AI Security Guard用于代理输入（MCP、网页、文档、电子邮件）的实时安全扫描器，可检测提示注入和数据泄露并返回裁决结果。

4.5 (6)

审阅者 Daniel Nikulshyn·更新 2026年6月

AI Security Prompt Injection Real-Time Scanning Agent Safety MCP Threat Detection Enterprise

概览

AI Security Guard 是一款实时安全扫描器，针对代理输入（MCP、网页、文档、电子邮件）检测提示注入和数据外泄。它提供判决和易于使用的 SDK，用于安全代理管理。该平台还包含教育和研究资源，帮助理解安全与代理创新的现状。该工具旨在为用户提供技能、工具与知识，以防御安全威胁。它包含供应链防护、设备硬化扫描以及对内容代理在生成与摄取时的运行时扫描等功能。 AI Security Guard提供用户友好的体验，拥有健康等级、安全态势和成本跟踪等功能。它能够检测并阻止恶意包，并提供对Agent活动的高度可视化。平台还提供一个免费的 AI Security Action Pack，包含 15 篇与 OWASP Agentic Top 10 对应的专业文章，以及 12 个可安装的 Claude Code 和 Cursor 技能。总体而言，AI Security Guard 为代理提供强大、持续的保护，重点在于易用性和教育。

主要功能

供应链保护
设备强化扫描
内容运行时扫描
令牌支出智能
预测风险分析
具有 14 阶段设备强化扫描的 Shield 扫描器

价格

模型: Free
分类: AI
评分: 4.5 / 5 (6)

使用场景

阻止 AI 代理中的提示注入

实时扫描来自 MCP 服务器、网页和工具的代理输入，以检测和阻止提示注入尝试，防止其到达模型。

防止数据泄露

检查代理处理的文档和电子邮件内容，以识别泄露尝试并返回裁决结果，以控制后续操作。

确保文档摄取管道的安全

在将不受信任的文档传递给 LLM 工作流程之前进行验证，确保文件中的恶意指令被标记并隔离。

保护电子邮件驱动的自动化

检查 AI 助手使用的入站电子邮件中是否包含隐藏的注入有效载荷，返回明确的裁决结果，以允许或阻止处理。

优点 & 缺点

优点

代理输入的实时安全扫描器
检测提示注入和数据泄露
用于安全代理管理的易用 SDK
用于理解安全和代理创新的教育和研究资源
具有健康评分、安全态势和成本跟踪的用戶友好體驗

缺点

定价信息有限
确切的功能和能力未完全详细说明

评测

4.5

6 个评分的平均值。

登录以留下评测。

Margaret Whitfield

May 27, 2026

Skeptical, then convinced

I went in skeptical — most tools in this space overpromise. It actually delivers on the core workflow, and support is responsive caught me off guard. Pricing gets steep at scale is why this isn't a perfect score, still, I'd recommend giving it a real trial.

Wei Chen

Mar 1, 2026

Years in this space

I've evaluated a lot of these over the years. What stands out here is the automation — handled better than most — and support is responsive. Worth the time if this is your use case.

Camille Laurent

Feb 10, 2026

Years in this space

I've evaluated a lot of these over the years. What stands out here is the onboarding — handled better than most — and it is genuinely easy to set up. Pricing gets steep at scale is my one real gripe. Worth the time if this is your use case.

Pierre Dubois

Aug 22, 2025

Use it every day

Honestly didn't expect to like it this much. The API is exactly what I needed, and it saves real time. but I reach for it almost every day now and it just clicks.

Fatima Zahra

Jul 4, 2025

Compared a few options

Evaluated this against two competitors. Where it wins: the dashboard and it saves real time. Where it lags: pricing gets steep at scale. On balance the feature set — especially the core workflow — justifies the 4 stars for our use case.

Jamal Carter

Jun 20, 2025

Use it every day

Honestly didn't expect to like it this much. The API is exactly what I needed, and it saves real time. I do wish a few rough edges remain, but I reach for it almost every day now and it just clicks.

问答

What is the cost of using AI Security Guard?

Information on pricing is limited and not fully detailed.

Asked by Kenji Watanabe · Nov 19, 2025

Is AI Security Guard easy to use?

Yes, AI Security Guard offers a user-friendly experience with features like health grade, security posture, and cost tracking, as well as easy-to-use SDKs for secure agent management.

Asked by Odalys Reyes · Oct 27, 2025

What types of threats can AI Security Guard detect?

AI Security Guard detects prompt injection and exfiltration, and can also detect and block malicious packages.

Asked by Sami Virtanen · Oct 1, 2025

What types of inputs does AI Security Guard scan?

AI Security Guard scans agent inputs from MCP, web, docs, and email in real-time.

Asked by Liam O’Connor · Aug 24, 2025