Agent Sandboxing & AI Governance

As AI agents move from demos to production, one hard question takes center stage: how do you run autonomous agents without handing them the keys to the kingdom?

This community meetup brings together AI practitioners, security engineers, and platform builders to tackle two sides of the same problem - sandboxing agents so a bad action can't blow up your environment, and governing them so you know what they're doing and can prove it.

Register Now! https://forms.gle/9oqgEe7SViqyDLW16

📢 Call for Speakers — We Want to Hear From You!
This agenda is built by the community. Got hands-on experience securing, sandboxing, or governing AI agents? Submit a talk. We're looking for:

• Agent sandboxing — microVM isolation, "a container is not a sandbox," runtime boundaries
• AI governance — policy enforcement, audit trails, SIEM integration, compliance
• MCP gateways — dynamic MCP, policy, and observability
• Agent security horror stories — what went wrong in production and what you learned
• Blast radius — limiting what an agent can touch when it misbehaves
• Anything else at the intersection of agents, security, and trust

Talk formats: Full talk (~25 min) or lightning Show-n-Tell demo (5–7 min).
Submit your proposal: https://forms.gle/ircEAQiGA48aybbEA
CFP closes: Wed, 15 July 2026

👥 Who Should Attend
Developers building multi-agent systems · Platform and DevOps engineers · Security folks hardening AI infra · Anyone shipping LLMs and agents into production. All skill levels welcome.

***

🎁 What You Get
Practical patterns you can take back to your team
· Real war stories from people doing this in production
· A chance to demo your own work
· Connection to the wider Collabnix community.

Session will be recorded and shared afterward.