Skip to main content
Glama
hallzyx

zero-slop-mcp-server

by hallzyx

zero-slop-mcp-server

MCP remoto con una tool: zero_slop_submit_assessment. Recibe el análisis del harness (Claude Code), valida, calcula total y lo guarda en Convex assessments vía evaluations.save.

Clerk es solo la web. Esta tool se autentica con Bearer → sha256mcpTokens. Nunca getUserIdentity().

githubUsername es el evaluado (él mismo, o un senior). Si el senior evalúa a alguien sin fila, Convex puede crear el developer — la cuenta Clerk se vincula después en el onboarding.

Setup

  1. Copia convex-dropin/mcp.tsconvex/mcp.ts y convex-dropin/evaluations.tsconvex/evaluations.ts. Despliega Convex.

  2. Al emitir tokens (en la web), guarda sha256(plaintext) en mcpTokens.tokenHash. No borres filas: usá revokedAt.

  3. Copia el skill skills/zero-slop-assessment/ al harness de Claude Code.

  4. Configura env y arranca:

cp .env.example .env
npm install
npm run dev
  • MCP: http://127.0.0.1:3000/mcp (Bearer obligatorio)

  • Health: http://127.0.0.1:3000/health (sin auth)

Related MCP server: Arkheia Hallucination Detection MCP

Cliente (Claude Code / Cursor)

{
  "mcpServers": {
    "zero-slop": {
      "url": "http://127.0.0.1:3000/mcp",
      "headers": {
        "Authorization": "Bearer DEVELOPER_MCP_TOKEN"
      }
    }
  }
}

El LLM puntúa con el skill (references/rubric.md) y llama la tool. No envía scores.total: avg(decisionUnderstanding, riskDetection), un decimal. explanationQuality se guarda y tiene feedback, no entra en el total. Cada llamada inserta (historial).

Sin el skillId correcto la tool rechaza. El agente debe buscar el skill en dos sitios: skills instaladas del harness, y este repo (skills/zero-slop-assessment/SKILL.md). Con uno alcanza. El MCP remoto no ve el disco; si no está en ninguno, te lo tiene que pedir.

Producción (Dokploy)

Dokploy ya trae Traefik + Let's Encrypt. No uses Caddy ni publiques 80/443 en el compose.

Lo más simple: una Application que construye el Dockerfile.

  1. DNS A/AAAA de mcp.tudominio.com al VPS de Dokploy.

  2. Nueva aplicación → Git o Compose. Si es Compose, el archivo es compose.yml (servicio mcp). Compose path en Dokploy: ./compose.yml.

  3. Environment:

    • HOST=0.0.0.0

    • PORT=3000

    • CONVEX_URL=https://….convex.cloud

    • ALLOWED_HOSTS=mcp.tudominio.com (Host header; Compose also accepts MCP_DOMAIN)

  4. Domains: host mcp.tudominio.com, servicio mcp, puerto interno 3000, certificado Let's Encrypt.

  5. Health path: /health (sin Bearer; ábrelo en el navegador para comprobar Traefik). MCP: https://mcp.tudominio.com/mcp (POST + Bearer; GET en el browser no es una página).

{
  "mcpServers": {
    "zero-slop": {
      "url": "https://mcp.tudominio.com/mcp",
      "headers": {
        "Authorization": "Bearer DEVELOPER_MCP_TOKEN"
      }
    }
  }
}

Local sigue siendo npm run dev.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.

No tool schema history has been recorded yet.

Maintenance

ActivityMaintained
ResponsivenessNo issues

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Connectors

Related MCP Servers

  • A
    license
    Not graded
    quality
    C
    maintenance
    Provides advanced evaluation tools for assessing AI safety, alignment, and performance of LLM outputs. Enables programmatic evaluation of quality, safety metrics like toxicity and PII detection, and operational metrics including carbon footprint and cost estimation.
    4
    Apache 2.0
  • A
    license
    Not graded
    quality
    B
    maintenance
    Enables LLM evaluation and observability by uploading documents, building test sets, running RAG pipelines, and automatically scoring answers for groundedness, hallucination risk, retrieval quality, latency, and cost, with tools exposed to MCP-compatible clients.
    1
    MIT

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/hallzyx/zero-slop-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server