Quiz 2 Question 10 of 20

A GitHub Copilot agent is designed to flag potential security vulnerabilities in pull requests. The team evaluates it solely on the number of flags it produces per PR. After one month, developers are ignoring the agent's feedback entirely. What evaluation misalignment caused this?

Select an answer to reveal the explanation.

Motivation