DataDoe MCP + Gemini CLI Template

This repository is a minimal example of using the DataDoe MCP server from Gemini CLI for Amazon-focused workflows.

Table of Contents

• What you can do with this repo
• What This Repo Includes
• Prerequisites
• How to get a DataDoe subscription and get MCP Key
• Configure DataDoe MCP in Gemini CLI
• Example prompt library starter pack
• Run Gemini CLI from Dedicated Launcher
• Gemini Settings (Official Model)
• DataDoe MCP Configuration Options
• How to get help

What you can do with this repo

• Connect Gemini CLI to DataDoe MCP in a secure way.
• Ask Amazon seller questions using DataDoe-backed data.
• Reuse this setup as a template for new Amazon-focused assistant projects.

What This Repo Includes

• Gemini CLI MCP setup guidance for a project-scoped datadoe server
• secure secret handling with .env and .env.example
• repository-specific assistant rules in GEMINI.md

Prerequisites

• Gemini CLI installed and working (gemini --version)
• A valid DataDoe subscription
• A generated DataDoe MCP key

If gemini --version shows nothing or gemini is not found, install and verify Gemini CLI:

npm install -g @google/gemini-cli
hash -r
gemini --version

If the command is still not found, add npm global binaries to your shell PATH (for zsh):

echo 'export PATH="$(npm config get prefix)/bin:$PATH"' >> ~/.zshrc
source ~/.zshrc
which gemini
gemini --version

If your organization uses Google Workspace subscription access for Gemini CLI, authenticate with your Workspace account:

1. Run gemini
2. Choose Sign in with Google
3. Complete browser sign-in with your Workspace account
4. If needed later, run /auth inside Gemini CLI to switch authentication method

How to get a DataDoe subscription and get MCP Key

1. Go to app.datadoe.com.
2. Create an account.
3. Purchase a subscription.
4. Accept the Terms and Conditions and Privacy Policy.
5. Go to the Integrations module.
6. Click the MCP tile (this navigates to /integrations/mcp).
7. Click MCP Key, then add a name and expiration date, and click Create.
8. Copy the key and store it in a secure secret manager or another safe location.

Configure DataDoe MCP in Gemini CLI

Warning

Direct gemini mcp add ... connection for DataDoe is currently unreliable in our setup. Use the launcher + mcp-remote workaround from this repository.

Direct command shown below is kept for reference only:

gemini mcp add --transport http --scope project --header "datadoe-mcp-key: YOUR_API_KEY" datadoe "https://api.datadoe.com/mcp/v1"

What this does (reference behavior):

• adds a project-scoped MCP server named datadoe
• creates/updates .gemini/settings.json for shared project configuration
• keeps setup consistent for all collaborators

Reference: Gemini CLI MCP docs and mcp-remote usage docs

Example prompt library starter pack

To help you start faster with an AI Agent + DataDoe MCP workflow, this repo includes a small prompt library at:

• .gemini/prompts/EXAMPLES.md

Use it as a starter pack:

1. Open .gemini/prompts/EXAMPLES.md.
2. Copy a prompt block and adjust placeholders (for example {{seller_name}}) to your account context.
3. Run the prompt in Gemini CLI shell chat with DataDoe MCP enabled.
4. Save your own high-performing prompts in the same file to build a reusable internal playbook.

Run Gemini CLI from Dedicated Launcher

This repository includes a dedicated launcher script:

Warning

scripts/start-gemini.sh is the protected launcher for this repository. Do not edit, replace, or "quick fix" it unless you are intentionally changing launcher behavior.

./scripts/start-gemini.sh

The launcher loads .env, exports DATADOE_MCP_KEY into the current process, syncs .gemini/settings.json to the repository mcp-remote configuration, and then starts Gemini CLI from the repository root.

Short manual:

# Interactive menu (recommended)
./scripts/start-gemini.sh

# Direct launch Gemini CLI
./scripts/start-gemini.sh --cli

# Validate env loading + Gemini CLI availability only
./scripts/start-gemini.sh --check

# Help
./scripts/start-gemini.sh --help

If needed, make it executable once:

chmod +x ./scripts/start-gemini.sh

Gemini Settings (Official Model)

Gemini CLI supports user and project settings scopes:

• User scope: ~/.gemini/settings.json (personal defaults across repositories)
• Project scope: .gemini/settings.json (shared with this repository team)
• Project memory/instructions: GEMINI.md (team-shared assistant guidance)

Recommended workflow:

1. Use /settings inside Gemini CLI to create or edit settings safely.
2. Keep team-wide MCP setup in .gemini/settings.json.
3. Use /model to inspect or set the active model for your session.
4. Use /mcp to verify MCP server status and discovered tools.

DataDoe MCP Configuration Options

You can configure DataDoe MCP in either of these ways.

Option A: repository-managed mcp-remote proxy config (recommended and currently supported):

{
  "mcpServers": {
    "datadoe": {
      "command": "bash",
      "args": [
        "-lc",
        "npx -y mcp-remote@latest https://api.datadoe.com/mcp/v1 --transport http-only --header \"datadoe-mcp-key:${DATADOE_MCP_KEY}\""
      ],
      "env": {
        "DATADOE_MCP_KEY": "$DATADOE_MCP_KEY"
      }
    }
  }
}

Option B: direct Gemini MCP HTTP setup (currently unstable in this environment; fallback only):

gemini mcp add --transport http --scope project --header "datadoe-mcp-key: YOUR_API_KEY" datadoe "https://api.datadoe.com/mcp/v1"

The recommended path is to always start with ./scripts/start-gemini.sh, which loads .env and re-syncs mcp-remote configuration automatically.

Caution

Treat DATADOE_MCP_KEY like a password. Do not publish repositories, screenshots, or logs that contain this key. Never commit real keys to git. If a key is exposed, rotate it immediately.

How to get help

• Email: contact@datadoe.com
• Gemini CLI docs: geminicli.com/docs
• Gemini CLI commands: geminicli.com/docs/reference/commands