Introduction to Agentic Development and AI Tooling
AI coding tools have moved from autocomplete to agents that plan, edit and run code across an entire repository. That shifts the developer's job more than the tooling suggests: the bottleneck is no longer typing, it's context, review and trust.
Details
This talk is a practical introduction to agentic development. We start with workflows that hold up in real projects — where to delegate, where to stay in the loop, and how to keep agent-generated changes small and reviewable.
The core of the talk is control: what the agent knows, and what it can do. On the knowing side — prompting, context enrichment, and memory. On the doing side — tools and skills: which tools an agent should get, how to use them, and how to build your own.
Then security, which most intro talks skip: prompt injection through untrusted content, secrets leaking into context, hallucinated dependencies, and what an agent should never have write access to — including the handful of cheap measures that get you most of the way to safe.
Takeaways: a mental model for what to delegate, a review checklist for AI-written code, practical context management, and a minimum security baseline for agentic coding.