Make One Small Agent Edit and Verify It

Today's Goals

1. Write a file scope and acceptance criteria for a small edit.
2. Ask Agent to state a plan before editing and run relevant checks.
3. Review the diff, test output, and untested risks.

Plain-Language Walkthrough

Define “small” before opening the editor

Yesterday's map is useful only if it tells you where to make the next move. Assume a real local copy confirms that booking/ui/Confirmation.js prints a fixed “morning” label while the booking object already contains the correct slot label. That premise is a teaching fixture, not an observed Cursor result. The D2 request is deliberately narrow: use the existing label in the confirmation view, add or adjust the nearest test, and leave domain, billing, and unrelated styling alone. A small patch changes the least behavior needed to satisfy a written acceptance condition. It is not merely a patch with few lines.

Write the acceptance conditions before asking Agent to edit. For an afternoon booking, the displayed message uses the afternoon label from the booking object. For a morning booking, the existing correct behavior remains. The edit does not change stored slot values, payment calculations, or the public API. The relevant test passes under the project's actual test command, and the final diff contains only files justified by the task. These conditions are observable. “Make it work nicely” is not. A reviewer can now reject a superficially plausible patch without debating taste.

Sign in and link WeChat to read

Sign in, then link your WeChat mini program account on the account page. Linked accounts share one free article per day. Unlock additional articles with rewarded videos in the mini program.

Scan in WeChat to open Dayward AI. Open the same course and chapter, choose the rewarded video, then refresh this page after completion.

Unlocked — refresh to readSign in / scan with WeChat →