ZTL Judge

Zero-trust logic judge: your AI writes a claim as a ZFL table, the ZTL core judges it.

¿Debería usar esto?

Calidad y seguridad

B
Calidad de la descripción
90%
Integridad del esquema
40%
Calidad de los nombres
80%
Riesgo de envenenamiento
100%
Coincidencia de permisos
100%
Cumplimiento del protocolo
100%

Hallazgos (1)

  • LOWTool 'language' description lacks action verben language

Basado en el análisis automatizado de las definiciones de herramientas y el cumplimiento del protocolo.

Costo de contexto

~429Tokens (definiciones de herramientas)
~528 BTamaño de respuesta típico
Impacto mínimo en la atención (0.34% del contexto de 128k)

Este es el número aproximado de tokens que se consumen cada vez que las herramientas del servidor se cargan en el contexto de un modelo. Los recuentos más altos reducen la atención disponible para otras tareas.

Instalar

Instalación con un clic

Agrega esto a tu archivo `claude_desktop_config.json`:

{
  "mcpServers": {
    "ztl-judge": {
      "url": "https://api.vitalyreznik.com/mcp"
    }
  }
}

Puntos de conexión remotos

https://api.vitalyreznik.com/mcpstreamable-http

Qué puede hacer

Inventario de herramientas

Herramientas (3)

🟢 Solo lectura🟡 Escritura🔴 Eliminación⚪ Desconocido
⚪language

The ZFL language: the columns of a row, the document fields, their meaning and rules.

Esquema de entrada

{
  "type": "object",
  "properties": {},
  "title": "languageArguments"
}

Esquema de salida

{
  "type": "object",
  "additionalProperties": true,
  "title": "languageDictOutput"
}
⚪examples

Worked examples: questions already written as ZFL documents, ready to judge.

Esquema de entrada

{
  "type": "object",
  "properties": {},
  "title": "examplesArguments"
}

Esquema de salida

{
  "type": "object",
  "additionalProperties": true,
  "title": "examplesDictOutput"
}
🟢judge(document)

Judge a ZFL document with ZTL, a zero-trust logic: three values, two-valued connectives. Args: document: The ZFL document, as an object or as JSON text: {"rows": [{"name": ..., "means": ..., "status": ..., "ground": ...}], "claim": ...}. Returns: The verdict with its disposition and grade, the receipt, the instruments that applied, the issues found, and what the core read. Read the verdict WITH its disposition: T EARNED = established; F REFUTED = false; F OPEN or Z OPEN = NOT ESTABLISHED, it could still turn either way (do not report it as false) — `why` and `unverified` say what to check; ON CREDIT = holds only on an unverified ground. A compound claim gets T or F; a claim that is a single name gets that name's own value, Z while unverified. The full report is returned whatever `ask` says.

Esquema de entrada

{
  "type": "object",
  "properties": {
    "document": {
      "anyOf": [
        {
          "additionalProperties": true,
          "type": "object"
        },
        {
          "type": "string"
        }
      ],
      "title": "Document"
    }
  },
  "required": [
    "document"
  ],
  "title": "judgeArguments"
}

Esquema de salida

{
  "type": "object",
  "additionalProperties": true,
  "title": "judgeDictOutput"
}

Comunidad

Califica este servidor

Evidencia

Observaciones recientes

verificadoversión no registrada3 herramientas