OpenText 主页。
解决方案

智能体应用安全

使用 Fortify 智能体分析器,发现基于规则的分析无法发现的问题

计算机上的 Fortify 智能体分析器的仪表板视图

概述

观看 Fortify 智能体分析器如何利用智能体 AI 发现复杂漏洞

Fortify 智能体分析器功能会利用 AI 智能体,模拟安全审查员的推理方式进行分析,随后与您的 OpenText™ Fortify™ SAST 检测结果进行去重处理,并将结果整合到您的团队在 Fortify 软件安全中心和 OpenText™ Fortify™ on Demand 中的工作流中。
阅读解决方案概述

主要优势

无需离开您当前运行的 Fortify 工作流,即可发现基于规则的分析所遗漏的漏洞。

  • 发现基于规则的分析无法检测到的关键漏洞

    智能体分析可跨 10 种语言捕获 12 个真实严重 CVE 中的 10 个,其中大多数 CVSS 评分达到 9.0 或更高——这些是逻辑和授权漏洞,且没有可追踪的受污染数据路径。

  • 在一个位置管理所有发现结果

    检测结果会根据您基于规则的 SAST 结果进行去重处理,映射到 Fortify 类别,并显示在 Fortify 软件安全中心和 Fortify on Demand 中。无需第二种工具,也无需第二个分类队列。

  • 无需新增 SKU 即可深化智能体能力

    Fortify 智能体分析器已包含在您现有的 Fortify SAST 许可证中。您可以接入任何受支持提供商的 LLM,并仅向该提供商支付使用费用。

商业影响

  • 逻辑和授权漏洞检测

    有些漏洞没有可供追踪的数据痕迹。排除误报(例如将消息字符串误认为硬编码密码),并揭示基于规则的分析无法追踪的逻辑和授权漏洞。

  • 针对高风险应用的更全面防护

    定期对您的关键应用运行智能体分析,就像安排渗透测试一样,同时让 SAST 继续作为您快速、日常的第一道防线。

  • 仅限全新发现

    团队在重新对 SAST 已发现的问题进行分类时会浪费时间。将 SAST 结果提交后,Fortify 智能体分析器会过滤掉重复内容,因此您的团队只需审核新增内容。

  • 一个记录系统

    检测结果将存储在 Fortify 软件安全中心和 Fortify on Demand 中,并按 Fortify 分类体系进行归类,同时以 SARIF 格式导出,因此智能体生成的结果将与您的 AppSec 计划中的其他内容一同呈现。

  • 扫描成本控制

    您可以使用自己的 LLM,因此可以掌控成本。设置每次运行的费用上限,为每个阶段选择模型和工作量,并在关键应用上进行更深入的分析,其他部分则适当简化。

探索解决方案的组成部分

产品

OpenText 将基于智能体和基于规则的分析整合到一个应用程序安全计划中。Fortify 智能体分析器(随 SAST 一起提供)能够发现基于规则的分析无法发现的问题;Fortify SAST 提供快速、确定性的覆盖;而 Fortify on Demand 则以 SaaS 形式提供该服务。

专业服务

OpenText 将端到端解决方案的实施与全面的技术服务相结合,以帮助改进系统。

Resources

  • Agentic AST is the category Gartner defines for AI-agent-driven code analysis. Rather than matching rules or tracing a data path, AI agents reason about your source code the way an expert reviewer would, following logic and intent to find vulnerabilities rules-based analysis can't reach. Fortify Agentic Analyzer applies this to your codebase to uncover flawed authorization design, business-logic weaknesses, and subtle injection and persistence patterns. It complements Fortify SAST, which runs a suite of deterministic analyzers fast enough to use on every build.

  • No. It complements SAST, it does not replace it. Fortify SAST stays your fast, broad, deterministic first line of defense on every build, covering a wide range of weakness classes. Fortify Agentic Analyzer adds a deeper, agentic pass you run selectively on higher-risk applications, the way you would schedule a penetration test. Used together, they cover ground neither reaches alone.

  • Yes. It is included under your existing Fortify SAST license at no additional charge. If an application is licensed for SAST analysis, you can analyze it with agentic analysis too, whether you connect through Fortify Software Security Center or Fortify on Demand. You bring your own LLM and pay your provider for usage: there is no separate SKU and no fixed AI surcharge from OpenText.

  • The flaws that turn on logic and intent rather than a traceable data path: broken authorization and authentication design, business-logic weaknesses, and subtle injection and persistence patterns. In a published study across 10 languages, Fortify Agentic Analyzer found 10 of 12 real-world CVEs, most rated CVSS 9.0 or higher, all disclosed after the analysis model's training cutoff, so the results reflect reasoning about the code rather than recall of a known bug.

  • You bring your own. Fortify Agentic Analyzer works with a wide set of LLM providers, connected through your own API key or a supported subscription, so you control model choice and cost. Usage is billed by your provider, not by OpenText. You can select the model and effort level for each stage of a scan to balance depth against spend.

  • They are distinct Fortify capabilities. Fortify Agentic Analyzer finds vulnerabilities rules-based analysis can't reach. Fortify Remediation Aviator triages findings and generates fixes to speed remediation. Fortify Agentic Analyzer does not require Remediation Aviator (which is an add-on). Fortify Agentic Analyzer is included with Fortify SAST.

  • Findings appear in Fortify Software Security Center and Fortify on Demand, reported in the Fortify taxonomy alongside your SAST, DAST, and SCA results, so agentic findings live in the same system of record your team already uses. Findings are deduplicated against your existing SAST results, so you review only what's net-new, and output is also available as SARIF for downstream tools.

  • Because you bring your own LLM, you control the spend. Set a per-run cost cap, choose the model and effort level for each stage, and run agentic analysis where it earns its keep: deeper on your critical applications, lighter elsewhere. There is no fixed AI surcharge from OpenText on top of your provider usage.

    Fortify Agentic Analyzer

    Watch the demo

    Fortify Agentic Analyzer

    Read the solution overview

    Fortify SAST

    Read the data sheet

    Fortify Agentic Analyzer

    Read the solution overview

    Fortify SAST

    Read the data sheet

    我们能提供什么帮助?

    脚注