Codex Desktop macOS: Computer Use skill loads but required node_repl tool is not exposed
Nobody has claimed this yet.
Assessment
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Newbie friendliness
- 45/100
- Issue type
- Bug
- Clarity
- Mostly clear
- Activity status
- Quiet
- Tech stack
- javascript, macos
- Domain
- desktop-dev, devtools
Research direction
Start by reproducing the failure through the @Computer entry point with the provided prompt, then inspect task-level provisioning for node_repl and mcp__node_repl__js. Compare the healthy Workspace Dependencies diagnostics and running bundled processes with the task’s callable tool panel. Done means the Computer Use skill receives the required tool and can list open applications without a fallback bridge.
Written by the indexing model from the issue text.
Description
What version of the Codex App are you using (From “About Codex” dialog)?
26.707.91948
What subscription do you have?
ChatGPT Plus
What platform is your computer?
Darwin 25.5.0 arm64 arm
What issue are you seeing?
The official Computer Use plugin loads, but the task is not given the node_repl JavaScript tool required by the Computer Use skill.
Using @Computer with this prompt:
List my currently open applications. Do not change anything. Do not use terminal commands or create a fallback bridge. If the required Computer Use tool is unavailable, stop and report the exact missing tool.
produces this exact response:
Cannot proceed: the required
node_replComputer Use tool is unavailable in this task. No changes were made.
The plugin is enabled, Workspace Dependencies diagnostics report healthy, and codex mcp get node_repl reports the server as enabled. Resetting and reinstalling Workspace Dependencies did not fix it.
What steps can reproduce the bug?
-
Enable the official
computer-use@openai-bundledplugin. -
Confirm Codex Workspace Dependencies are enabled.
-
Run Workspace Dependencies diagnostics and confirm they report healthy.
-
Start a new Codex Desktop task.
-
Invoke
@Computer. -
Enter:
List my currently open applications. Do not change anything. Do not use terminal commands or create a fallback bridge. If the required Computer Use tool is unavailable, stop and report the exact missing tool. -
The Computer Use skill loads, but reports that its required
node_repltool is unavailable.
Session/task ID:
019f6c3b-3dba-7bd2-a43f-d59aab5aed5d
I also reset and reinstalled Workspace Dependencies, restarted Codex, and reproduced the failure again.
What is the expected behavior?
Invoking the official Computer Use plugin should expose the Node REPL JavaScript tool required by its skill—apparently mcp__node_repl__js.
The skill should then initialize its bundled Computer Use client and list the open applications without using terminal commands or a fallback bridge.
Additional information
Environment:
- Codex Desktop: 26.707.91948
- Codex CLI: 0.144.5
- Computer Use plugin: 1.0.1000387
- Reproduced: July 16, 2026, approximately 2:40–3:50 PM EDT
- Workspace Dependencies diagnostics: Healthy
- Workspace Dependencies reinstalled: Yes, with no effect
node_replMCP configuration: Enabled- Bundled Node REPL and Computer Use processes: Running
I manually registered the bundled Computer Use client as a direct stdio MCP server with mcp as its argument. The underlying server exposed tools such as list_apps, get_app_state, click, and type_text, and a terminal-side bridge successfully called list_apps.
However, neither the normal Computer Use JavaScript tool nor the manually registered direct MCP tools were exposed natively in the task's callable tool panel.
This suggests that the underlying Computer Use service is healthy and that the failure occurs during task-level tool provisioning.
Possibly related Browser Use report: https://github.com/openai/codex/issues/20579
- Dominant language
- Rust
- Stars
- 125k
- Forks
- 19.5k
- Avg merge
- 1m
- Merged PRs (30d)
- 1k
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from openai/codex
-
enhancement remote
Difficulty 2/5 1-3 hours Newbie friendliness 68/100
-
bug CLI windows-os
Difficulty 2/5 1-3 hours Newbie friendliness 76/100
-
macOS sandbox blocks hw.optional.arm64 sysctl, causing Flutter to misdetect Apple Silicon as x64 Openbug CLI sandbox
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
-
bug CLI TUI
Difficulty 2/5 1-3 hours Newbie friendliness 90/100
-
CLI config enhancement skills
Difficulty 2/5 1-3 hours Newbie friendliness 84/100
Similar issues
-
Difficulty 2/5 1-3 hours Newbie friendliness 86/100
kwakseongjae/auto-hwp#319 ·
-
area:cli bug filter-quality good first issue priority:medium
Difficulty 2/5 1-3 hours Newbie friendliness 84/100
-
Difficulty 1/5 Under an hour Newbie friendliness 72/100
bevyengine/bevy#25861 ·
-
comp-datalake
Difficulty 2/5 1-3 hours Newbie friendliness 88/100
ClickHouse/ClickHouse#121222 ·
-
A-linter
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
oxc-project/oxc#26863 ·