Everything OpenAI Codex - An open-source workflow OS for OpenAI Codex.

Open-source operating system for OpenAI Codex workflows: agents, skills, hooks, rules, memory, safety gates, and cross-harness adapters. - mturac/everything-openai-codexEOC turns raw OpenAI Codex into a maintained engineering environment: 60 agents, 230 skills, 110 rules, hook gates, session memory, and adapters for Cursor, OpenCode, Gemini, Zed, Copilot, and Trae. Not a prompt dump — a field-tested workflow system. MIT.

Hey Hunters, I'm Mehmet Turac, the maker. I built EOC because every serious OpenAI Codex project I worked on kept re-discovering the same gaps: no session memory, no quality gates, no reusable skills, no cross-tool portability. Every team was rebuilding the same operating layer from scratch. EOC packages that layer once, properly: - 60 specialized agents (planner, code-reviewer, security-reviewer, language-specific reviewers for TS/Python/Go/Java/Kotlin/Rust/C++/Perl) - 230 skills covering backend, frontend, security, ML, docs, ops, release work, content, and investor materials - 110 rules across 12 language ecosystems - Hook gates (no-verify blocks, supply-chain scanners, MCP health checks) - Session memory + compaction + handoff-friendly logs - Cross-harness adapters for Cursor, OpenCode, Gemini, Zed, Copilot, Trae, and Codex app + CLI v2.0.0-rc.1 ships with a desktop dashboard (Tkinter), the Hermes operator workflow, and an EOC 2.0 Rust control-plane alpha. MIT licensed. Field-tested on daily software work, not a prompt collection. Would love to hear from anyone running Codex at scale — what breaks first in your setup? That's usually where EOC's next skill comes from. GitHub: github.com/mturac/everything-openai-codex

Everything OpenAI Codex - An open-source workflow OS for OpenAI Codex.

Replies