vitest seems to be an issue now that we have code agents

[rhysburnie]

Clear and concise description of the problem

vitest on its own is vitest watch which I have always thought was great.

However, with so much ai agent support now I have noticed it can be an issue if you are allowing the agent to run tests etc. In an agent flow a watch isn't great and I have found it lead to issues for obvious reasons.

Obviously the answer is to make your test script in package.json use vitest run instead but I feel like run may be a better default now?

Like I said, I love that watch is default, but now Im starting to dip toe in ai assist it feels wrong as default because it can block the ai flow.

Suggested solution

Change default to vitest run so ai agents especially in yolo mode actually get the signal that tests have completed.

Alternative

Leave as is.
This is just an observation from someone who isn't even using AI heavily right now but feels like it interupts the ai flow.

Additional context

• Cursor
• VSCode w/ agent

Sometimes especially in yolo mode agent can run test to check if code working and if it defaults to watch then it can get hung up.

Validations

• Follow our Code of Conduct
• Read the Contributing Guidelines.
• Read the docs.
• Check that there isn't already an issue that request the same feature to avoid creating a duplicate.

[hi-ogawa]

This is being discussed in #7673.