> For the complete documentation index, see [llms.txt](https://docs.tuannvm.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.tuannvm.com/docs/todo.md).

# Codex MCP Server - TODO

## Features from Codex CLI v0.98.0

These features were introduced/stabilized in Codex CLI v0.98.0 but are not yet implemented in this MCP server.

### High Priority

#### \[ ] Steer Mode Support

* **Status**: Stable & enabled by default in Codex CLI v0.98.0
* **Description**: Allow redirecting agents during execution without stopping them
* **CLI Flag**: `--steer` (now default)
* **Implementation Notes**:
  * Add `steerMode` parameter to CodexToolSchema
  * Pass `--steer` flag to codex exec commands
  * Consider whether MCP needs to handle streaming input for steering
* **Reference**: [v0.98.0 Release Notes](https://github.com/openai/codex/releases/tag/rust-v0.98.0)

### Medium Priority

#### \[ ] Collaboration Mode

* **Status**: Naming unified in v0.98.0
* **Description**: Multi-agent parallel collaboration support
* **Implementation Notes**:
  * Add `collaborationMode` parameter (enum: `none`, `collaborate`)
  * Update command flags accordingly
* **Reference**: Collaboration mode naming synced across prompts, tools, and TUI

#### \[ ] Enhanced Structured Content

* **Status**: Text + image content items for dynamic tool outputs in v0.98.0
* **Description**: Better support for dynamic tool outputs with mixed content
* **Implementation Notes**:
  * Current `structuredContent` support is partial
  * May need enhancement to handle text + image content items
* **Reference**: #10567

### Low Priority

#### \[ ] Personality Mode

* **Status**: Pragmatic restored as default in v0.98.0
* **Description**: Control Codex's response personality
* **Options**: `pragmatic` (default), `verbose`
* **CLI Config**: `personality = "pragmatic"` or `personality = "verbose"`
* **Implementation Notes**:
  * Add `personality` parameter to CodexToolSchema
  * Pass via `-c personality="..."`
* **Reference**: #10705

***

## Implemented in v1.3.4+

### ✅ GPT-5.3-Codex Model

* **Status**: Implemented
* **Description**: New default model
* **Changes**:
  * Updated `DEFAULT_CODEX_MODEL` constant to `'gpt-5.3-codex'`
  * Updated tool definitions to reflect new default
  * Single source of truth for model updates

### ✅ Reasoning Effort: 'none' and 'xhigh'

* **Status**: Implemented (commit 448fa3c)
* **Description**: Extended reasoning effort options
* **Changes**:
  * Added `'none'` and `'xhigh'` to reasoningEffort enum
  * Full range: `none`, `minimal`, `low`, `medium`, `high`, `xhigh`

### ✅ Web Search Tool

* **Status**: Implemented
* **Description**: Web search capability using Codex CLI
* **Changes**:
  * Added `websearch` tool with `query`, `numResults`, and `searchDepth` parameters
  * Leverages Codex CLI's natural web search capability through crafted prompts
  * Supports streaming progress updates for long-running searches

***

## Future Considerations

### Model Version Management

* Consider adding a `getAvailableModels()` tool to query Codex CLI for available models
* This would make the server more resilient to future model additions

### Configuration File Support

* Codex CLI supports config files (`.codexrc.toml`)
* Consider whether MCP server should expose config file options

### Streaming Support

* Codex CLI supports SSE streaming for responses
* Consider adding streaming support for long-running tasks

***

## References

* [Codex CLI Releases](https://github.com/openai/codex/releases)
* [Codex Changelog](https://developers.openai.com/codex/changelog/)
* [v0.98.0 Release](https://github.com/openai/codex/releases/tag/rust-v0.98.0)


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.tuannvm.com/docs/todo.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.