Local dev needs an exit ramp

Starting a local app is easy. Getting back to a clean baseline is where the development loop shows its quality.

I want the lifecycle to be legible from the command line:

• start the process
• stop the process
• reset local state
• inspect the logs

Those are different operations. A stop command should not quietly delete data, and a reset command should not leave the old process holding the port.

pnpm dev:stop
pnpm dev:reset
pnpm dev:logs

The names are not the important part. The ownership is. dev:stop should end the process it started. dev:reset should say which cache, temporary data, lockfile, or socket it removes. dev:logs should make the failed run inspectable instead of sending me back through a browser tab and three terminal windows.

The useful failure cases are ordinary:

• the port is still occupied
• a migration stopped halfway through
• a stale cache changes the behavior
• the previous session left an orphaned process

Each one needs a documented path back to normal. “Restart the app” is not a recovery procedure when the thing that needs restarting is unclear.

An exit ramp is part of the local development interface. Without one, every small failure becomes archaeology. With one, stopping and recovering are routine operations instead of tribal knowledge.