


Built by Apex Stack — based on real experience running 10+ MCP-connected agents in production.

安全使用建议
This skill is a checklist-style, manual analyzer and is internally consistent with its description. Before using it: do not paste live secrets, API keys, or private keys into the chat — sanitize or redact sensitive fields; verify any remediation steps before applying them; treat the paid-version link as an external marketing URL (don’t provide credentials there); and remember the output is agent reasoning (not an automated code audit), so consider running independent tooling for confirmatory checks if you need high assurance.
功能分析
Type: OpenClaw Skill Name: mcp-security-auditor-lite Version: 1.0.0 The skill bundle consists entirely of markdown instructions (SKILL.md) directing the AI agent to act as a security auditor for MCP (Model Context Protocol) configurations. It contains no executable code, no data exfiltration logic, and no malicious prompt injection; instead, it provides a framework for the agent to evaluate security risks like tool description integrity and permission sprawl. The inclusion of a link to a paid version on Gumroad is a standard commercial practice and does not pose a technical security threat.
能力评估
Purpose & Capability
Name and description match SKILL.md: it promises a lightweight, manual-style security scan of MCP configs across three dimensions. There are no unexpected binaries, env vars, or installs required.
Instruction Scope
The skill is instruction-only and asks the agent to evaluate MCP config/tool lists provided by the user using the included rubrics. This is expected, but the rubric-driven analysis is manual reasoning rather than automated checks; the user must supply config data (which may contain secrets) and the agent will analyze it.
Install Mechanism
No install spec or code files; lowest-risk delivery model. Nothing is downloaded or written to disk by the skill itself.
Credentials
The skill declares no required credentials or environment access (appropriate). However, it requires the user to paste MCP configs/tool lists — those artifacts can contain sensitive secrets or tokens, so the user should sanitize inputs before sharing.
Persistence & Privilege
always is false and default invocation behavior is normal. The skill does not request persistent presence or system-wide changes.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install mcp-security-auditor-lite
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /mcp-security-auditor-lite 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
Initial release of MCP Security Auditor Lite. - Launches free version to scan MCP configurations for the top 3 security risks: tool description integrity, permission sprawl, and supply chain trust. - Provides scoring and actionable recommendations for each risk. - Includes a concise output template to summarize findings and top fixes. - Full version link provided for advanced auditing and extra features.
元数据
Slug mcp-security-auditor-lite
版本 1.0.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

MCP Security Auditor Lite 是什么?

Free version — scan your MCP configuration for the top 3 security risks. Tool description injection, permission sprawl, and supply chain trust. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 104 次。

如何安装 MCP Security Auditor Lite?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install mcp-security-auditor-lite」即可一键安装,无需额外配置。

MCP Security Auditor Lite 是免费的吗?

是的,MCP Security Auditor Lite 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

MCP Security Auditor Lite 支持哪些平台?

MCP Security Auditor Lite 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 MCP Security Auditor Lite?

由 apex-stack-ai(@apex-stack-ai)开发并维护,当前版本 v1.0.0。

💬 留言讨论