Today, every task the agentic mode needs to run a yarn lint to verify if there’s any error within project, as lints or even TS, but those errors usually already are showing at the editor, because lint usually is configured and reporting those errors at editor.
Cursor actually can’t access it? It needs to perform a lint
every time to handle those errors? Asking because this could make the execution steps way faster, as it would be able to see problems while writing the code.
Hey, you can enable a linting step in the settings, that lets Composer fix the linting errors that might have shown up from the changes it makes.
However, many people have projects that already contain low priority linting errors and we don’t want to have Composer get brought off topic by the constant linting errors it can see, so we don’t show them to the Composer in normal use!
Does the linting step not do what you are looking for?
Hey @danperks , thanks by taking time to answer
Yeah, linting does work for me indeed, works very well! The thing here is usually cursor apply changes, without tanking care of linting issues, them it runs a command yarn lint
, npm lint
, catchup the issues and then keeps going - this takes time… My point here, is - what if while he’s creating the code, he’s aware of the linting errors/rules in order that doesn’t need to refact every change.
As i was writing this aswer i’ve realized the best approach here would i just pass the eslint file so it creates the code with the correct syntax lowering the refactors with eslint… This migth work 