ECX plugin
ECX ties Claude Code and Codex together with shared memory so each sees the other latest messages, plus 19 skills and 21 hooks.
- 89 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- Hooks fire in every session, run tests and edit Claude Code and Codex settings
- `ecx duo` lets two agents work on a repository, including file edits and commits
- Notes and a memory copy go to a private GitHub repo and the graph can send text to a cloud model
Install
- Works in
Manual install
git clone https://github.com/saksham10arora-dotcom/everything-claude-codex ~/ecxThen `cd ~/ecx && ./install.sh`. For the full stack add `--vault ~/vault --name Ada --with-search --with-cron --with-phone`; `--dry-run` only shows the plan.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Hooks fire in every session, run tests and edit Claude Code and Codex settings
- `ecx duo` lets two agents work on a repository, including file edits and commits
- Notes and a memory copy go to a private GitHub repo and the graph can send text to a cloud model
This is third-party code. Review the repository files before installing.
What it does
ECX, short for everything claude codex, installs next to your own skills and adds a shared layer for two agents. The `agent-sync` hook hands each agent the latest lines of the other chat before every message, Codex voice transcripts included. Both start from one `hot.md` file and shared memory, and can mail each other through `sync.py send`. The set has 19 skills for the working week (`/boot`, `/daily`, `/weekly`, `/catchup`, `/standup`, `/ship`, `/wiki` and more), 21 hooks and three note search engines: by meaning through ChromaDB, by relationship through a LightRAG graph and by image through CLIP. A tests hook sends an agent back while checks fail. A danger guard blocks `rm -rf ~` and asks before `git push --force`. `ecx duo` runs Claude Code and Codex on one task on a separate branch while you watch their shared chat.
Who it is for. People who work in both Claude Code and Codex and are tired of retelling one agent what the other decided.
Good fit when
- A task started in Claude Code is easier to finish in Codex and you do not want to retell the context
- Notes live in an Obsidian vault and you need search by meaning, relationship and slide images
- An agent says "done" while tests fail and you need a hook that sends it back
- Notes from your phone with the laptop shut should reach the vault through a private GitHub repo
- You want to give a hard task to two agents at once and compare their plans before work begins
Not a fit when
- You only need a skills pack for one agent: ECX is built around two agents and shared hooks
- You have no notes or vault: most features depend on your vault
- Note text must not leave your machine: the relationship graph uses an Ollama cloud model by default
Example request
Run `/catchup` from my Codex window, continue the same task in Claude Code and run the tests before saying it is done.Limitations
Python 3.9 or newer is required. Shared memory and agent sync run locally, while the relationship graph sends note text through an Ollama cloud model by default; set `ECX_LLM_MODEL` for a fully local setup. Phone mode syncs notes and a memory copy to your private GitHub repo and needs a one-time Codex connector setup in a ChatGPT account. The Codex app-server protocol behind `ecx duo` is marked experimental. The author calls the six-task SWE-bench result a first signal and not proof.
How to disable. Run `./install.sh --uninstall --vault ~/vault` with the same vault you installed with. The installer merged settings with a `.ecx-bak` backup, which you can restore by hand before removing.
FAQ
How do I install ECX?
Clone the repository into `~/ecx` and run `./install.sh`. It asks where your notes live and your first name, merges settings without overwriting your files and saves backups. Then open Claude Code in your vault and type `/boot`.
Do I need Codex?
No. Everything except `agent-sync` works with Claude Code alone. Codex is needed for the shared chat of two agents and `ecx duo` mode.
What does the install do to my config?
It merges rather than overwrites: it skips what exists and saves a `.ecx-bak` copy before editing. `./install.sh --dry-run` lists all files in advance and changes nothing.
What does the danger guard block?
It blocks `rm -rf ~`, disk formats and fork bombs and asks before `git push --force`, `reset --hard` and `DROP TABLE`. The author tested it against the real Claude Code hook payload.
How does ecx duo work?
`ecx duo "task"` starts both agents on a new branch; they write plans blind first and then critique each other. You approve the plan and dangerous actions wait for your confirmation on a card.
Discussion
No questions or reviews yet.
Related
PluginsA plugin that syncs Claude Code, Codex CLI and Antigravity on one project and lets them hand work to each other.
One config for Claude Code and Codex in every repo: shared rules, skill packs, second-model review and a destructive-command guard
An engineering system for coding agents: 71 agents, 311 skills, hooks, memory, rules and the AgentShield security scanner
Shared memory for coding agents: messages in SQLite, old context compacted into a summary tree and MCP tools to search it.
Axel runs named agents on Claude Code, Codex and more in the Rex terminal, and task threads outlive a closed laptop
Shared memory for Claude Code, Codex and Cursor: a decision made in one session is recalled in another tool or on another machine.
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| ECX (this card) | no | 89 | |
| obsidian-second-brain | no | 5,313 | agent memory in an Obsidian vault for one agent, with no message exchange between two |
| Claude Code Memory Setup: Obsidian-based memory | no | 1,065 | a guide with ready files for keeping decisions between sessions |
README in short
The page opens with a composition table: 19 skills, 21 hooks, three search engines, two agents with shared memory, phone access and more than 16 health checks. It then compares with ordinary setups, shows an architecture diagram and walks through seven sections: memory, agent sync, search, phone access, self-healing hooks, duo mode and weekly skills. An FAQ on data, config and removal closes it.