from sklearn.linear_model import LogisticRegressionCV
from sklearn import datasets
iris = datasets.load_iris()

LogisticRegressionCV(scoring="f1").fit(iris.data, iris.target)

This runs and I think it should error.
By default f1 uses binary averaging, which makes no sense for multiclass IIRC.
I'm sure @jnothman knows more about this than me.
I'd argue this is a bug in the f1_score which should error if average='binary' but the targets are not binary.
Though I vaguely remember that we discussed that we want to enable this behavior to compute the f1-score for a single class that is selected by pos_label but that seems to be covered by
this

The class to report if average='binary' and the data is binary. If the data are multiclass or multilabel, this will be ignored; setting labels=[pos_label] and average != 'binary' will report scores for that label only.

The docstring seems exactly right to me but the behavior is different.