Luis Vieira
Free Lesson

Improve Your Coding Agent with Skills and Evals

30 min
Oct 15, 2026 12:00 PM

By continuing, you agree to Maven's Terms and Privacy Policy.

What you'll learn

Turn a recurring mistake into a testable skill

Identify a failure in your agent's work, define the expected result, and capture a procedure that could prevent it.

Compare your agent with and without the skill

Build a small evaluation using realistic tasks and executable checks

Decide which changes are worth keeping

Understand the effectiveness of your skills and different models using metrics such as: task success, time, and cost

Why this topic matters

You add a skill. Rewrite your project instructions. Try a different model. However You can't tell if your coding agent is actually producing bette results. In this lesson, we'll work through a coding example, turn an observed failure into a focused skill, and compare versions using a simple evaluation. You'll get the tools to decide whether your change actually made your agent better as agents models and skills constantly change and evolve.

You'll learn from

Luis Vieira

Luis Vieira

Staff Engineer & Agentic AI Educator

See all products from Luis
Get free access