📕 LLM Security: 50+ Adversarial Probes you need to know.
Download the guide
KnowledgeBlog
September 24, 2024
Blanca Rivera Campos

Evaluating LLM applications: Giskard Integration with NVIDIA NeMo Guardrails

Giskard has integrated with NVIDIA NeMo Guardrails to enhance the safety and reliability of LLM-based applications. This integration allows developers to better detect vulnerabilities, automate rail generation, and streamline risk mitigation in LLM systems. By combining Giskard with NeMo Guardrails organizations can address critical challenges in LLM development, including hallucinations, prompt injection and jailbreaks.
Giskard integrates with NVIDIA NeMo

Factoring in the safety and reliability of LLM-based applications is critical as more industries, from healthcare to finance, deploy them in their operations. Having this in mind, we worked on a new integration between Giskard and NVIDIA NeMo Guardrails, an open-source toolkit for easily adding programmable guardrails to LLM-based conversational systems.

LLMs are susceptible to issues that traditional models don't face, such as hallucinations and adversarial attempts including jailbreaks and prompt injections. Developers often struggle with where to start, what issues to focus on, and how to implement effective tests for these risks. Errors in LLMs can lead to legal liability, reputational damage, and costly service disruptions.

Giskard’s integration with NeMo Guardrails equips developers to create more secure and robust LLM-based applications, helping address the need for comprehensive LLM evaluation.

Why integrate with Giskard?

By integrating Giskard for use in NeMo Guardrails, users gain a powerful synergy that helps significantly improve the development and deployment of LLM applications. Developers will be able to:

  1. Better detect LLM vulnerabilities: Giskard’s scanning capabilities can identify potential weaknesses and failure modes in LLM applications.
  2. Automate rail generation: The combination of technologies can, based on the vulnerabilities detected by Giskard, automatically generate Colang rules to address the issues.
  3. Streamline risk mitigation: By combining Giskard’s testing with NeMo Guardrails’ control mechanisms, developers can better anticipate and prevent potential problems before deployment. This helps result in more secure LLM-based applications while simplifying the workflow of identifying issues and implementing protective measures.
Colang file with detected vulnerabilities

How it works

The integration process is straightforward. After running a Giskard scan on an LLM application, users can export the detected vulnerabilities as Colang rules using a simple Python command. These generated rules can then be incorporated into the NeMo Guardrails configuration, providing immediate protection against the identified risks. The integration supports both Colang 1.0 and 2.x versions, offering flexibility for different development environments.

Giskard - NVIDIA NeMo workflow

More about the Giskard’s integration with NeMo Guardrails

Giskard’s integration with NVIDIA NeMo Guardrails helps in creating safer, more reliable LLM-based applications. By combining Giskard's advanced testing capabilities with NeMo Guardrails' robust control mechanisms, developers can now more easily identify and mitigate potential risks in their AI systems. We invite developers and organizations to explore this integration and experience the benefits of enhanced LLM application security.

For more details about this integration, read the documentation.

At Giskard we are developing a holistic platform to address LLM risks across quality, security, and compliance domains. Our platform allows AI teams to automate test creation, enabling efficient model validation, detailed reporting, and optimized review procedures.

‍Reach out to us today to learn more about how we can help you to ensure your LLM-based applications are safe and reliable.

Continuously secure LLM agents, preventing hallucinations and security issues.
Book a Demo

You will also like

Who judges the LLM-as-a-Judge? Meta-Evaluation of an LLM vulnerability scanner

Who judges the LLM-as-a-Judge? Meta-Evaluation of an LLM vulnerability scanner

When your LLM vulnerability scanner detects a threat, it relies on an LLM judge to decide whether the attack succeeded. Using one LLM to evaluate another can bring some failures into your evaluation pipeline (false positives, model drift, or context blindness). This article walks through how we meta-evaluated our own LLM-as-a-judge using giskard-checks to freeze expected verdicts, replay attack traces, and detect evaluator regressions in CI.

View post
Giskard + Databricks integration

Partnership announcement: Bringing Giskard LLM evaluation to Databricks

Giskard has integrated with Databricks MLflow to enhance LLM testing and deployment. This collaboration allows AI teams to automatically identify vulnerabilities, generate domain-specific tests, and log comprehensive reports directly into MLflow. The integration aims to streamline the development of secure, reliable, and compliant LLM applications, addressing key risks like prompt injection, hallucinations, and unintended data disclosures.

View post
Giskard Guards

Securing AI agents with Europe's first sovereign guardrail platform

Giskard has launched Guards, the first independent, EU-sovereign guardrail platform that allows regulated enterprises to secure their AI applications on-premise. Moving beyond the limitations of generic content filters, the platform is context-aware and built specifically for modern AI agents, securing the full execution chain via a Policy-as-Code framework that includes ready-to-use EU AI Act and OWASP Top 10 LLM compliance packs. To showcase this new standard in enterprise AI security, our technical team is hosting an in-depth live session on May 13, 2026.

View post
Get AI security insights in your inbox
Subscribe

Get AI security insights in your inbox

No vulnerabilities found?

We refund the assessment.
Limited-Time Offer: available until the end of September.