$ loading_
Forward MCP tools to coding agents for session-based development tasks.
The material indicates this tool mainly forwards normalized MCP tools to Codex or Claude Code for agentic coding tasks. It requires no secrets, declares no remote endpoint, and is open-source under MIT, so overall risk is relatively low; however, caution is still warranted because it is tied to code-execution-style capabilities and has weak community/maintenance signals.
The material explicitly states that no keys or environment variables are required, and there is no request for API tokens, account credentials, or other sensitive authentication material, so credential exposure appears limited.
No remote endpoints are declared in the material, and it does not describe uploading user data to third-party services. It only says it forwards MCP tools to Codex or Claude Code, but provides no concrete evidence of actual network targets or egress mechanisms.
The system flags this tool as executes-code, and the description mentions agentic coding tasks, starting new sessions, continuing sessions, and checking status, indicating ordinary local code-execution or process-control capabilities. This is an inherent capability of such tools and warrants caution in a restricted environment.
The description does not explicitly list readable or writable paths, but as an MCP tool for agentic coding, it would typically interact with project code, session state, or related local workspace data. The current material does not prove overbroad access, but it is reasonable to assume potential access to data relevant to coding tasks.
Positive factors include open-source code, an MIT license, and auditability. However, it comes from a third-party registry, has no README content, shows 0 stars, and has unknown maintenance status, which reduces verifiability and maturity; therefore the supply-chain posture warrants caution rather than high risk.
Copy the install command and let the AI configure it · recommended for beginners
No copy-paste install info for "codemcp" yet — see the docs or source repo.
Use codemcp to send this task to Claude Code: initialize an authentication module for a Node.js API project with JWT login, auth middleware, and basic tests, then return the session ID.
Returns the new session details and starts generating code for the authentication module.
Use codemcp to continue session abc123, finish the remaining integration tests, fix the current test failures, and summarize the files changed.
Continues work in the existing session and returns progress, fixes, and a summary of changes.
Use codemcp to check the current status of session xyz789, tell me whether the task is complete, what the latest step was, and whether you need more context from me.
Returns the session status, recent execution log, and any information needed for the next step.
Delegate coding tasks to Codex via MCP with security and result checks.
Enable Claude Code to perform coding tasks through the OpenAI Codex CLI.
Get technical advice, code reviews, and explanations via Codex CLI.
Query Claude Code transcript analytics for cost, safety, audit, and efficiency insights.
Delegate coding and debugging from Claude Code to OpenAI Codex with auto-recovery.
Coordinate multiple AI coding agents to build, review, and remember together.