OpenWolf CLI
openwolf
A shared local project memory for Claude Code, Codex and OpenCode: task context, a project map, token accounting and a localhost dashboard
- 2,383 ★ on the repository
- Last commit
- AGPL-3.0
Broad access
Why this level
- Installs hooks into agents and adds instructions to their context
- Writes to the project's working folder
- The AGPL-3.0 license has copyleft conditions
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/openwolfDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Installs hooks into agents and adds instructions to their context
- Writes to the project's working folder
- The AGPL-3.0 license has copyleft conditions
This is third-party code. Review the repository files before installing.
What it does
OpenWolf keeps a shared local folder in the project with task context, a project map and known fixes so agents continue saved work. It accounts for tokens by agent and model, works locally without extra model calls and provides a dashboard on `127.0.0.1` with a project token. Install with `npm install -g openwolf`, then in the project run `openwolf init`: the command finds installed agents and connects them, and you can set the list with the `--agent claude codex opencode` flag. After that the agents ask you to confirm trust in the hooks. Integrations: Claude Code (lifecycle hooks, read hints, Bash output control, skills), Codex CLI (hooks), OpenCode (a plugin), compatible Grok Build hooks, and for Cursor, Gemini CLI and Antigravity only the context is available.
Who it is for. For people who work on one project from several agents and want them to share context and the project map and not lose found fixes.
Good fit when
- You want Claude Code, Codex and OpenCode to continue each other's work from a shared context
- You need token accounting by agent and model on a local dashboard
- You need a project map and a list of known fixes in one folder
Not a fit when
- You need a license free of copyleft: OpenWolf is under AGPL-3.0
- You are not ready for hooks in agents that read and add instructions
Example request
Connect OpenWolf to the project, give Claude Code and Codex one project map and show the token spending of eachLimitations
OpenWolf installs hooks into agents and writes to the project's working folder, so it has wide access. The author himself notes that long-term instruction injection needs independent review by an administrator: read what lands in the context. The AGPL-3.0 license has copyleft conditions. For Cursor, Gemini CLI and Antigravity only the context is available with no hooks. The dashboard listens only on `127.0.0.1` and is protected by a project token. Version 2.5.2 from 15.09.2026.
How to disable. Remove the OpenWolf hooks from the agents' settings, run npm uninstall -g openwolf and delete the OpenWolf working folder from the project.
FAQ
What does OpenWolf do?
It keeps a shared local project context folder for several agents: tasks, a project map, known fixes and token accounting.
Which agents are supported?
Claude Code, Codex CLI and OpenCode with hooks or a plugin, compatible Grok Build hooks, and for Cursor, Gemini CLI and Antigravity only the context.
How do I install it?
npm install -g openwolf, then openwolf init in the project.
What is the license?
AGPL-3.0: copyleft terms must be considered when distributing changes.
Discussion
No questions or reviews yet.
Related
CLI toolsA local context store for coding agents that captures project knowledge and reuses it across Codex, Claude Code and OpenCode
Orchestrator plugin for OpenCode and Codex CLI: multiple agents, hooks, built-in MCPs, ultrawork mode and hash-anchored edits
A context database for agents: cross-session memory, documents and skills in one viking:// file tree, reachable through MCP and hooks
A self-hosted control center for coding agents: OpenHands, Claude Code, Codex and Gemini on local and remote backends
A local dashboard for many Claude Code sessions: each job gets its own context and todo, agents spawn child agents and message each other
CCS
CCS - Claude Codex Switch
A CLI that switches Claude Code between accounts, providers and local models without editing configs
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| OpenWolf (this entry) | no | 2,383 | ||
| agent-memory | no | 3,403 | long-term memory on Markdown files with a SQLite index | |
| MCP Knowledge Graph | no | not required | 888 | memory in a local knowledge graph, .aim files |
| MemoraX Code | no | 2,182 | a memory plugin for architecture and repository rules, version 0.1.x |
README in short
The README describes a shared local project memory for coding agents: context, a project map, token accounting, a localhost dashboard and integrations with Claude Code, Codex, OpenCode and others. Installation through npm. AGPL-3.0 license.