OpenSSF Prague Summit Convenes 200 Practitioners Around AI Security and EU Regulation
The OpenSSF Community Day Europe 2026 brought together maintainers, policymakers, and security leaders to address AI supply chains, the EU Cyber Resilience Act, and systemic challenges in open source software transparency.
More than 200 participants converged in Prague for the OpenSSF Community Day Europe 2026, assembling across two concurrent programming tracks to examine how security integrates into the software development lifecycle at scale. The Main Track (South Hall 3B) centered on community strategy, policy, and tooling, while the SCORED Track (South Hall 3A) hosted academic researchers presenting empirical studies and technical supply chain analysis.
Main Track: AI, Regulation, and Infrastructure
Artificial Intelligence in Open Source Systems
The conference opened with sessions on securing AI within open source infrastructure. Dmitry Tantsur of Red Hat and Tuomo Tanskanen from Ericsson discussed security analysis applied to Metal3 and OpenStack Ironic. Sheng Sun and Sarah Evans at Dell Technologies introduced verifiable AI provenance to address gaps in machine learning supply chain attestation. Sal Kimmich examined the scope of AI autonomy through tooling designed to manage how AI agents interact with kernel systems.
European Regulatory Compliance
Policy alignment dominated the main hall agenda. Michael Schuster delivered an update on the EU Cyber Resilience Act (CRA), followed by Annika Kristin Niemann's examination of supply-chain liability under the CRA and Product Liability Directive. A panel including Roman Zhukov (Red Hat), Daniel Appelquist, Madalin Neag (OpenSSF), and Megan Knight explored how to operationalize CRA requirements across policy, practice, and community. Gergely Csatari and Timo Perala then outlined practical approaches for integrating CRA obligations into routine open source due diligence.
Infrastructure Hardening and Supply Chain Transparency
Securing foundational software and improving visibility into dependencies emerged as central themes. Kris Borchers (OpenSSF) and Carlos O'Donell (Red Hat) discussed funding mechanisms for secure GLIBC and GNU toolchain infrastructure, while Francis Perron (Linux Foundation) introduced Project Akrites for vulnerability management. Jeff Diecks (OpenSSF) and Laura Guazzelli (Linux Foundation) demonstrated bug discovery using OSS-CRS, Hannah Braswell (Red Hat) presented structured security approaches via Gemara, and Adolfo García Veytia analyzed the SLSA tooling landscape.
Software bill of materials (SBOM) discoverability proved critical to the discussion. Philippe Ombredanne and Stephen Augustus (Bloomberg) outlined work toward unified open data infrastructure, while Mario Fahlandt and Koray Oksay stressed that SBOMs lack utility without effective discovery mechanisms. Yuta Kiyoumi (Honda) and Akihiko Takahashi (Fujitsu) demonstrated applying VEX standards for vulnerability sharing across multi-tier automotive supply chains, and Felix Lange (SAP) shared techniques for identifying abandoned open-source projects.
Global perspectives rounded out the main track. Yunseong Choi (OpenSSF Ambassador) and Donghwa Lee discussed open source security policy in Korea, while Adrianne Marcum and Christopher "CRob" Robinson of OpenSSF hosted an interactive game show segment.
SCORED Track: Research and Empirical Analysis
AI Agent and Model Security
The dedicated research track featured Luca Galli examining why large language models require continuous software composition analysis and static analysis testing. Run Hao presented SandScope, a behavioral audit layer for Model Context Protocol (MCP) tools used in LLM agent supply chains. Gale Fagan demonstrated surfacing supply chain attack campaigns within AI agent skill registries using document similarity techniques, and Jens Dietrich highlighted deserialization vulnerabilities in Hugging Face model loaders.
Component Security and SBOM Empirical Studies
Rigorous empirical research dominated the SCORED track's technical sessions. Aman Sharma detailed trusting-trust binary attacks targeting Linux distributions, Timothy Brennan presented predictive risk scoring methodologies, Yibin Sun covered verification of Python package builds, Dominic Tassio examined malware detection timelines and release age patterns in npm, and Romain Lefeuvre demonstrated detecting zero-day vulnerabilities in forked repositories.
Large-scale SBOM studies formed another research pillar. Lukas Gehrke presented empirical findings on how SBOM generation differs across tools and approaches regulatory alignment. Philippe Boinot and Alan Prado analyzed how specification ambiguities produce divergent SBOMs, Sanchit Sahay showed how monitoring the build process enhances SBOM accuracy, and Abhishek Reddypalle (Purdue University) alongside Dennis Roellke (Bloomberg) demonstrated improving SBOM integrations by navigating uncharted files. Linus Kühl drew lessons from physical supply chain risk management, while Ranindya Paramitha reviewed the erosion of inexpensive trust signals.
Sponsors and Community Support
The event was made possible through support from Silver Sponsor Honda, Microsoft Azure, Akrites (evening reception host), Google (SCORED Track sponsor), and Dell Technologies (session recording sponsor). The Linux Foundation and its corporate partners, along with the attending community members, enabled the exchange of insights and collaborative problem-solving.
Getting Involved
OpenSSF Community Day Europe 2026 underscored that open source security depends on open collaboration, empirical research, and shared tools. The foundation welcomes contributions from participants of all backgrounds and experience levels.