Axiopera
Lighter, more efficient, easier agent collaboration with minimal setup changes.
Axiopera connects Claude Code, Codex, and Grok Build.
In Claude Code, ask:
Have Codex and Grok review these changes, then fix the issues they find.
Advantages:
- One executable + one skill.
- Uses your existing agent sign-ins and usage allowances.
- Leaves agent configuration unchanged by default.
- About one line for skill discovery; instructions load on demand.
- All three sub-agents support Steer: add instructions and change direction while they run.
- All three lead agents can receive sub-task messages while working and follow up.
Install
Install and sign in to the agent CLIs you want to use.
1. Install the executable
Download and extract your platform's archive from Releases. Put axiopera or axiopera.exe on your PATH.
With Rust: cargo install axiopera --locked.
Windows and Linux tested; macOS awaits validation. Test coverage
2. Install the skill (recommended)
Run the command for your host agent, then start a new session:
| Agent | Command |
|---|---|
| Claude Code | axiopera skill install claude |
| Codex | axiopera skill install codex |
| Grok Build | axiopera skill install grok |
For all supported harnesses: axiopera skill install all.
The skill is bundled with the executable and installs offline. Locations, replacement, and removal.
For Codex TUI, the skill is enough. Axiopera pushes messages into the active turn directly.
Desktop and VS Code normally receive notifications after the current turn ends. The plugin delivers them after the next tool call through a hook.
The plugin includes the skill. Install through the CLI, trust its hooks, then start a new session:
Documentation
Usage and configuration · Permissions and messaging · Development. Currently in Chinese.