Skip to main content
Glama
deenesjakoruzh
SkinnyMonk

EAG MCP Demo

by SkinnyMonk
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Local

EAG v3 — Demo de MCP con interfaz Prefab

Un servidor FastMCP con tres herramientas que utilizan (1) internet, (2) CRUD de archivos locales y (3) interfaz generativa a través de Prefab. Diseñado para ejecutarse dentro de Claude Desktop.

Herramientas

Herramienta

Tipo

Propósito

web_search(query)

internet

API de respuesta instantánea de DuckDuckGo: devuelve resumen + temas relacionados

notes_file(action, key, content)

CRUD local

una herramienta, con acciones (create / read / update / delete / list) sobre ./notes.json

show_dashboard()

UI

renderiza las notas guardadas como una cuadrícula de tarjetas Prefab (usa @mcp.tool(app=True))

Related MCP server: MCP + CrewAI Agentic Integration

Configuración

Requiere uv (brew install uv) y Claude Desktop.

cd /Users/nitingangwar/Documents/Code/eagv3/mcp
uv sync

Conectar a Claude Desktop

Edita ~/Library/Application Support/Claude/claude_desktop_config.json y añade (o fusiona) el bloque mcpServers:

{
  "mcpServers": {
    "eag-mcp-demo": {
      "command": "/opt/homebrew/bin/uv",
      "args": [
        "run",
        "--directory",
        "/Users/nitingangwar/Documents/Code/eagv3/mcp",
        "python",
        "server.py"
      ]
    }
  }
}

Reinicia Claude Desktop. El indicador de MCP debería mostrar eag-mcp-demo con tres herramientas.

Alternativa con Python estándar (sin uv)

python3 -m venv .venv
.venv/bin/pip install -e .

Luego, en la configuración de Claude Desktop, usa:

"command": "/Users/nitingangwar/Documents/Code/eagv3/mcp/.venv/bin/python",
"args": ["/Users/nitingangwar/Documents/Code/eagv3/mcp/server.py"]

Prompt de demostración

Pega esto en Claude Desktop después de conectar el servidor:

Busca los detalles de propiedad de Tata Sons mediante búsqueda web, guarda un resumen de 2 a 3 frases en mi archivo de notas bajo la clave tata_sons, luego abre el panel de control para que pueda verlo.

Comportamiento esperado:

  1. web_search("Tata Sons ownership") → Resumen de DDG.

  2. notes_file("create", "tata_sons", "<resumen>") → escribe en notes.json.

  3. show_dashboard() → renderiza una cuadrícula de tarjetas Prefab en línea con una tarjeta.

Vuelve a ejecutarlo con otra empresa (p. ej., "Reliance Industries") y el panel de control mostrará ahora dos tarjetas.

Prueba de humo

# Server starts and blocks on stdin (correct):
uv run python server.py

# Or interactive inspector:
npx @modelcontextprotocol/inspector uv run python server.py

En el inspector: lista las herramientas, llama a web_search("Tata Sons"), luego llama a notes_file("create", "x", "y") y confirma que notes.json se actualiza en el disco.

Alternativas

  • El renderizador de Prefab no se activa (Claude Desktop devuelve JSON en lugar de UI): cambia el decorador show_dashboard de @mcp.tool(app=True) a @mcp.tool simple y haz que devuelva una tabla Markdown construida a partir de notes.json.

  • DuckDuckGo devuelve resultados vacíos para la consulta: DDG solo tiene respuestas instantáneas para entidades conocidas. Usa empresas, personas o lugares famosos para la demostración.

Archivos

server.py        # FastMCP server with the 3 tools
notes.json       # backing store for notes_file (starts as {})
pyproject.toml   # deps: fastmcp[apps], prefab-ui==0.19.1, httpx

Available Tools

3 tools
notes_fileA

CRUD on local notes.json.

Actions:

  1. create - add a new note: action='create', key='foo', content='...'

  2. read - fetch one note: action='read', key='foo'

  3. update - replace existing: action='update', key='foo', content='...'

  4. delete - remove a note: action='delete', key='foo'

  5. list - list all keys: action='list'

ParametersJSON Schema
NameRequiredDescriptionDefault
actionYes
keyNo
contentNo

Output Schema

ParametersJSON Schema
NameRequiredDescription
resultYes

TDQS

A3.8/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations are provided, so the description must fully disclose behavioral traits. It lists actions but does not mention potential destructive effects (update/delete), error handling, atomicity, or file locking. The transparency is minimal beyond basic CRUD semantics.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is concise, with a single introductory line followed by a list of actions and examples. Every sentence serves a purpose, and the structure is front-loaded with the core purpose.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's simplicity (3 parameters, output schema present), the description covers the action parameter well. However, it does not clarify behavior for empty 'key' or 'content' defaults, nor potential side effects. Still, it is largely complete for a basic CRUD tool.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The schema has 0% description coverage, but the description adds meaningful context by providing examples for each action, clarifying how 'key' and 'content' are used per action. This adds value beyond the schema definition.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states 'CRUD on local notes.json' and enumerates each action with examples, making the tool's purpose unmistakable. It is distinct from siblings 'show_dashboard' and 'web_search'.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description implicitly conveys when to use the tool (to manage notes), but does not provide explicit guidance on when not to use it or discuss alternatives. Given the unrelated siblings, no confusion arises, but explicit exclusions are absent.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

show_dashboardA

Render saved notes as a Prefab card grid.

Call this AFTER saving notes via notes_file to display them visually to the user. Reads notes.json fresh on each call.

ParametersJSON Schema
NameRequiredDescriptionDefault

No parameters

TDQS

A4.6/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Despite no annotations, the description discloses that it reads notes.json fresh on each call, informing the agent about data source and potential performance. Missing edge cases like missing notes.json, but sufficient for a simple tool.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Two sentences with no redundancy. Front-loaded purpose, then usage and behavior. Every sentence adds value.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a zero-param, no-output-schema tool, the description covers main action, usage context, and a behavioral trait. Misses what happens when no notes exist, but overall adequate.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

No parameters exist, so schema coverage is 100%. The description provides no extra param info, but baseline for zero-param tools is 4.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool renders saved notes as a Prefab card grid. It uses a specific verb-resource combination and distinguishes itself from siblings (notes_file for saving, web_search unrelated).

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Explicitly instructs to call AFTER saving via notes_file, establishing a clear ordering and naming the prerequisite sibling. This guides the agent on when to invoke the tool.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

  1. 3 tool updatesv0.1.0
    • First observednotes_file
    • First observedshow_dashboard
    • First observedweb_search
Deploy Server
Shareable, User-Owned Context BackendPeople, applications, and agents read, write, and coordinate together in one context lake. Reach it over MCP, a CLI, a Python SDK, or REST.fulcradynamics.comAd

TDQS

A4.1/5.0

Scored across 3 tools

Disambiguation5/5

The three tools cover completely different areas: local note CRUD (notes_file), visual dashboard (show_dashboard), and external web search (web_search). There is no overlap or ambiguity.

Naming Consistency4/5

All tool names use snake_case and are descriptive, though the pattern varies slightly (notes_file is a noun, show_dashboard is verb+noun, web_search is noun+verb). Still consistent and readable.

Tool Count4/5

With only 3 tools, the set is small but well-suited for a demo server. Each tool has a clear purpose and none feel redundant. A few more tools might be expected for a full-featured service, but for a demo it's appropriate.

Completeness5/5

The notes_file tool provides full CRUD and listing, show_dashboard adds visualization, and web_search covers external search. There are no obvious gaps for the stated demo purpose.

Maintenance

ActivityInactive
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers

  • Custom Context MCP Server

    jiliangmason
    A
    license
    Not graded
    quality
    D
    maintenance
    A FastMCP-based server that supports SSE real-time event streaming and tool calling, with features for text-to-JSON conversion and grouping text by JSON templates.
    MIT

  • Forgetful

    ScottRBK
    A
    license
    A
    quality
    A
    maintenance
    A storage and retrieval MCP server for AI agents using FastMCP, enabling persistent knowledge base with semantic search and automatic linking.
    3
    231 PyPI
    302
    MIT

  • mnemos

    ncz-os
    A
    license
    Not graded
    quality
    B
    maintenance
    MCP server providing persistent memory, semantic search, versioned storage, webhook fanout, and unified LLM routing for AI agents via FastAPI runtime with multiple backend options.
    35
    Apache 2.0