Security by Design Meetup: AI Agents in Code Security
Details
Microblink invites you to a new Security by Design Meetup, this time dedicated to AI agents in code security.
Agentic coding has moved from experiment to everyday practice. Developers now hand entire tasks to tools like Codex, Copilot and Cursor, and AI agents review, test and even fix code on their own. On the security side the progress is real: agents are uncovering vulnerabilities in widely used software that years of conventional scanning missed. Yet the same technology is creating problems just as fast. AI-generated code that compiles flawlessly still regularly fails basic security checks, and some bug bounty programs have had to pause under a flood of low-quality AI-generated reports. Meanwhile, agents themselves have become a target, as malicious skills and insecure MCP servers turn a helpful assistant into an attacker’s tool.
At the next Security by Design meetup, a panel of practitioners from Apatura, Ascalia, Infinum, Memonia and Microblink will take on the open questions this shift has raised. We’ll ask whether AI coding tools leave us with more or fewer vulnerabilities, and whether AI agents can outperform traditional methods in code analysis and penetration testing or simply generate more noise. The panel will also look at trust and control: how much autonomy agents should have in the development pipeline, and who is accountable when they get it wrong. Finally, we’ll explore what happens when AI becomes both attacker and target, from prompt injection to compromised agent supply chains. Expect concrete examples from real projects rather than vendor slides, and plenty of room for questions from the audience.
The meetup is in person only and will not be recorded, and it will be followed by food and drinks. Attendance is free, but space is limited. To register your interest, please fill out this form.
