Google Gemini Chat — SKILL.md

---
name: "Google Gemini Chat"
description: "Use Google Gemini via browser automation — chat, create images, music, video, use Gems, upload files. For browser automation of gemini.google.com."
version: "1.0.0"
author: "skynet"
category: "dev"
agents: ["claude-code"]
tags: ["gemini", "google", "ai", "browser-automation", "chat"]
tools_required: ["chrome-mcp"]
---

# Google Gemini Chat

# Google Gemini Chat

Browser automation skill for gemini.google.com.

## Navigation

Navigate to `gemini.google.com/app` to access the Gemini chat interface.

## Core Interface Elements

### Prompt Input
- The main prompt textbox has placeholder text "Enter a prompt for Gemini"
- This is a real textbox — `form_input` works directly
- Type your prompt and click the **Send message** button to submit

### Mode Picker
- Located near the prompt input
- Switch between different Gemini modes/models

### File Upload
- Upload files directly to the conversation
- Use the upload button near the prompt input

### Quick Tools
Gemini offers quick action buttons:
- **Create image** — generate images from text descriptions
- **Create music** — generate music tracks
- **Create video** — generate video content
- **Write anything** — general writing assistance

## Sidebar Navigation

The sidebar contains:
- **New chat** — start a fresh conversation
- **My stuff** — access saved content and history
- **Gems** — access custom Gems (specialized AI personas/tools)
- **Chat history** — browse previous conversations
- **Settings** — configure Gemini preferences

## Tips
- Gemini responses may take time for media generation (images, music, video)
- Use Gems for specialized tasks — they provide pre-configured AI behavior
- The sidebar can be toggled open/closed

## No Project System Warning

Gemini has NO project system or persistent memory. Agents must include full context in every prompt:
- Who the user is and their account details
- What they are working on and why
- The complete goal and any relevant history
- All necessary URLs, credentials references, and prior results

Never assume the agent remembers anything from previous interactions.

## Running on Mac Minis

These browser automation tasks run on Mac Minis via mac-control on port 8200:
- **Vault**: 192.168.86.27:8200
- **Bots**: 192.168.86.50:8200
- **Jarvis**: 192.168.86.51:8200

Use the mac-control API to send browser automation commands. Example:
```
POST http://192.168.86.27:8200/execute
```