CrowdStrike Falcon AIDR Secures AI Agents with NVIDIA NeMo Guardrails

CrowdStrike Falcon AIDR now integrates NVIDIA NeMo Guardrails (v0.20.0) to provide enhanced security for AI agents, addressing the challenge of limiting agent scope and preventing abuse as these agents transition from experimental projects to mainstream business tools. A compromised agent can expose customer data, execute unauthorized transactions, or violate compliance requirements. Falcon AIDR with NeMo Guardrails enables organizations to manage agentic data access, control responses, and ensure policy compliance and safety controls. This integration allows organizations to confidently move AI agents from development to production by defining guardrails that constrain the capabilities of Large Language Models (LLMs).

Attack Chain

1. An attacker crafts a malicious prompt containing a prompt injection attack designed to manipulate the AI agent's behavior.
2. The user submits the prompt to the AI agent via a chat interface or API endpoint.
3. Falcon AIDR intercepts the prompt and analyzes it using its built-in classification rules and NVIDIA NeMo Guardrails.
4. If a prompt injection attack or other malicious content is detected, Falcon AIDR blocks the prompt from being processed by the AI agent.
5. If the prompt contains sensitive data, such as PII, Falcon AIDR redacts the data before it is processed by the AI agent.
6. The AI agent processes the sanitized prompt and generates a response.
7. Falcon AIDR analyzes the AI agent's response to ensure it does not contain any sensitive data or malicious content.
8. The AI agent delivers the response to the user, completing the interaction without exposing sensitive information or executing unauthorized actions.

Impact

Compromised AI agents can lead to customer data exposure, unauthorized transactions, and compliance violations. Successful prompt injection attacks could manipulate agent behavior to bypass security controls and access sensitive resources. Without proper guardrails, AI agents in financial services, healthcare, and customer service organizations are vulnerable to attacks that could compromise sensitive data and disrupt critical business processes. Protecting AI agents is essential for maintaining data security, ensuring compliance, and preventing financial losses.

Recommendation

• Implement CrowdStrike Falcon AIDR with NVIDIA NeMo Guardrails (v0.20.0) to protect AI agents against runtime attacks.
• Utilize Falcon AIDR's built-in classification rules to detect and block prompt injection attacks, redact sensitive data, defang malicious content, and moderate unwanted topics.
• Create custom Falcon AIDR policies tailored to specific security requirements, focusing on chat input sanitization, chat output filtering, RAG data ingestion, and agent tool invocation.
• Deploy the Sigma rules below to your SIEM to detect potential prompt injection attempts targeting AI agents.