OpenAI Acquires Promptfoo for AI Security
OpenAI acquired Promptfoo, an AI security startup, to stress-test AI agents for vulnerabilities like prompt injection and data leakage.
Promptfoo, founded in 2024, offers tools to evaluate AI prompts and models, addressing risks like prompt injection and data leakage. Their platform supports integrations with development tools and CI/CD pipelines, enabling local testing and red-team simulations. OpenAI plans to integrate Promptfoo's technology into its Frontier platform for building and operating AI coworkers. This will enhance agent behavior testing and risk detection before deployment. Promptfoo's tools are already used by over 25% of Fortune 500 companies. The company has raised $23 million in funding, valuing it at $86 million after its Series A round in July 2025. The acquisition highlights the growing need for security measures as AI agents interact with real-world systems. These agents can access internal data, call APIs, and trigger actions, creating security risks that traditional tools don't address. OpenAI is also developing its own AI-powered security tools, such as Aardvark, an AI agent designed to identify vulnerabilities and propose fixes. They also launched Trusted Access for Cyber, an initiative to enhance digital defenses and manage risks associated with AI systems.
