Agentic Engineering
Community/About Agentic Engineering
This skill sets up a working method for engineering projects where AI agents do most of the implementation and humans stay in charge of quality and risk. It covers four core principles: define completion criteria before starting, break work into small independently verifiable units, pick the right model tier for each task, and measure results with evals and regression checks.
AI-generated listing based on the project's public information. It may contain errors: check the original source.
/What Agentic Engineering is for
Break down a feature like user authentication into small, verifiable units
Set up an eval-first loop before implementing a new capability
Decide which model tier to use for a given coding task
Review AI-generated code for edge cases, security, and hidden coupling
Track token usage, retries, and cost per development task