$ loading_
Enable Claude Code to perform coding tasks through the OpenAI Codex CLI.
The available materials suggest this is an open-source MIT-licensed MCP wrapper with no extra secrets required and no declared standalone remote endpoints, with no clear high-risk red flags. The main concern is that it executes/orchestrates Codex CLI locally, while sparse documentation leaves the real execution scope, data access boundaries, and possible indirect network behavior to be verified from source and runtime behavior.
The materials explicitly state that no secrets or environment variables are required; there is no indication that this tool itself requests or stores additional credentials. Note that the underlying OpenAI Codex CLI may have its own authentication requirements, but that is not declared in the provided materials.
No standalone remote endpoint is declared, so the MCP server itself has no clearly documented egress target in the materials. However, its purpose is to let Claude Code interact with Codex CLI, and the underlying CLI may perform network communication. With no README, the materials do not sufficiently confirm whether user data is sent out indirectly via the CLI and to which destination.
The system checks explicitly mark this tool as executes-code, and the description indicates it interacts with the local OpenAI Codex CLI, implying the ability to start or control local CLI processes. This is a typical but security-relevant MCP capability, so execution scope should be constrained and ideally isolated.
The materials do not specify which files or directories can be read or written, but as an MCP tool that executes a local CLI, its effective data access scope will typically inherit the permissions of the account running the process. There is no evidence of excessive authorization beyond the stated function, but the boundary is unclear due to sparse documentation.
Positive factors are that it is open source, auditable, and MIT-licensed, which materially lowers supply-chain risk. However, it comes from a third-party registry, has 0 stars, unknown maintenance status, and no README, so maturity and maintenance signals are weak and the source code and dependencies should be reviewed before trust.
Copy the install command and let the AI configure it · recommended for beginners
No copy-paste install info for "codex-cli-mcp-tool" yet — see the docs or source repo.
Use codex-cli-mcp-tool to call the Codex CLI and generate a Python script that reads a CSV file, aggregates sales by city, and outputs the top five cities.
A runnable Python script that completes the requested data aggregation task.
Use codex-cli-mcp-tool to call the Codex CLI to refactor the JavaScript utility functions in the current project, remove duplicated logic, and add brief comments for each function.
Refactored code suggestions or applied changes that improve readability and reusability.
Use codex-cli-mcp-tool to call the Codex CLI to inspect the Python code in the current repository, identify obvious errors, style issues, and potential risks, and provide fix suggestions.
A summary of code review results with identified issues and corresponding fix suggestions.
Use Codex in Claude Code for execution, debugging, and web search.
Connect OpenAI Codex CLI to MCP clients for headless coding sessions.
Get technical advice, code reviews, and explanations via Codex CLI.
Delegate coding tasks to Codex via MCP with security and result checks.
Connect IDEs or AI assistants to Codex CLI for safe automation and code analysis.
Securely run Claude Code CLI via MCP for tasks, memory, and IDE integration.