What is claudish CLI in the context of this skill?

Claudish is a command-line interface that allows Claude Code to communicate directly with external AI providers like Gemini, OpenAI, and MiniMax through Bash commands.

Can I save my preferred models for specific tasks?

Yes, model preferences are learned per context (e.g., debugging or research) and saved in .claude/multimodel-team.json for automatic reuse.

How does the error escalation protocol work?

If an external model fails, the skill stops and reports the exact stderr error to the user. It never silently substitutes a different model to ensure output integrity.

Does this skill require manual provider prefixes for models?

No, since version 5.4.0, claudish handles routing internally. You can pass bare model names like 'grok-code-fast-1' and the system auto-resolves the provider.

Multi-Model Proxy Integration

Name: Multi-Model Proxy Integration
Author: MadAppGang

byMadAppGang

•

数据科学与机器学习

Integrates and manages external AI models within Claude Code using deterministic CLI routing for multi-model workflows.

This skill provides a comprehensive framework for invoking external LLMs like Gemini, OpenAI, and specialized regional models via the claudish CLI. It ensures reliable model routing, manages persistent context-based model preferences, and enforces a strict error escalation protocol. By utilizing direct CLI calls rather than prompt-based delegation, it enables deterministic and parallel model execution for complex tasks such as cross-model code reviews and deep research, ensuring Claude stays synchronized with the best available external intelligence.

主要功能

01Deterministic external model invocation via claudish CLI

02Persistent context-based model preferences and automatic reuse

03Strict error escalation protocol to prevent silent model substitution

04Automatic model routing without required provider prefixes

05Parallel model execution for high-speed multi-agent workflows

063 GitHub stars

使用场景

01Debugging integration issues using specialized external models via CLI

02Conducting multi-model code reviews to gather diverse LLM perspectives

03Automating research tasks requiring specific non-native model capabilities

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add madappgang/magus proxy-mode-reference

For use in Claude.ai and ChatGPT

主要功能

01Deterministic external model invocation via claudish CLI

02Persistent context-based model preferences and automatic reuse

03Strict error escalation protocol to prevent silent model substitution

04Automatic model routing without required provider prefixes

05Parallel model execution for high-speed multi-agent workflows

063 GitHub stars

使用场景

01Debugging integration issues using specialized external models via CLI

02Conducting multi-model code reviews to gather diverse LLM perspectives

03Automating research tasks requiring specific non-native model capabilities

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add madappgang/magus proxy-mode-reference

For use in Claude.ai and ChatGPT