MCP servers not detected in Codex VS Code extension (but working in Codex CLI)

[PangeranSilaen]

What version of the VS Code extension are you using?

0.4.38

What subscription do you have?

Plus

Which IDE are you using?

VS Code

What platform is your computer?

Microsoft Windows NT 10.0.26100.0 x64

What issue are you seeing?

Hi, I'm encountering an issue where Codex CLI successfully connects to local and remote MCP servers (e.g. Playwright, Supabase), but the Codex VS Code extension fails to detect or list any MCP resources/tools.
This seems to be a session isolation problem between Codex CLI and the Codex IDE extension.

Environment

• OS: Windows 11 (Build 26100.x)
• Codex CLI version: v0.56.0
• Codex VS Code extension: Latest (from OpenAI official marketplace)
• Node.js: v22.13.0
• Playwright MCP: latest via @playwright/mcp@latest
• Supabase MCP: https://mcp.supabase.com/mcp
• config.toml path: C:\Users\hi\.codex\config.toml

What steps can reproduce the bug?

1. Create a .codex/config.toml file:

model = "gpt-5-codex"
model_reasoning_effort = "low"
sandbox = "danger-full-access"

[features]
rmcp_client = true

[mcp_servers.playwright]
command = "cmd"
args = [
  "/c",
  "C:\\Program Files\\nodejs\\npx.cmd",
  "@playwright/mcp@latest",
  "--browser", "chrome",
  "--extension",
  "--no-sandbox"
]
startup_timeout_sec = 120
tool_timeout_sec = 600

[mcp_servers.playwright.env]
APPDATA = "C:\\Users\\hi\\AppData\\Roaming"
LOCALAPPDATA = "C:\\Users\\hi\\AppData\\Local"
HOME = "C:\\Users\\hi"
SystemRoot = "C:\\Windows"
NODE_OPTIONS = "--dns-result-order=ipv4first"

[mcp_servers.supabase]
url = "https://mcp.supabase.com/mcp?project_ref=${SUPABASE_PROJECT_REF}"
bearer_token_env_var = "SUPABASE_ACCESS_TOKEN"
startup_timeout_sec = 120
tool_timeout_sec = 600

2. Run in Codex CLI:

   codex
   /mcp
   

   ✅ Output:
   Playwright detected, tools listed, working normally.

3. Open Codex Chat in VS Code (GPT-5-Codex model).

4. Run:

   list_mcp_servers
   

   or

   list_mcp_resources
   

   ❌ Output: empty (no MCP servers detected).

---

What is the expected behavior?

Expected Behavior

Codex IDE should inherit the same MCP configuration and session from ~/.codex/config.toml, or at least spawn MCP processes defined there (just like Codex CLI does).
Playwright MCP should appear in list_mcp_resources and be callable inside Codex IDE.

Actual Behavior

• Codex CLI: Works perfectly, detects MCP Playwright and Supabase.
• Codex IDE (VS Code extension): list_mcp_resources always empty.
• In terminal logs, IDE session doesn’t appear to start MCP processes or read environment variables.
• Restarting VS Code, reloading configuration, and setting env vars globally have no effect.

Additional information

Codex CLI output:

> codex
The supabase MCP server is not logged in. Run `codex mcp login supabase`.

MCP Tools
• playwright
  Status: enabled
  Command: npx @playwright/mcp@latest --browser chrome --extension --no-sandbox
  Tools: browser_click, browser_tabs, browser_wait_for, ...
• supabase
  Status: enabled
  Auth: OAuth
  URL: https://mcp.supabase.com/mcp

Codex VS Code chat output:

There are still no MCP servers connected in this session — list_mcp_resources remains empty.

---

Hypothesis / Possible Cause

The Codex VS Code extension may be launching its own isolated agent session without sharing MCP process context or ~/.codex/config.toml with the CLI environment.
It seems MCP initialization is only performed in CLI-based Codex sessions.
There is currently no manual command (codex mcp start) or extension setting to sync the two.

---

Suggested Fix

• Make the Codex IDE extension automatically detect and initialize MCP servers from the same configuration file (~/.codex/config.toml).
• Optionally, allow Codex IDE to connect to an already-running MCP process from Codex CLI.
• Add a debug log or indicator for MCP connection status in the IDE.

---

Workaround

Currently, the only working approach is:

1. Run codex CLI and keep it open (MCP servers run there).
2. Then interact with MCP commands through that CLI session only.
   The VS Code Codex chat cannot yet call those tools.

This makes MCP development and testing inside VS Code Codex extension impractical, even though the CLI works fine.
Please confirm if MCP integration for Codex IDE is expected to be functional yet, or if this is a known limitation.

[github-actions]

Potential duplicates detected. Please review them and close your issue if it is a duplicate.

• Codex not finding MCP servers using its go-to codex.list_mcp_resources (VS Code Extension) #6215
• MCP Servers not detected in .vscode\mcp.json #5929

Powered by Codex Action

[PangeranSilaen]

Just to clarify, this is not a duplicate of #5929 or #6215.
Both of those refer to older Codex releases and .vscode/mcp.json integration issues on macOS.

My report (#6465) is about Codex VS Code extension (v0.56.0) failing to detect MCP servers already working in Codex CLI, even though both share the same ~/.codex/config.toml configuration.

Reproduction shows CLI detects Playwright/Supabase MCP successfully, while list_mcp_resources in IDE remains empty — indicating the VS Code extension isn't spawning or connecting to the same MCP runtime session.

Tested on Windows 11 + Node v22, so this is cross-platform and not related to .vscode/mcp.json.

Please keep this open — it's a session isolation bug between Codex CLI and IDE, not a config syntax issue.

[Milad]

I have the same problem:
codex-cli 0.57.0
Codex VScode extension: 0.4.38

My ~/.codex/config.toml:

model = "gpt-5-codex"
model_reasoning_effort = "medium"

[projects."/path/to/back-end"]
trust_level = "trusted"

[mcp_servers.postgres]
command = "/usr/local/bin/toolbox"
args = ["--prebuilt","postgres","--stdio"]
env = {"POSTGRES_HOST"= "localhost", "POSTGRES_PORT"= "43256", "POSTGRES_DATABASE"= "my_db_name", "POSTGRES_USER"= "my_db_usernamename","POSTGRES_PASSWORD"= "my_db_pass"}

Codex CLI can access the MCP server and can list the tables of the DB.
While Codex VScode extension reports this:

resources/list failed: unknown MCP server 'postgres'

[MatthewGranfield]

Can confirm i have the same issue on a fresh install Win 11 pro.

CLI v0.57.0
VSC extention v0.4.39

Using CLI i can access larvel boost and use the tools, however i cannot through the extention.

[crstamps2]

Also not working on Mac.

CLI v0.58.0
VSC extension 0.4.40

When prompting the chat to check the status of the tools, it's thinking always outputs:

MCP client for figma failed to start: handshaking with MCP server failed: expect initialized result, but received: Some(EmptyResult(EmptyObject))

[XFairesV1]

I have the same problem:

CLI v0.57.0
Extensão VSC v0.4.39

Linux

[Robert-Phan]

I also have the same problem:

CLI v0.58.0
Extension v0.4.40

My config.toml:

model = "gpt-5.1-codex"
model_reasoning_effort = "medium"

[mcp_servers.svelte]
command = "npx"
args = ["-y", "@sveltejs/mcp"]
startup_timeout_sec = 30

Using Windows with WSL. CLI works fine, the extension times out.

[nicolascanese]

I'm having the same issue in Win11 and in WSL.

Extension v0.4.43
CLI v0.58.0

[leike0813]

I'm having the same issue in Win11, WSL not tested.

Extension v0.4.43
CLI v0.58.0

[RyanRiz]

I'm having the same issue in VS Code Insiders Linux
Extension 0.5.46
CLI v0.60.1

.codex/config.toml:

model = "gpt-5.1-codex-max"
model_reasoning_effort = "high"

[mcp_servers.laravel-boost]
command = "php"
args = ["artisan", "boost:mcp"]

chat:

what mcp that currently installed that you can access?

I don’t see any MCP servers/resources exposed in this environment right now—both the resource and resource-template lists came back empty. If you expected some to be available, let me know their names or restart/connect them so I can access them.

but codex recognized installed mcp in CLI

[FelixKuehl]

Same here on MacOS. Any workaround / fixes?