gemini-flow
Google Gemini API integration and model orchestration
When & Why to Use This Skill
Gemini Flow is a powerful Claude skill designed for seamless Google Gemini API integration and advanced model orchestration. It enables developers to leverage multimodal capabilities—including text generation, multi-turn conversations, and sophisticated image analysis—while achieving over 90% context savings through efficient model routing and architectural optimization.
Use Cases
- Multimodal Data Processing: Utilize the analyze-image and describe-image tools to extract insights from visual content and integrate them into automated text workflows.
- Cost-Efficient LLM Orchestration: Implement intelligent routing between different Gemini models (like Pro and Vision) to optimize performance and reduce API overhead for complex agentic tasks.
- Advanced RAG and Search: Use the embedding tool to create high-dimensional vector representations of data, enhancing the accuracy of knowledge retrieval and semantic search within AI applications.
- Automated Model Management: Dynamically list and query model capabilities using list-models and model-info to ensure the agent always uses the most appropriate version for the specific task.
| name | gemini-flow |
|---|---|
| description | Google Gemini API integration and model orchestration |
| version | 1.0.0 |
| allowed-tools | [Bash, Read, WebFetch] |
Gemini Flow Skill
Overview
Google Gemini API integration. 90%+ context savings.
Requirements
- GOOGLE_API_KEY environment variable
- gemini-pro or gemini-pro-vision model access
Tools (Progressive Disclosure)
Generation
| Tool | Description |
|---|---|
| generate | Generate text content |
| chat | Multi-turn conversation |
| embed | Generate embeddings |
Vision
| Tool | Description |
|---|---|
| analyze-image | Analyze image content |
| describe-image | Describe image |
Models
| Tool | Description |
|---|---|
| list-models | List available models |
| model-info | Get model capabilities |
Agent Integration
- llm-architect (primary): Model orchestration
- model-orchestrator (primary): Multi-model routing
- developer (secondary): API integration