fix: Adapt the test running in E2E to new VS Code task pick logic

It seems that it will not try to fuzzy match the items anymore but
rather rely on the user to pick the correct task type/group first,
before showing the relevant options.

This fixes the issue by explicitly specifying the "hardhat" task type.

test/e2e/tests/commands/clean.test.ts

@@ -3,7 +3,6 @@ import { existsSync, mkdirSync } from "fs";
 import path from "path";
 import * as vscode from "vscode";
 import { waitForUI } from "../../helpers/editor";
-import { sleep } from "../../helpers/sleep";
 import { getRootPath } from "../../helpers/workspace";
 
 suite("task - clean", function () {
@@ -20,22 +19,11 @@ suite("task - clean", function () {
 
     // Run clean task
     await vscode.commands.executeCommand("workbench.action.tasks.runTask", {
-      task: "hardhat: clean main",
+      type: "hardhat",
+      task: "clean",
     });
     await waitForUI();
 
-    await sleep(1000);
-
-    await vscode.commands.executeCommand(
-      "workbench.action.acceptSelectedQuickOpenItem"
-    );
-    await sleep(1000);
-
-    await vscode.commands.executeCommand(
-      "workbench.action.acceptSelectedQuickOpenItem"
-    );
-
-    // Wait for task to finish
     await new Promise((resolve) => vscode.tasks.onDidEndTask(resolve));
 
     // Assert the artifacts were removed