Metadata-Version: 2.4
Name: gc-feedback-mcp
Version: 1.0.0
Summary: MCP server for BizChat Gov Cloud quality-through-feedback data
License-Expression: MIT
Requires-Python: >=3.10
Requires-Dist: azure-identity>=1.15.0
Requires-Dist: azure-keyvault-secrets>=4.7.0
Requires-Dist: azure-storage-blob>=12.19.0
Requires-Dist: httpx>=0.25.0
Requires-Dist: mcp>=1.0.0
Description-Content-Type: text/markdown

# GC-Feedback MCP Server

Manifest-driven MCP server for BizChat Gov Cloud **quality-through-feedback** data.

Same architecture as `gc-eval-mcp` — reads a `manifest.json` from Azure Blob Storage at startup, dynamically registers tools per dataset type (HTML reports, Markdown docs, JSON, SQLite).

## Architecture

```
┌──────────────┐    ┌──────────────────┐    ┌────────────────────┐
│ Copilot CLI  │───▶│ gc-feedback-mcp  │───▶│ Azure Blob Storage │
│ (MCP client) │    │  (this server)   │    │ (feedback-data)    │
└──────────────┘    └──────────────────┘    └────────────────────┘
                           │                         ▲
                           ▼                         │
                    ┌──────────────┐          ┌─────────────┐
                    │ Key Vault    │──────────▶│ Storage Key │
                    │ (az login)   │          └─────────────┘
                    └──────────────┘
```

## Adding Data

1. Add files to the `feedback-data` blob container
2. Update `manifest.json` to describe the new dataset
3. Upload the updated manifest

No code changes needed — the server picks up new datasets automatically.

## Local Development

```bash
cd quality-through-feedback/mcp-server
pip install -e .
gc-feedback-mcp
```

## Publishing

Tagged releases (e.g., `feedback-v1.0.0`) auto-publish to PyPI via GitHub Actions.
