Agent skill
gemini-genai
Google python-genai SDK for Gemini 3 Flash, Gemini 3 Pro, and Gemini models. Use when building with Google's Gemini API, google-genai, implementing thinking/reasoning, structured outputs, function calling, image generation, or multimodal. Triggers on "gemini", "google ai", "genai".
Install this agent skill to your Project
npx add-skill https://github.com/cuba6112/skillfactory/tree/main/.claude/skills/gemini-genai
SKILL.md
Google Gemini python-genai SDK
Model IDs
| Model | ID |
|---|---|
| Gemini 3 Flash | gemini-3-flash-preview |
| Gemini 3 Pro | gemini-3-pro-preview |
| Gemini 3 Pro Image | gemini-3-pro-image-preview |
| Gemini 2.5 Flash | gemini-2.5-flash |
| Gemini 2.5 Pro | gemini-2.5-pro |
Gemini 3 Flash Capabilities
Token Limits: 1,048,576 input / 65,536 output (confirmed 1M context)
Inputs: Text, Image, Video, Audio, PDF
Supported: Batch API, Caching, Code execution, File search, Function calling, Search grounding, Structured outputs, Thinking, URL context
Not Supported: Audio generation, Image generation (use gemini-3-pro-image-preview), Live API, Grounding with Google Maps
Knowledge cutoff: January 2025
Critical: Temperature for Gemini 3
Keep temperature=1.0 - Lower values cause response looping.
Thinking Levels (Gemini 3)
config=types.GenerateContentConfig(
thinking_config=types.ThinkingConfig(thinking_level="high") # minimal|low|medium|high
)
minimal: No thinking, lowest latency (Flash only)high: Maximum reasoning (default)
Async Pattern
async with genai.Client().aio as client:
response = await client.models.generate_content(...)
# Async chat
chat = client.aio.chats.create(model="gemini-3-flash-preview")
Media Resolution (v1alpha required)
client = genai.Client(http_options=types.HttpOptions(api_version='v1alpha'))
types.Part(
inline_data=types.Blob(mime_type="image/jpeg", data=image_bytes),
media_resolution={"level": "media_resolution_high"} # low|medium|high|ultra_high
)
Tokens: low=280, medium=560, high=1120
Built-in Tools (Gemini 3)
config=types.GenerateContentConfig(
tools=[{"google_search": {}}], # Web search
# tools=[{"url_context": {}}], # Fetch URL content
# tools=[{"code_execution": {}}], # Run code
)
Enum Response
from enum import Enum
class Category(Enum):
A = "A"
B = "B"
config={
"response_mime_type": "text/x.enum",
"response_schema": Category,
}
Image Generation
response = client.models.generate_content(
model="gemini-3-pro-image-preview",
contents="A cyberpunk city",
config=types.GenerateContentConfig(
response_modalities=["IMAGE"],
image_config=types.ImageConfig(aspect_ratio="16:9", image_size="4K"),
),
)
image = response.parts[0].as_image()
image.save("out.png")
Grounded (with search): Add tools=[{"google_search": {}}]
Response Helpers
response.text # Text content
response.parsed # Auto-parsed JSON (when using response_json_schema)
response.function_calls # List of function calls (cleaner than candidates drilling)
Pricing (Gemini 3 Flash)
Input: $0.50/1M | Output: $3.00/1M | Context caching: up to 90% reduction
Resources
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
prompt-engineering
Comprehensive prompt engineering techniques for Claude models. Use this skill when crafting, optimizing, or debugging prompts for Claude API, Claude Code, or any Claude-powered application. Covers system prompts, role prompting, multishot examples, chain of thought, XML structuring, long context handling, extended thinking, prompt chaining, Claude 4.x-specific best practices, and agentic orchestration including subagents, agent loops, skills, MCP integration, and multi-agent workflows.
adk-rag-agent
Build RAG (Retrieval-Augmented Generation) agents with Google ADK and Vertex AI RAG Engine. Use when implementing document Q&A, knowledge base search, or citation-backed responses. Covers VertexAiRagRetrieval tool, corpus setup, and citation formatting.
headless-cli-agents
Build agentic systems using Claude CLI in headless mode or the Claude Agent SDK. Use when building automation pipelines, CI/CD integrations, multi-agent orchestration, or programmatic Claude interactions. Covers CLI flags (-p, --output-format), session management (--resume, --continue), Python SDK (claude-agent-sdk), custom tools, and agent loop patterns.
notion-knowledge-capture
Capture conversations and decisions into structured Notion pages; use when turning chats/notes into wiki entries, how-tos, decisions, or FAQs with proper linking.
mcp-builder
Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).
gh-fix-ci
Inspect GitHub PR checks with gh, pull failing GitHub Actions logs, summarize failure context, then create a fix plan and implement after user approval. Use when a user asks to debug or fix failing PR CI/CD checks on GitHub Actions and wants a plan + code changes; for external checks (e.g., Buildkite), only report the details URL and mark them out of scope.
Didn't find tool you were looking for?