Wasu Model Onboarding Assistant

Problem Addressed

Connecting a third-party LLM endpoint often involves account login, API key retrieval, base URL handling, model capability review, and writing the right settings into different agent frameworks. wasu-models turns that flow into a wasu-cli-driven workflow for engineers who need to configure Wasu LLM endpoints in clients such as OpenCode, Claude Code, WorkBuddy, OpenClaw, Codex, Kilo CLI, or Hermes.

How the Skill Works

• Environment check: verifies that the wasu CLI is available and meets the required version.
• Login and credentials: completes Wasu platform login through a QR code or link, obtains the LLM endpoint, and saves apiKey and baseUrl to wasu-config.json without printing the key in chat.
• Model recommendation: uses the wasu --models result together with reference files to summarize available text-generation and multimodal models, applicable scenarios, and preferred or fallback combinations.
• Agent configuration: detects the target agent, reads the existing config, and merges in the new endpoint across JSON, TOML, and YAML files, then advises restarting the client.

Boundaries and Notes

The skill does not query billing details, configure MCP servers, or handle tool plugins. If the user already has another provider's credentials, or the target agent is unsupported and cannot be customized, use the native provider flow instead. Agent configuration must preserve existing settings, model IDs must keep their original names, and sensitive keys must never appear in logs or conversation history.