gemini-skill

Automate Gemini web actions via CDP and MCP server.

827|120|Updated Mar 12, 2026
One-click install
npx skills add https://github.com/WJZ-P/gemini-skill --skill gemini-skill-wjz-p
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gemini-skill
Source: https://github.com/WJZ-P/gemini-skill/tree/main
Command: npx skills add https://github.com/WJZ-P/gemini-skill --skill gemini-skill-wjz-p

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Manual interaction with the Gemini web UI is slow and hard to automate, especially for repeatable tasks like image generation, chat workflows, and retrieving generated images.

Core Features & Use Cases

  • AI Image Generation: Generate images from prompts, optionally using reference images for image-to-image workflows, and return saved outputs to your local filesystem.
  • Text Conversations: Send multi-turn text prompts and retrieve Gemini’s latest or all text responses.
  • Image Retrieval & Delivery: Extract generated images (base64 or full-size CDP download) and automatically remove Gemini watermarks where supported.
  • Session & Model Management: Create new/temp chats, switch models, and adjust thinking depth through provided MCP tools.
  • On-demand Browser Daemon: Automatically start and reuse a managed browser session via a lifecycle TTL so you don’t have to handle setup each run.

Quick Start

Start using the MCP tool by asking: "Generate an image of a cute cat in a specific style and return the saved file path."

Frequently Asked Questions about gemini-skill

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate Gemini image generation and return saved files locally?

Automate Gemini image generation by controlling the Gemini website through a managed browser daemon using CDP and an MCP server. It processes your prompts, extracts generated images, removes watermarks, and saves the output files directly to your local filesystem.

Can I use reference images for image-to-image workflows with Gemini automation?

Gemini automation supports reference image uploads for image-to-image workflows. The MCP server controls the browser session to upload your reference images, apply the prompt modifications, and extract the newly generated images via CDP download.

What is the best way to remove watermarks from Gemini generated images?

Removing watermarks from Gemini generated images is handled automatically during the image retrieval process. The Skill extracts generated images using full-size CDP download and automatically removes Gemini watermarks where supported before saving the output.

Does MCP browser automation support multi-turn text conversations with Gemini?

MCP browser automation supports multi-turn text conversations with Gemini. It sends text prompts through the controlled browser session, retrieves Gemini's latest or all text responses, and manages session state including creating new chats and switching models.

Why does my Gemini browser automation session require a lifecycle TTL?

Gemini browser automation requires a lifecycle TTL to automatically start and reuse a managed browser session. This on-demand daemon approach eliminates manual setup each run, maintaining an active CDP connection for synchronous blocking operations with strict timeout handling.

Are there limitations when using CDP for Gemini session and model management?

CDP for Gemini session management requires mandatory MCP-first tool execution and synchronous blocking for long operations with strict timeout handling. It provides protected alternatives for diagnostics and requires user consent before exposing specific CDP connection details.