Reading
A proxy layer designed for OpenAI Codex MCP servers that enables parallel execution via readOnlyHint injection, supporting model whitelisting, reasoning intensity control, and custom tool extensions.
Section 01
Codex Proxy MCP is a lightweight proxy layer designed for OpenAI Codex MCP servers. Its core function is to enable parallel execution of Claude Code by injecting readOnlyHint, while supporting model whitelisting, reasoning intensity control, custom tool extensions, and graceful degradation mechanisms—addressing the latency issue caused by the serial execution of the native Codex MCP server.
Section 02
The Model Context Protocol (MCP) is a standard interface for AI assistants to interact with external tools, extending AI capabilities. However, when Claude Code communicates with Codex servers via MCP, the native execution mode is serial—even independent calls are processed one at a time, leading to latency in complex tasks.
Section 03
Codex Proxy MCP sits between Claude Code and the native Codex MCP server. Its core design principle is to not alter the underlying tool implementation, only modify the way tools are described. By intercepting tools/list requests in the MCP protocol, it injects the readOnlyHint: true annotation into all Codex tools, informing the client that the tools are read-only and can be executed in parallel (though they may actually modify state—this is a "white lie").
Section 04
Section 05
Written in Node.js and run as an independent stdio process. It intercepts two types of messages: 1. tools/list responses: Inject readOnlyHint and modify parameter modes to add whitelist and reasoning intensity options; 2. Tool call requests: Route custom tools to user processors, and pass standard Codex tools through to the underlying server. It maintains MCP protocol compatibility while extending capabilities.
Section 06
Section 07
Section 08
Codex Proxy MCP is an elegant solution to Claude Code's serial execution bottleneck. It enables parallel capabilities via a lightweight proxy layer, demonstrating open-source community innovation. It significantly improves efficiency for power users, especially in scenarios requiring multiple information retrieval or validation. Additionally, it provides a proxy pattern reference for the MCP ecosystem, highlighting the value of metadata.
Keep going with more reads from the same topic.
Splinter is a minimalist, high-performance key-value (KV) and vector storage system enabling zero-latency inter-process communication via shared memory and atomic operations. With only 766 lines of core code, it supports millions of operations per second and 768-dimensional vector storage, offering a new architectural approach for local LLM inference and data-intensive applications.
Folkering OS is the world's first AI-native bare-metal operating system, entirely written in Rust no_std without relying on Linux, POSIX, or libc. It can generate commands from scratch, compile them into WASM, and run them in 10 seconds, achieving true self-evolution.
This project explores how to use Large Language Models (LLMs) to detect logical vulnerabilities in Ethereum smart contracts, especially business logic flaws that are difficult for traditional static analysis tools to capture, such as economic manipulation and execution order errors.
Cross-generational hardware testing based on NVIDIA A10G and H100 GPUs, comparing and analyzing the throughput, latency, and concurrent scalability of three mainstream LLM inference frameworks. SGLang achieves a 3.4x performance advantage over vLLM on H100, while Ollama faces architectural bottlenecks in high-concurrency scenarios.