Skip to main content
Glama
enesjakozh
stat-guy

Retrieval-Augmented Thinking MCP Server

by stat-guy
OverviewSchemaSourceRelated ServersDiscussionsAdmin
JavaScript

Servidor MCP de pensamiento aumentado por recuperación

Una implementación de servidor MCP (Protocolo de Contexto de Modelo) que mejora las capacidades de los modelos de IA con procesos de pensamiento estructurados y de recuperación aumentada. Este servidor permite cadenas de pensamiento dinámicas, rutas de exploración paralelas y ciclos de refinamiento recursivos para mejorar el razonamiento y la resolución de problemas.

Características

  • Cadenas de pensamiento adaptativas : mantiene flujos de razonamiento coherentes con capacidades de ramificación y revisión.

  • Generación iterativa de hipótesis : implementa ciclos de validación para la prueba de hipótesis

  • Coherencia del contexto : preserva el contexto a lo largo de caminos de razonamiento no lineales

  • Ajuste dinámico del alcance : admite exploración y refinamiento flexibles

  • Evaluación de calidad : evaluación en tiempo real de los procesos de pensamiento

  • Gestión de sucursales : gestiona rutas de exploración paralelas

  • Seguimiento de revisiones : gestiona ciclos de refinamiento recursivos

Related MCP server: Sequential Thinking MVP Server

Instalación

npm install @modelcontextprotocol/server-retrieval-augmented-thinking

Uso

Línea de comandos

mcp-server-retrieval-augmented-thinking

Uso programático

import { Server } from '@modelcontextprotocol/sdk/server';
import { StdioServerTransport } from '@modelcontextprotocol/sdk/server/stdio';

// Initialize and run the server
const server = new Server({
  name: 'retrieval-augmented-thinking',
  version: '0.1.0'
});

// Connect transport
const transport = new StdioServerTransport();
await server.connect(transport);

Configuración de herramientas

El servidor proporciona una herramienta con los siguientes parámetros:

  • thought (cadena): paso de razonamiento actual

  • thoughtNumber (número): Posición en la cadena de razonamiento

  • totalThoughts (número): Alcance estimado

  • nextThoughtNeeded (booleano): señal de continuación de cadena

  • isRevision (booleano, opcional): marca los pasos de refinamiento

  • revisesThought (número, opcional): Referencias al pensamiento objetivo

  • branchFromThought (número, opcional): punto de origen de la rama

  • branchId (cadena, opcional): identificador de rama

  • needsMoreThoughts (booleano, opcional): Señal de expansión del alcance

Funciones avanzadas

Análisis de la cadena de pensamiento

El servidor rastrea varias métricas para la calidad de la cadena de pensamiento:

  • Eficacia de la cadena

  • Impacto de la revisión

  • Tasa de éxito de la sucursal

  • Calidad general

  • Métricas de pensamiento individual (complejidad, profundidad, calidad, impacto)

Reconocimiento de patrones

Analiza patrones de pensamiento para:

  • Estructuras de razonamiento

  • Preservación del contexto

  • Validación de hipótesis

  • Coherencia de la solución

Desarrollo

# Build
npm run build

# Watch mode
npm run watch

Contribuyendo

¡Agradecemos sus contribuciones! Lea nuestras normas de contribución y envíe solicitudes de incorporación de cambios.

Licencia

Instituto Tecnológico de Massachusetts (MIT)

Available Tools

1 tool
ratB

A context-aware reasoning system that orchestrates structured thought processes through dynamic trajectories.

Core Capabilities:

  • Maintains adaptive thought chains with branching and revision capabilities

  • Implements iterative hypothesis generation and validation cycles

  • Preserves context coherence across non-linear reasoning paths

  • Supports dynamic scope adjustment and trajectory refinement

Reasoning Patterns:

  • Sequential analysis with backtracking capability

  • Parallel exploration through managed branch contexts

  • Recursive refinement via structured revision cycles

  • Hypothesis validation through multi-step verification

Parameters: thought: Structured reasoning step that supports: • Primary analysis chains • Hypothesis formulation/validation • Branch exploration paths • Revision proposals • Context preservation markers • Verification checkpoints

next_thought_needed: Signal for continuation of reasoning chain thought_number: Position in current reasoning trajectory total_thoughts: Dynamic scope indicator (adjustable) is_revision: Marks recursive refinement steps revises_thought: References target of refinement branch_from_thought: Indicates parallel exploration paths branch_id: Context identifier for parallel chains needs_more_thoughts: Signals scope expansion requirement

Execution Protocol:

  1. Initialize with scope estimation

  2. Generate structured reasoning steps

  3. Validate hypotheses through verification cycles

  4. Maintain context coherence across branches

  5. Implement revisions through recursive refinement

  6. Signal completion on validation success

The system maintains solution integrity through continuous validation cycles while supporting dynamic scope adjustment and non-linear exploration paths.

ParametersJSON Schema
NameRequiredDescriptionDefault
thoughtYesYour current thinking step
branchIdNoBranch identifier
isRevisionNoWhether this revises previous thinking
thoughtNumberYesCurrent thought number
totalThoughtsYesEstimated total thoughts needed
revisesThoughtNoWhich thought is being reconsidered
branchFromThoughtNoBranching point thought number
needsMoreThoughtsNoIf more thoughts are needed
nextThoughtNeededYesWhether another thought step is needed

TDQS

B3.1/5.0
Behavior2/5

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

No annotations are provided, so the description carries full burden. It describes capabilities like branching and revisions but does not disclose side effects, statefulness, or what the tool returns when invoked. The behavior is described conceptually rather than practically.

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

Conciseness3/5

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

Structured with sections and bullet points, making it skimmable. However, it is somewhat verbose with overlapping sections (Core Capabilities vs Reasoning Patterns). Could be tightened without losing meaning.

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

Completeness2/5

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

The description is thorough about reasoning patterns but omits what the tool actually returns or how the agent should interpret the result. With no output schema and no annotations, this is a significant gap for correct invocation and result handling.

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?

Schema covers all 9 parameters with basic descriptions. The description adds value by categorizing what each parameter supports (e.g., 'thought' supports hypothesis formulation, branch exploration, etc.) and providing domain context for fields like branch_id and is_revision.

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

Purpose4/5

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

Clearly states it is a context-aware reasoning system that orchestrates structured thought processes. The verb 'orchestrates' and resource 'thought processes' are explicit, but the tool's operational purpose for an agent is somewhat abstract; no siblings to differentiate from.

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?

Provides an Execution Protocol with steps, implying a multi-step reasoning workflow. However, it does not explicitly state when to use this tool versus other tools, as there are no siblings, nor does it give conditions for non-use.

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. 1 tool updatev0.1.0
    • First observedrat
Deploy Server
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

B3.1/5.0

Scored across 1 tool

Disambiguation5/5

Only one tool exists, so there is no possibility of confusion or misselection. The tool's purpose, while broad, is clearly the sole entry point.

Naming Consistency2/5

The single tool name 'rat' is vague and does not follow a clear verb_noun pattern or any recognizable convention. With only one tool, there's no consistent pattern to infer, and the name appears arbitrary.

Tool Count3/5

Exposing just one tool is on the lower end of acceptable, borderline 'thin.' Although the tool is highly capable, a server focused on 'Retrieval-Augmented Thinking' might benefit from separate tools for retrieval and reasoning sub-tasks.

Completeness2/5

The server name implies both retrieval and thinking, but the tool only covers the reasoning aspect, missing retrieval or external context fetching. This leaves a significant gap in the expected functionality, and the single tool is overloaded with parameters rather than modularly covering the domain.

Maintenance

ActivityInactive
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers

Appeared in Searches