# Model Config

> Configure model providers and default model selection in X-OmniClaw. Use when the user asks to switch models, set a default model, or add or edit provider configs (including self-hosted OpenAI-compatible endpoints as a new `models.providers` key). For X-OmniClaw, model settings live in /sdcard/.xomniclaw/xomniclaw.json under agent.defaultModel and models.providers.

- Skill: `oppo-mente-lab/model-config` (Agent Skill)
- Install (CLI): `npx skillmds@latest add oppo-mente-lab/model-config`
- Raw SKILL.md: https://api.skillmd.com/api/skills/oppo-mente-lab/model-config/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: oppo-mente-lab (https://skillmd.com/u/oppo-mente-lab)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/oppo-mente-lab/model-config

---


# Model Config

For X-OmniClaw, model configuration is stored in:

- `/sdcard/.xomniclaw/xomniclaw.json`

Relevant paths:

- `agent.defaultModel`
- `models.providers.<providerName>.baseUrl`
- `models.providers.<providerName>.apiKey`
- `models.providers.<providerName>.api`
- `models.providers.<providerName>.authHeader`
- `models.providers.<providerName>.models`

## Workflow

1. Read `/sdcard/.xomniclaw/xomniclaw.json` first.
2. Check current `agent.defaultModel` and existing providers under `models.providers`.
3. If the target provider already exists, update only the necessary fields.
4. If the target provider does not exist, create a provider entry under `models.providers`.
5. For self-hosted or other OpenAI-compatible endpoints, add a **new** `models.providers.<yourKey>` entry (any stable key name) with:
   - `api: "openai-completions"`
   - provider-specific `baseUrl`
   - `apiKey`
   - one or more model entries in `models`
6. Update `agent.defaultModel` to the requested model.
7. Preserve unrelated providers and settings.
8. If asked to verify, send a simple test prompt after config is saved.

## Notes

- Do not overwrite the full config if only one provider/model needs to change.
- If the user provides a compatible endpoint/key, add it as a new provider block under `models.providers` with `api: "openai-completions"` (or another supported API type per app).
- Keep model IDs exact.