diff --git a/internal/lsp/cmd/capabilities_test.go b/internal/lsp/cmd/capabilities_test.go
index 02f7fab828..aebed67e7a 100644
--- a/internal/lsp/cmd/capabilities_test.go
+++ b/internal/lsp/cmd/capabilities_test.go
@@ -130,15 +130,11 @@ func TestCapabilities(t *testing.T) {
 		t.Fatal(err)
 	}
 	for _, item := range list.Items {
-		// We expect the "editor.action.triggerParameterHints" command for functions and methods.
-		if item.Kind == protocol.MethodCompletion || item.Kind == protocol.FunctionCompletion {
-			continue
-		}
 		// All other completion items should have nil commands.
 		// An empty command will be treated as a command with the name '' by VS Code.
 		// This causes VS Code to report errors to users about invalid commands.
 		if item.Command != nil {
-			t.Errorf("unexpected command for non-function completion item")
+			t.Errorf("unexpected command for completion item")
 		}
 		// The item's TextEdit must be a pointer, as VS Code considers TextEdits
 		// that don't contain the cursor position to be invalid.
diff --git a/internal/lsp/completion.go b/internal/lsp/completion.go
index 0c3cc97795..cd94bbf0b4 100644
--- a/internal/lsp/completion.go
+++ b/internal/lsp/completion.go
@@ -119,15 +119,6 @@ func toProtocolCompletionItems(candidates []source.CompletionItem, rng protocol.
 			Preselect:     i == 0,
 			Documentation: candidate.Documentation,
 		}
-		// Trigger signature help for any function or method completion.
-		// This is helpful even if a function does not have parameters,
-		// since we show return types as well.
-		switch item.Kind {
-		case protocol.FunctionCompletion, protocol.MethodCompletion:
-			item.Command = &protocol.Command{
-				Command: "editor.action.triggerParameterHints",
-			}
-		}
 		items = append(items, item)
 	}
 	return items