Metadata-Version: 2.4
Name: au-selection-analytics-mcp
Version: 2.0.0
Summary: MCP server for AU Selection Analytics Portal — query CRB parity, GCOR brands, LT Review, GMS models from Kiro AI chat
Author-email: AU Retail BI <cbalakri@amazon.com>
License-Expression: MIT
Requires-Python: >=3.10
Requires-Dist: requests>=2.28
Description-Content-Type: text/markdown

# AU Selection Analytics MCP Server v2.0

Read-only access to AU Selection Analytics portal data from Kiro AI chat.

## Install (PyPI)

```json
{
  "mcpServers": {
    "au-selection-analytics": {
      "command": "uvx",
      "args": ["au-selection-analytics-mcp@latest"]
    }
  }
}
```

Add the above to your `~/.kiro/settings/mcp.json`. That's it.

## Prerequisites

- Amazon VPN connected
- `mwinit` run (valid Midway session)
- No AWS credentials needed

## Available Tools (10 — all read-only)

| Tool | What it does |
|------|-------------|
| `get_parity_data` | GL/Channel/PF parity metrics (gaps, CTC, parity %) |
| `get_gcor_brands` | Brand-level GCOR data per PF tab (gaps, status, actionability) |
| `search_brand` | Search a brand across all 4 PFs |
| `get_top10` | Top 10 brand opportunities |
| `get_lt_review` | LT Review tables — Top 10/20 brands per PF (Net New, Existing, Holdout) |
| `get_strategic_brands` | VM Bottom-Up/Input strategic brands per PF |
| `get_gms_models` | GMS Opportunity Model data (predicted opportunity, GVS) |
| `get_feedback_list` | Recent feedback items |
| `get_weeks` | Available data weeks |
| `ask_analytics` | Natural language → SQL → results (AI-powered, SELECT only) |

## Example Prompts in Kiro

```
"Show me OHL LT Review — top brands by gaps"
"What's the status of Omron across all PFs?"
"Get Top 10 brands with highest CRBHS gaps"
"Which Cons brands have Net New actionability with >50 gaps?"
"What weeks of data are available?"
"Show me GMS opportunity for Beauty brands"
```

## Security

- **Read-only** — no write/edit/delete operations exposed
- **Midway auth** — same as opening the portal in browser
- **No AWS access** — MCP calls the portal REST API, not AWS directly

## Publishing (maintainer)

```bash
cd mcp/
pip install build twine
python -m build
twine upload dist/*
```

## Version History

| Version | Date | Changes |
|---------|------|---------|
| 2.0.0 | 2026-06-29 | Added LT Review, GMS Models, Strategic Brands, Feedback, Weeks. All read-only. PyPI packaging. |
| 1.0.0 | 2026-06-22 | Initial: parity, GCOR, search, top10, analytics |
