Add external features to build my own metric function with Caret

[ghost]

I am building a classification algorithm with the Caret package. I would like to know how to add external features in the metric function to compute my own metric.

I would like to compute a score based on both the prediction (pred) and an external feature, let's call it score, which doesn't belong to the original features used by the classification algorithm.

It would give something like this:

library(caret)
library(C50)

data(iris)
dataset<-iris

compute_metric = function(data, lev = NULL, model = NULL) {
score = sum(data$pred * data$Petal.Width)
c(metric_score = score)
}

objControl =trainControl(method="repeatedcv",number=10,repeats=3,summaryFunction=compute_metric,classProbs=TRUE)

objModel <- train(x=iris[,c("Sepal.Length","Sepal.Width")], y=iris$Species,
method="C5.0",trControl=objControl,
metric="score",tuneLength=3)
However, it doesn't work.

Does anyone have an idea how to make it work?