codex-mcp-server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {} |
| logging | {} |
| prompts | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| ask-codexA | Use OpenAI Codex to analyze, review, edit, or generate code. Call this tool whenever the user mentions "codex", "use codex", or wants a second opinion from OpenAI models on code. Supports file references with @ syntax (e.g. @src/), model selection (see list-models), reasoning effort, and structured edits via changeMode. Read-only by default; set sandbox: true to let Codex edit files. For reviewing git changes prefer the review tool. |
| batch-codexA | Run multiple tasks through OpenAI Codex in batch. Use when the user wants Codex to handle several tasks sequentially — mass refactoring, bulk code changes, or automated transformations. |
| reviewA | Run a Codex code review (codex review) on git changes: uncommitted work, a branch against a base, or a single commit. Returns prioritized findings with file:line references. Read-only. |
| list-modelsA | List the model IDs and reasoning effort levels the installed Codex CLI supports. Call this before passing a model or reasoningEffort to other Codex tools. |
| pingC | Echo |
| HelpC | receive help information |
| versionA | Display version and system information |
| brainstormB | Brainstorm ideas using OpenAI Codex with structured frameworks (SCAMPER, design-thinking, lateral, etc.). Use when the user wants creative ideation, brainstorming, or idea generation via Codex. |
| fetch-chunkA | Retrieves cached chunks from a changeMode response. Use this to get subsequent chunks after receiving a partial changeMode response. |
| timeout-testA | Test timeout prevention by running for a specified duration |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| ask-codex | Execute Codex CLI with optional changeMode |
| batch-codex | Execute multiple atomic Codex tasks in batch mode for efficient automation |
| review | Review git changes with Codex |
| list-models | Show models and reasoning levels supported by the local Codex CLI |
| ping | Echo test message with structured response. |
| Help | receive help information |
| version | Get version information for Codex CLI and MCP server |
| brainstorm | Create structured brainstorming with chosen methodology and analysis |
| fetch-chunk | Fetch the next chunk of a response |
| timeout-test | Test the timeout prevention system by running a long operation |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 10 tools
Core Codex tools are mostly distinct: ask-codex handles general requests, batch-codex handles multiple sequential tasks, review targets git changes, and brainstorm targets ideation. There is some overlap because ask-codex is broad enough to cover review, editing, and idea generation, but the descriptions provide clear routing guidance.
Most names use kebab-case, but the set mixes verb_noun names like ask-codex and list-models with single-word names like review, brainstorm, ping, and version, plus the capitalized Help. The convention is readable but not fully consistent.
Ten tools is a reasonable size for a Codex integration, with six domain tools and four generic utility tools. The count is well within range, though ping, timeout-test, and Help are boilerplate rather than core Codex capabilities.
The surface covers the main Codex workflows: general requests, batch processing, code review, brainstorming, model discovery, and chunk retrieval. Minor gaps exist around task status, cancellation, or conversation continuation, but core workflows are covered.