Missing Tool Name in Agent Approval Dialog (Playwright MCP & Browser Automation)

[shaggy2626 (JG)]

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

When an agent attempts to execute a tool (specifically observed with the Playwright MCP or the built-in Browser Automation), the approval dialog in the chat interface fails to display the name of the tool or the specific action being requested. The UI shows the “Executing browser tasks” header and the “Skip/Run” buttons, but the middle section where the tool name/parameters should be is blank. After I click “Run” the tool works fine, its just that the requested tool name is not displayed so I have no idea what tool its trying to call.

Steps to Reproduce

1. Enable the Playwright MCP server OR enable the built-in Browser Automation (configured for either “Google Chrome” or “Browser Tab”).

2. Prompt the agent to perform an action requiring the enabled tool (e.g., “Open a new browser tab and go to example.com”).

3. Observe the approval prompt that appears in the sidebar.

4. Note that the tool name/action description is missing, leaving only the “Skip” and “Run” buttons.

Expected Behavior

The approval dialog should clearly state which tool is being called (e.g., browser_navigate) and show the arguments being passed so the user can verify the action before clicking “Run.”

Screenshots / Screen Recordings

Operating System

Windows 10/11

Version Information

Version: 2.4.22 (user setup)
VSCode Version: 1.105.1
Commit: 618c607a249dd7fd2ffc662c6531143833bebd40
Date: 2026-01-26T22:51:47.692Z
Build Type: Stable
Release Track: Default
Electron: 39.2.7
Chromium: 142.0.7444.235
Node.js: 22.21.1
V8: 14.2.231.21-electron.0
OS: Windows_NT x64 10.0.26200

Does this stop you from using Cursor

No - Cursor works, but with this issue

[deanrie]

Hey, thanks for the report.

This is a known issue and the team is aware. When the Agent calls MCP tools or Browser Automation, the confirmation dialog doesn’t show the tool name and parameters where they should appear.

Like you noticed, the tool itself works fine after you click Run. The issue is only with the UI dialog. For now, you’ll need to rely on the chat context to understand which tool is about to run.

Your report helps us prioritize this. I’ll message you when there are updates.

[shaggy2626 (JG)]

Is It also noted that you can choose “auto run everytime” even though the MCP Tools protection is toggled off? I have to constantly click Run manually each time.