🤖 Skill: gemini-sdk-expert (v1.3.0)

Executive Summary

gemini-sdk-expert

@google/genai

📋 Table of Contents

1. Core Capabilities
2. The "Do Not" List (Anti-Patterns)
3. Quick Start: JSON Enforcement
4. Standard Production Patterns
5. Advanced Agentic Patterns
6. Context Caching Strategy
7. Multimodal Integration
8. Safety & Responsible AI
9. Reference Library

🚀 Core Capabilities

• Strict Structured Output: Leveraging

  responseSchema

  for 100% reliable JSON generation.
• Agentic Function Calling: enabling models to interact with private APIs and tools.
• Long-Form Context Management: Using Context Caching for massive datasets (2M+ tokens).
• Native Multimodal Reasoning: Processing video, audio, and documents as first-class inputs.
• Latency Optimization: Strategic model selection (Flash vs. Pro) and streaming responses.

🚫 The "Do Not" List (Anti-Patterns)

responseSchema

@google/generative-ai

@google/genai

GOOGLE_GENAI_API_VERSION

⚡ Quick Start: JSON Enforcement

import { GoogleGenerativeAI, Type } from "@google/genai";

// Optional: Set API Version via env
// process.env.GOOGLE_GENAI_API_VERSION = "v1beta1";

const schema = {
  type: Type.OBJECT,
  properties: {
    status: { type: Type.STRING, enum: ["COMPLETE", "PENDING", "ERROR"] },
    summary: { type: Type.STRING },
    priority: { type: Type.NUMBER }
  },
  required: ["status", "summary"]
};

// Always set MIME type to application/json
const result = await model.generateContent({
  contents: [{ role: 'user', parts: [{ text: "Evaluate task X..." }] }],
  generationConfig: {
    responseMimeType: "application/json",
    responseSchema: schema
  }
});

🛠 Standard Production Patterns

Pattern A: The Data Extractor (Flash)

• Model:

  gemini-3-flash

• Config: High

  topP

  , low

  temperature

  for deterministic extraction.

Pattern B: The Complex Reasoner (Pro)

• Model:

  gemini-3-pro

• Config: Enable Strict Mode in schemas for 100% adherence.

🧩 Advanced Agentic Patterns

Parallel Function Calling

Semantic Caching

💾 Context Caching Strategy

• Warm-up Phase: Initial context upload.
• Persistence Phase: Referencing the cache via

  cachedContent

  .
• Cleanup Phase: Managing TTLs to optimize storage costs.

📸 Multimodal Integration

• Video: Scene detection and temporal reasoning.
• Audio: Sentiment, tone, and environment detection.
• Document: Visual layout and OCR.

📖 Reference Library

• Structured Output: Nested schemas and validation.
• Function Calling: Tools, execution loops, and security.
• Context Caching: Reducing cost and latency.
• Multimodal 2026: Video, audio, and PDF mastery.