agentic patterns

how it works

• each lesson is one pattern, with a minimal implementation.
• examples are plain typescript pseudocode — no sdk to install, no account to create.
• every pattern ends with when it breaks — the failure mode you'll hit in production, and how to notice before it's expensive.
• lesson seven is evals: how you know a pattern is working, not just demoing well.

lessons

1. 01 what are agentic patterns the shape of the problem. why patterns, not frameworks. the four primitives you'll see in every later lesson. 10m
2. 02 skills — the smallest unit a tool call plus the prompting, validation, and return-shape glue around it. reusable across agents. 14m
3. 03 extensions — wrapping the base loop memory injection, retrieval, safety, telemetry. the pattern equivalent of middleware. 12m
4. 04 personas — identity for agents a bundle of prompt, voice, allowed skills, and guardrails. swap it and the same loop behaves differently. 12m
5. 05 handoffs — multi-agent composition how one agent passes control to another with just enough context. most 'the agent got confused' bugs live here. 14m
6. 06 loops — react, reflection, and when to stop the control flow inside a single agent turn. stopping criteria, budget, and failure modes. 13m
7. 07 evals — knowing the pattern works offline scorecards, trace-based checks, and the gap between demo and production. 15m