Meta Agent Tools
PT

Back to the catalog

memoriant-mcp-data-skill

MCP server for conversational data access in Claude Code. Query CRM systems, ticket trackers, and databases using natural language. Includes

Open source Repository Open in the app JSON README (API)

About

MCP server for conversational data access in Claude Code. Query CRM systems, ticket trackers, and databases using natural language. Includes schema discovery, relationship mapping, and query building. Built on the Model Context Protocol for seamless Claude Code integration.

Details

Kind
Plugins
Topic
AI, RAG & memory
Publisher
nathanmaine
Origin
marketplace
Category
ferramentas
Last push
2026-03-26T05:02:35Z
Repository state
ativo
Language
Python
License
MIT
Added
2026-08-30 01:48:58
Updated
2026-08-30 01:48:58
Origin id
nathanmaine/memoriant-mcp-data-skill/memoriant-mcp-data-skill

README

<p align="center">
  <img src="https://img.shields.io/badge/claude--code-plugin-8A2BE2" alt="Claude Code Plugin" />
  <img src="https://img.shields.io/badge/skills-4-blue" alt="4 Skills" />
  <img src="https://img.shields.io/badge/agents-2-green" alt="2 Agents" />
  <img src="https://img.shields.io/badge/license-MIT-green" alt="MIT License" />
</p>

# Memoriant MCP Data Skill

A Claude Code plugin for conversational data access via the Model Context Protocol (MCP). Ask questions in plain English and get answers from CRM, ticket systems, and databases — all at once, without logging into each system separately.

**No servers. No Docker. Just install and use.**

## Install

```bash
/install NathanMaine/memoriant-mcp-data-skill
```

## Cross-Platform Support

### Claude Code (Primary)
```bash
/install NathanMaine/memoriant-mcp-data-skill
```

### OpenAI Codex CLI
```bash
git clone https://github.com/NathanMaine/memoriant-mcp-data-skill.git ~/.codex/skills/mcp-data
codex --enable skills
```

### Gemini CLI
```bash
gemini extensions install https://github.com/NathanMaine/memoriant-mcp-data-skill.git --consent
```

## Skills

| Skill | Command | What It Does |
|-------|---------|-------------|
| **Query Data** | `/query-data` | Natural language → parallel MCP queries → unified answer |
| **Discover Schema** | `/discover-schema` | Map all connected data sources: fields, types, relationships |
| **Explore CRM** | `/explore-crm` | Lookup, list, filter, and summarize CRM accounts and contacts |
| **Query Tickets** | `/query-tickets` | List, filter, count, and summarize support tickets |

## Agents

| Agent | Best Model | Specialty |
|-------|-----------|-----------|
| **Data Query Agent** | Sonnet 4.6 | Multi-source NL query, parallel tool calls, answer synthesis |
| **Schema Explorer** | Sonnet 4.6 | Schema discovery, relationship detection, data map generation |

## Quick Start

```bash
# Discover what data sources you have
/discover-schema

# Ask a plain-English question across all sources
/query-data

# Look up a specific customer
/explore-crm

# List open high-priority tickets
/query-tickets
```

## The Core Value

Instead of logging into 5 different systems to answer one customer question, ask this plugin one question:

```
"What is the current status, open tickets, and recent orders for Acme Corp?"
```

**Result:**
```
Acme Corp (CRM: Active, $45,000/year, Rep: Jane Smith)

Open Tickets (2):
  #9871  High  "API integration failing"   3 days
  #9799  Med   "Export missing columns"   21 days

Recent Orders:
  Order #10234  $12,500  2026-03-15  Delivered
  Order #10198   $8,200  2026-02-28  Delivered

Sources: CRM (1 record), Tickets (2 open), Legacy DB (2 orders) — 0.8s
```

## MCP Tool Pattern

This plugin follows the MCP (Model Context Protocol) server pattern:

```
User question
  ↓
Natural language parsing (subject, verb, object, filters)
  ↓
Tool routing (crm / tickets / legacy_db / all)
  ↓
Parallel MCP tool calls
  ↓
Response synthesis
  ↓
Unified plain-language answer
```

## Connected Data Sources

| Tool | Data | Example Questions |
|------|------|-------------------|
| `crm` | Accounts, contacts, reps, contracts | "Who is Acme Corp's account rep?" |
| `tickets` | Support tickets, status, priority | "List all critical open tickets" |
| `legacy_db` | Orders, transactions, products | "Show Q1 orders over $10,000" |

## Schema Discovery

```
/discover-schema
```

Outputs `mcp-schema.md` with:
- All available tools and their field types
- Detected cross-source relationships (e.g., `crm.name ↔ tickets.customer`)
- Example queries for each data source
- Sample records to validate connectivity

## Use Cases

- Support team: complete customer picture before a call
- Sales: account status + recent tickets + order history in one view
- Operations: cross-system reporting without manual data assembly
- Engineering: natural language database exploration during development
- Management: aggregate counts and summaries across all systems

## Using the Actual Tool

The full source code from [NathanMaine/mcp-conversational-data-agent](https://github.com/NathanMaine/mcp-conversational-data-agent) is bundled in `src/`. It implements an MCP server with three data tools: `crm`, `tickets`, and `legacy_db`.

### Install

```bash
# Requires Python 3.8+
cd src
pip install -r requirements.txt
```

### Run the MCP Server

```bash
python src/server.py
```

The server exposes three MCP tools over stdio:

| Tool | Description |
|------|-------------|
| `crm` | Query CRM accounts, contacts, reps, and contracts |
| `tickets` | Query and filter support tickets by status, priority, customer |
| `legacy_db` | Query orders, transactions, and product data |

### Connect to Claude Code

Add the server to your Claude Code MCP config:

```json
{
  "mcpServers": {
    "conversational-data": {
      "command": "python",
      "args": ["/path/to/src/src/server.py"]
    }
  }
}
```

### Configuration

Edit the tool files in `src/src/tools/` to point at your actual data sources:

- `crm.py` — connect to your CRM (Salesforce, HubSpot, CSV, etc.)
- `tickets.py` — connect to your ticketing system (Jira, Zendesk, etc.)
- `legacy_db.py` — connect to your database (Postgres, SQLite, etc.)

The tool implementations are intentionally simple stubs — replace the return values with real queries to your systems.

### Full Documentation

See the [mcp-conversational-data-agent repo](https://github.com/NathanMaine/mcp-conversational-data-agent) for the full architecture guide and examples.

## Source Repository

Built from [NathanMaine/mcp-conversational-data-agent](https://github.com/NathanMaine/mcp-conversational-data-agent).

## License

MIT — see [LICENSE](LICENSE) for details.

More