Name: Your newest developer is an AI agent. Who’s governing it?
Start: 2026-12-01T16:00:00.000Z
End: 2026-12-01T16:45:00.000Z

AI agents are able to install packages, hold credentials, and publish to registries. Cloudsmith and Docker will walk through the steps required to harness this power while protecting your data and infrastructure. You’ll leave with a playbook for your next agent rollout.


Stéphane Montri

Senior Partner Solutions Engineer

Esteban Garcia

Principal Engineer

Meghan McGowan

Product Marketing Manager

Your newest developer works around the clock and writes code faster than anyone on the team. It takes on every ticket and shrugs off changing priorities. What they haven't learned yet is which packages your team trusts.

That’s because your newest developer is an AI agent. It pulls dependencies, holds credentials and publishes to registries. That’s work you’d only trust a new engineer with after onboarding and a few weeks of supervision.

So what does onboarding look like for an agent?

In this 45-minute panel, Docker and Cloudsmith walk through the process step by step. You’ll leave with a playbook for your next agent rollout: give it a safe workspace, set the rules for what it installs, start it on hardened images, and keep the records your auditors will ask for.

Your newest developer is an AI agent. Who’s governing it? 

AI agents add dependencies to builds faster than anyone can review them. Cloudsmith and Aikido Security's product leaders take one package request and show you which controls can help keep malicious packages out of your builds – at the registry, and at runtime.

Alison Sickelka

VP of Product

Mackenzie Jackson

Field CTO

AI coding agents now install open source dependencies without a person reviewing each package they pull. Every install is a trust decision, and review processes built around a developer at a keyboard no longer cover them. Strict supply chain governance requires controls that run automatically where a package is requested, installed and executed.

Alison Sickelka, VP of Product at Cloudsmith, and Mackenzie Jackson, Field CTO at Aikido Security, follow one package request from request to production so you can see which checks belong where. Alison starts at the registry, where policy blocks, flags or conditionally allows the package. Mackenzie picks it up on the assumption it passed those checks, and shows what Aikido Security detects on the developer machine, in CI, and at runtime.

The two end on the handoff: which signals a registry can pass to downstream detection, and which detection signals can return to policy. You should leave able to name the control points you have, the ones you're missing, and how each keep you more secure. Live Q&A at the end.

Registry-to-runtime handoff for securing AI-driven supply chains

The CRA's reporting obligations start September 11 and SBOM requirements land in 2027. Cloudsmith and AWS on producing the evidence without slowing releases.

James Matchett

Head of Security

Jonah Craig

Senior Solutions Engineer

The CRA's reporting obligations start on September 11. Manufacturers now have 24 hours to report an actively exploited vulnerability to ENISA. December 2027 adds SBOMs, CE marking, and secure-by-design principles.

CRA obligations build on the foundation of knowing what is in your software, and being able to prove it. Providing that proof within the established reporting windows is a governance issue before it is a paperwork issue. Manually auditing systems will likely take more than the 24 hours teams have.

James Matchett, Head of Security at Cloudsmith, and Jonah Craig, Sr. Solutions Architect at AWS, cover what the CRA asks of engineering teams and how to produce the evidence without slowing releases. They will discuss mechanisms that automate policy enforcement at the point of ingestion, SBOMs generated on every push, and logging that holds up during an audit. If your program already runs SCA or other scanning, you'll learn where ingestion controls fit and what they add to your security posture.

Compliance needs proof: Securing the software supply chain under the CRA

Open source attacks are showing up in new places – GitHub repos, npm packages, VS Code extensions. Join OpenSourceMalware and Cloudsmith to break down these new attack vectors, and what companies can do to mitigate these more advanced bad actors.

Jenn Gile

Founder

Nigel Douglas

Head of Developer Relations

Malicious open source packages keep finding new doors. Compromised GitHub repositories, state-linked npm packages, and attacks leveraging VS Code extensions installed by developers without a second thought – the attack surface keeps expanding, and most teams are trying to figure out how to keep up.

As co-founder of OpenSourceMalware, she runs the community-driven threat intel that companies rely on, watching malicious packages, domains, and infrastructure in real time across npm, GitHub, and beyond. In this session, she joins Cloudsmith’s Head of Developer Relations, Nigel Douglas, to go through what's changed in the last month or two, why it matters, and what to do before it becomes your postmortem. Then the floor is yours: stay for a live AMA and ask what's on your mind.

Inside the open source malware attack landscape

This webinar explores how platform engineering teams are uniquely positioned to operationalize CRA compliance - turning regulatory obligations from the security and compliance teams into frictionless, automated workflows that developers will actually adopt.

Your newest developer is an AI agent. Who’s governing it?

Things you'll learn

Speakers

Summary

Registry-to-runtime handoff for securing AI-driven supply chains

Compliance needs proof: Securing the software supply chain under the CRA

Inside the open source malware attack landscape

How platform engineering unlocks CRA readiness