What seems like a straightforward calculation of group means and standard errors doesn’t work and I can’t figure out why. I get NA for the std. errors

data("AddHealth", package = "heplots")
library(dplyr)

# use a function
std.error <- function(x, na.rm = TRUE) {
  sd(x, na.rm = na.rm) / sqrt(length(x))
}

# try it two ways
AddHealth |>
  group_by(grade) |>
  summarise(depression = mean(depression),
            anxiety = mean(anxiety),
            n = n(),
            se.dep = std.error(depression),
            se.anx = sd(anxiety, na.rm = TRUE) / sqrt(n)) |> 
  print()

I get NA for the standard error whether I use my function or do the calculation directly.
This should be relatively simple. Why doesn’t this work? How can I get what I want?

# A tibble: 6 × 6
  grade depression anxiety     n se.dep se.anx
  <ord>      <dbl>   <dbl> <int>  <dbl>  <dbl>
1 7          0.881   0.751   622     NA     NA
2 8          1.08    0.804   664     NA     NA
3 9          1.17    0.934   778     NA     NA
4 10         1.27    0.956   817     NA     NA
5 11         1.37    1.12    790     NA     NA
6 12         1.34    1.10    673     NA     NA