OpenDomain CLI
OpenDomain keeps a project's business concepts and rules in repository Markdown so agents rely on them instead of guessing.
- 297 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- The `init` command edits `AGENTS.md` and adds skills under `.codex/skills/`.
- The agent reads project materials and conclusions may land as flawed candidates.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/opendomainDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The `init` command edits `AGENTS.md` and adds skills under `.codex/skills/`.
- The agent reads project materials and conclusions may land as flawed candidates.
This is third-party code. Review the repository files before installing.
What it does
OpenDomain is a CLI for Codex and other agents that maintains a domain layer in the repository: concepts, rules, lifecycles, events and evidence. Everything is Markdown and YAML, so it lives next to the code under Git. The `opendomain init --tools codex` command adds an `opendomain/` folder, a managed `AGENTS.md` block and three Codex skills: explore, model and review. Knowledge an agent inferred goes into candidates and becomes accepted only after a human decision. Before changing code the agent checks accepted rules through `opendomain assure`.
Who it is for. Teams with long-lived products where an agent edits business logic and must know what an order cancellation or a payment status really means.
Good fit when
- An agent is about to change order cancellation logic and needs the rules first.
- Business rules are scattered across tickets and chats and should become reviewable.
- An agent's conclusions about the domain need human review before acceptance.
- A new hire or a new agent session starts by reading the accepted model.
Not a fit when
- The project is small and every rule fits in one README.
- You need a delivery plan for changes, which OpenSpec and Spec Kit cover.
- The team will not review candidates and the model would go stale without an owner.
Example request
Organize the order cancellation rules from the project materials, put conclusions into candidates and show the evidence for review.Limitations
The integration is described for Codex: `opendomain init --tools codex` generates skills under `.codex/skills/`. Other agents need manual setup. Node.js 20 or newer is required. The last repository commit is from September 2026. Model quality depends on people: until candidates are reviewed, the agent relies on incomplete knowledge. The `init` command edits `AGENTS.md`, so review the diff before committing. Inferred domain knowledge can be wrong, which is why a human accepts it.
How to disable. Delete the `opendomain/` folder, the managed block in `AGENTS.md` and the `opendomain-*` skills under `.codex/skills/`. Remove the package with `npm uninstall --global @echopath-labs/opendomain`.
FAQ
What does `opendomain init` create?
An `opendomain/` folder with config and subfolders, a managed `AGENTS.md` block and three skills under `.codex/skills/`. The project `package.json` is left alone.
Does the agent accept knowledge by itself?
No. Agent conclusions stay candidates and only a human can accept them after reviewing evidence and conflicts.
How do I update a workspace?
Upgrade the package and run `opendomain update --json`, then `opendomain doctor --json` and `opendomain validate --json`. No model migration is needed.
Does it replace OpenSpec?
No. OpenSpec describes a change while OpenDomain stores lasting domain knowledge. An agent can pass a request that references accepted IDs.
Which Node.js version is required?
Node.js 20 or newer. A standalone binary exists as an alternative to npm.
Discussion
No questions or reviews yet.
Related
CLI toolsA layer on top of OpenAI Codex that adds project memory, planning and completion checks before a task counts as done
A code graph and guardrails for agents: one call shows what a task touches, and a check stops an unsafe change
A kit for porting a Claude Code workflow to OpenAI Codex CLI for multi-tenant PHP/Bitrix24-Marketplace projects
Git-native project context for agents: specs, decisions and rules live in Markdown beside the code and are read before edits
A CLI that gives six coding agents project context, over 80 skills, verification and resumable tasks across sessions
Tasks, docs, specs and decisions live as markdown files in your repository, and the agent reads them through the CLI and MCP
README in short
The docs cover the Markdown knowledge layer, boundaries with OpenSpec and EchoPath, installation through Codex or npm and the 0.1.0 upgrade. Apache-2.0, stable status.