This attack path highlights important considerations for the future of agent-assisted development. Extended supply chain risk: Traditional supply chain attacks focus on injecting malicious code directly. In agentic environments, a compromised dependency can also redirect the agent itself, extending familiar supply chain risks into a new dimension, such as injecting subtle delays that cause performance degradation or denial-of-service scenarios. Instruction following under adversarial conditions: When the agent followed injected configuration directives, including instructions to conceal its
Mitigating Indirect AGENTS.md Injection Attacks in Agentic Environments | NVIDIA Technical BlogAGENTS.md files help Codex and similar AI tools understand project-specific instructions, coding conventions, and organizational structures. They can reside anywhere within a Codex container, providing valuable context to AI agents. Like other project configuration files, these instructions are treated as trusted context by the agent. This trust model is by design, but it creates an interesting attack surface when a malicious dependency is able to write or modify these files at build time.
Mitigating Indirect AGENTS.md Injection Attacks in Agentic Environments | NVIDIA Technical Blog…By making cuOpt available as an agent skill, the LLM can hand off the mathematical heavy-lifting to the GPU, while focusing on understanding the business problem, gathering data, and returning actionable…
…They’re drawn from a broader post-training corpus of 40 million samples, which cover reasoning, instruction following, coding, safety, and multi-step agent tasks. This stage establishes the behavioral foundation that…
…vehicles with agentic AI are expected to grow to 70 million by 2035 from around 5 million in 2025. Delivering these experiences within the vehicle—where strict latency, safety, and privacy requirements…
…His work centers on optimized, reliable AI inference for safety-critical robotics and automotive edge systems. Before joining NVIDIA in 2018, Michael created and led several floating-point-focused verification tools at…
…The threat model of a code agent Using an SLM to generate code that runs on the host introduces obvious security and safety risks, including: Dangerous function access. The SLM generates os…
…Step 3: The agent generates a complete application with multiple files — model download scripts, the pipeline app, inference config file, and more. Let’s focus on the files that matter for model…
…solutions architect focusing on NVIDIA generative AI. With over 15 years covering AI, ML, full-stack accelerated computing, he designs AI solutions for real-world application, and leads partner agentic AI integration…
…Agentic AI / Generative AI | Trustworthy AI / Cybersecurity | General | Intermediate Technical | Deep dive | AI Red Team | Security for AI About the Authors About Joseph Lucas Joe is a Principal Offensive Security Researcher focused…
…In autonomous agents, attackers hijack the agent’s goals, ensuring continued pursuit of attacker-defined goals. Persistence enables attackers to repeatedly exploit hijacked states, increasing the likelihood of downstream impact. In agentic…
…When building agentic systems with VLMs, continue evaluating them based on their autonomy level and threat modeling . Explore the family of NVIDIA VLMs . Discuss (0) Discuss (0) Tags Agentic AI / Generative AI…
