The object_usage_linter won't acknowledge bindings from @import or @importFrom

Describe the bug
The object_usage_linter that comes installed with vscode-R's lintr won't recognize functions added to package scope via @import and @importFrom calls. The linter that runs via devtools::check() or calling lintr::lint() directly, however, will. As such, there seems to be a bug in the lintr that comes packaged with vscode-R but not in other instances of lintr.

To Reproduce
Steps to reproduce the behavior:

1. Create an R package via the R console usethis::create_package("Test")
2. Run usethis::use_mit_license()
3. Open the package in VSCode where you have the R extension installed
4. Import deplyr and ggplot2 in the DESCRIPTION file

Package: Test
Title: What the Package Does (One Line, Title Case)
Version: 0.0.0.9000
Authors@R:
person("First", "Last", , "first.last@example.com", role = c("aut", "cre"))
Description: What the package does (one paragraph).
License: MIT + file LICENSE
Imports:
dplyr,
ggplot2,
Encoding: UTF-8
Roxygen: list(markdown = TRUE)
RoxygenNote: 7.3.2

5. Create an R file called test in the R directory R/test.R
6. Paste these functions into test.R

#' Group rows by name
#'
#' This function groups the rows of a data frame by the column "name"
#' @param df A data frame.
#'
#' @import ggplot2
#' @return A grouped data frame.
#' @export
group_by_name <- function(df) {
  data_grouped <- df |>
    dplyr::group_by(.data$name)
  data_grouped
}

#' Group rows by size
#'
#' This function groups the rows of a data frame by the column "size"
#' @param df A data frame.
#'
#' @importFrom ggplot2 .data
#' @return A grouped data frame.
#' @export
group_by_size <- function(df) {
  data_grouped <- df |>
    dplyr::group_by(.data$size)
  data_grouped
}

7. Run devtools::document() and devtools::load_all()
8. Go to problems panel or hover over .data in each of the functions and note the no visible binding for global variable '.data' warnings.

9. Run devtools::check() and see that no errors, warnings, or notes arise
10. Run lintr::lint("R/test.R") and see ℹ No lints found.

Can you fix this issue by yourself? (We appreciate the help)

No, attempting to reinstall the R, the vscode-R extension, and rlanguageserver have not helped. Here is my settings.json file:

{
    "git.autofetch": true,
    "git.confirmSync": false,
    "[r]": {
        "editor.defaultFormatter": "REditorSupport.r"
    },
    "r.rterm.mac": "/usr/local/bin/R",
}

Expected behavior
An @import or @importFrom call should update the global namespace such that the imported function is defined for the object_usage_linter in the lintr package included by the vscode-R extension.

Screenshots

Environment (please complete the following information):

• OS: macOS
• VSCode Version: 1.99.0 (Universal)
• R Version: R version 4.4.2 (2024-10-31)
• vscode-R version: 2.8.4

Additional context
Folks have a lot of issues with the object_usage_linter in lintr but I didn't find any that pertain directly to this, and as mentioned above, this doesn't seem to be an issue when invoking lintr without the vscode-R extension.

Reproduce the warning with the package setup and R/test.R example, then compare the VSCode Problems panel with devtools::check() and lintr::lint("R/test.R"). Trace how the vscode-R extension invokes the packaged lintr object_usage_linter and how @import and @importFrom are handled. Done means imported bindings no longer produce the false .data warnings in VSCode while the existing lintr checks continue to pass.