OpenAI's Promptfoo Deal Plugs Agentic AI Testing Gap

OpenAI is stepping up its push to bolster security around its enterprise AI ecosystem, announcing the acquisition of agentic security testing firm Promptfoo. The purchase followed an interview conducted on 9 March with Jamieson O’Reilly, OpenClaw’s security advisor, who warned that more ways were needed to scan AI tools for human-language malware.

Promptfoo, founded in July 2024 by Ian Webster and Michael D’Angelo, provides open source tools to test and evaluate large language models and AI agents, including scanning vulnerabilities in LLMs, red‑teaming AI tools, evaluating prompts and models, and a secure proxy for model context protocol servers.

According to OpenAI’s 10 March announcement, Promptfoo’s tools are used by over 25% of Fortune 500 companies; the startup has raised $23m to date, with $18.4m from Insight Partners in July 2025 and participation from Andreessen Horowitz. OpenAI said Promptfoo’s suite will remain open source and that its technology will be integrated into OpenAI Frontier to enhance security testing, governance and accountability across enterprise AI coworkers.