gemini-context-bridge

# Gemini Context Bridge

## 🚨 MANDATORY: Invoke gemini-cli-docs First

> **STOP - Before providing ANY response about GEMINI.md syntax:**
>
> 1. **INVOKE** `gemini-cli-docs` skill
> 2. **QUERY** for the specific context topic
> 3. **BASE** all responses EXCLUSIVELY on official documentation loaded

## Overview

This skill bridges the gap between **Claude Code** and **Gemini CLI** by providing workflows to synchronize project context and strategic guidance on agent delegation. It ensures that project conventions defined in `CLAUDE.md` are understood by Gemini (via `GEMINI.md`) and helps users leverage the unique strengths of each agent.

## When to Use This Skill

**Keywords:** sync context, share memory, copy instructions, migrate context, bridge agents, gemini.md, claude.md, delegation strategy, agent selection

**Use this skill when:**

- **Onboarding Gemini:** You have a `CLAUDE.md` and want to initialize a `GEMINI.md` for Gemini CLI.
- **Syncing Instructions:** You want to ensure both agents follow the same coding styles and conventions.
- **Deciding Agent:** You are unsure whether to use Claude or Gemini for a specific task (e.g., "Who should run this migration?").
- **Comparing Context:** You want to see if there are conflicting instructions between the two agents.

## Delegation Strategy (Gemini vs Claude)

When orchestrating tasks, use this guide to choose the right agent:

### Delegate to **Gemini CLI** for

1. **Google Ecosystem Auth**: Tasks requiring authenticated access to GCP, Vertex AI, or Firebase.
2. **Long Context / Low Cost**: Operations best suited for Gemini Flash (large context) or Pro (very large context).
3. **Interactive Tools**: Running interactive CLIs (e.g., `top`, `vim`, `git rebase`) via `enableInteractiveShell`.
4. **Smart Edits**: Using the `edit` tool with the specific "Smart Edit" strategies of Gemini.
5. **Memory/Todos**: Utilizing the built-in programmable `save_memory` and `write_todos` tools for state tracking.

### Delegate to