​Agents don’t fail because the model is bad, they fail because context is unmanaged.

​As teams rely more on coding agents, prompts, docs, rules, and examples quietly become production infrastructure. But unlike code, context is rarely versioned, reviewed, tested, or rolled out safely.

​In this session, we’ll introduce a practical framework for treating context like software. You’ll learn how to package it, version it, distribute it across teams, and prevent it from drifting as your codebase evolves.

​This isn’t theory. It’s a repeatable “context supply chain” you can apply whether you’re a platform team, product team, or developer enablement lead.

​What you’ll learn

• ​Why context should be treated as a first-class software artifact

• ​How to package prompts, rules, docs, and examples into reusable units

• ​Versioning and rollout strategies that won’t break agents in production

• ​How to detect and prevent context drift over time

• ​A concrete workflow you can adapt to your org immediately

​Speakers

​Dru Knox

​Head of Product, Tessl

​​Dru leads Product at Tessl. He brings deep experience in platform and ecosystem development, having helped build two of the largest developer platforms in the world, Android and the web. His work sits at the intersection of product design, developer experience, and systems thinking. Outside of work, he’s drawn to design, game theory, and a bit of armchair philosophy.

​Patrick Debois

​Head of Product DevRel, Tessl

​Patrick Debois is the leads Product DevRel at Tessl. In 2009 he coined the word DevOps by organizing the first devopsdays event, as is now often known as one of the grandfathers of DevOps. He organized conferences all over the world, including AI Native DevCon, to collect and spread new ideas.

​Who this is for

​Platform engineers, staff+ developers, developer experience teams, and anyone responsible for making agents work reliably at scale.