# Sub-Agents Skills

[](https://developers.openai.com/codex/cli) [](https://claude.ai/code) [](https://www.kimi.com/code/en) [](https://agentskills.io) [](https://opensource.org/licenses/MIT)

Run task-specific agents on different AI coding backends from one parent tool.

Use Codex, Claude Code, Cursor CLI, GLM, Kimi, Grok Build, Gemini CLI, and OpenCode as sub-agents in one workflow. Agent definitions are Markdown files, and each agent can select its execution backend.

```mermaid graph LR A["Your AI tool<br/>(Codex, Claude Code, Cursor...)"] --> B["run_subagent.py"] B --> C["Codex"] B --> D["Claude Code"] B --> E["Cursor CLI"] B --> H["Grok Build"] B --> G["GLM"] B --> K["Kimi"] B --> F["Gemini CLI"] B --> I["OpenCode"] I --> J["Configured provider/model<br/>(API · gateway · local)"] style B fill:#f5f5f5,stroke:#333 ```

## Why?

Most AI coding tools provide sub-agents tied to their own models. Claude Code delegates to Claude, and Codex delegates to GPT. Their built-in delegation does not provide a portable way to route a task to another provider's model.

This skill lets each agent select a supported backend:

- **Backend selection per task:** Choose Codex for a quick edit, Claude Code for a deeper pass, or another supported backend for a separate implementation pass. - **Portable definitions:** Plain Markdown agent files work with Codex, Claude Code, Cursor CLI, GLM, Kimi, Grok Build, Gemini CLI, VS Code, and [30+ other tools](https://agentskills.io) that support the Agent Skills format. - **Direct provider billing:** Choose which model handles each task and pay the provider at