OpenSSF to showcase agentic AI security standards at San Jose conference
The Open Source Security Foundation will sponsor AGNTCon + MCPCon North America in October 2026, presenting workshops and talks on securing AI systems through the Secure Agentic Framework.

The Open Source Security Foundation (OpenSSF) has secured a Silver Sponsor position at AGNTCon + MCPCon North America, taking place in San Jose, CA, on October 22–23, 2026. Agentic AI security represents the year's most pressing technical challenge, and the conference offers an ideal platform for establishing critical open source security standards as artificial intelligence matures. The industry needs standardized protocols as production-ready systems enter deployment across the ecosystem.
The OpenSSF is positioning the Secure Agentic Framework (SAF) as a foundational industry standard for building interoperable and secure agentic AI systems. Designing these systems with security as a core requirement becomes essential as frameworks become embedded in software supply chains. The foundation is inviting open source developers and contributors to participate in hands-on workshops and technical sessions led by its experts.
OpenSSF Community Sessions at the Conference
Playing the Agentic AI Supply Chain Game
Sarah Evans and Christopher Robinson will use a board game format to examine the software supply chain through an agentic AI lens. Despite rapid movement in the agentic AI space, proven security lessons from decades of software engineering remain applicable. The session demonstrates how to transfer established supply chain security practices to agentic AI contexts, enabling developers and users to construct more robust and dependable systems using familiar concepts and standards.
Hands-On Workshop: Secure Agentic Framework Implementation
Sarah Evans, Jautau "Jay" White, Frederick Kautz, and Laura Guazzelli will lead an intensive workshop on the Secure Agentic Framework (SAF). Participants will work through two separate implementations: one designed for the Model Context Protocol (MCP) and another for Kubernetes runtimes. The workshop includes analysis of a compromised environment where attendees will identify and remediate security threats including tool poisoning and command injection attacks. For the OpenSSF community, this initiative marks a significant advance in securing AI system integration, moving beyond generic frameworks to deliver protocol-specific security controls tailored to contemporary agentic AI requirements.
OpenSSF Booth and Networking
The OpenSSF will maintain a booth throughout the conference where attendees can connect with team members and community participants. Visitors can collect branded merchandise, including the foundation's distinctive Honk and CRAfish keychains, and discuss collaboration opportunities for strengthening open source ecosystem security.
PyTorchCon Co-Location
AGNTCon + MCPCon shares the venue with PyTorchCon, providing a chance to explore the broader AI and machine learning landscape. The OpenSSF considers securing AI and machine learning tools a major community priority, and the co-location arrangement simplifies connections between advanced AI development and security best practices.
Registration is now open for the San Jose event. The OpenSSF newsletter provides additional details on these and upcoming events.