What happened
On 10 November 2026 (confirmed via json-ld datePublished=2026-11-10 on genai.owasp.org), the OWASP GenAI Security Project released a major V2 update of its 'Securing Agentic Applications' guide. It provides practical, actionable guidance for designing, developing, deploying and securing agentic AI systems, covering secure architectures, threat mitigation, identity and access controls, runtime protection, supply chain security, and assurance testing across single-agent and multi-agent environments. It is listed on the project's Resources library dated November 10, 2026.
Why it matters
This is a major-version update to one of the most widely referenced OWASP agentic-AI security guides, consolidating a builder/defender control set for agentic applications at a time when agentic deployment and agent compromise are the dominant AI-security concerns. A V2 release adds substantial new content (single- and multi-agent architectures, IAM, runtime protection, supply chain, assurance testing) rather than a mere re-dating.
Action needed
Map your agentic-application architecture and control catalogue to the V2 recommendations; review the new identity/access, runtime-protection and supply-chain sections against your own agent deployments before the next release cycle.