Liran Tal researches AI Security with a focus on hardening agentic developer workflows and the Model Context Protocol (MCP). Over the past years, Liran has uncovered and responsibly disclosed CVEs in MCP servers, AI frameworks, and open-source libraries, publishing practical write-ups that show how issues like command injection, path traversal, and SQL read-only bypasses emerge in real systems and how to mitigate them. His work includes open-source tools for auditing software supply chains and research into how poisoned MCP tool metadata can steer agents toward unsafe actions. He also develops practical guardrails for build pipelines, IDE assistants, and CI to help prevent vulnerabilities in AI-generated code.

Beyond AI, Liran is a longstanding Node.js developer and secure-coding advocate. He authored books on Node.js security, contributed to the Node.js Security Working Group, and led community education through OWASP initiatives and the OpenSSF Secure Coding Labs. He's recognized as a GitHub Star and received the OpenJS Foundation JavaScriptLandia "Pathfinder for Security" award. Today, he leads Developer Relations at Snyk, where he turns research into approachable education with guidance—frameworks, demos, policies, and tooling that teams can adopt to safely use LLMs, agentic systems, and secure code in production.