create-server-json

Estrelas0

Forks1

Atualizado10 de março de 2026 às 01:27

Create a valid MCP Registry server.json file for an MCP server. Use when the user wants to generate, scaffold, or write a server.json for any MCP server project.

Instalação

Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.

Fonte

pulsemcp

pulsemcp/agentic-engineering

Abrir repositório GitHub Ver repositórios do creator

Download

Ocupações relacionadasSOC

Baseado na classificação ocupacional SOC

Desenvolvedores de softwareInformática e Matemática·SOC 15-1252

SKILL.md

readonly

name	create-server-json
description	Create a valid MCP Registry server.json file for an MCP server. Use when the user wants to generate, scaffold, or write a server.json for any MCP server project.
user-invocable	true
argument-hint	["server-name-or-url"]

create-server-json

Sequencing Checklist

Fetch the latest schema from GitHub and save to /tmp/server.schema.json
Read the downloaded schema to understand valid fields and constraints
Pull the generic-server-json.md reference into context for examples
Check for an existing server.json on registry.modelcontextprotocol.io
Check for an existing server.json on pulsemcp.com/servers
Fetch any additional context provided by the user
Construct the server.json (set $schema to the schema URL, add top-level metadata, packages, and/or remotes)
Validate the generated file against the schema using ajv
Fix any validation errors and re-validate until it passes

Create a valid MCP Registry server.json file for an MCP server project. The user provided this context about the server:

$ARGUMENTS

What is server.json?

A server.json is a standardized self-description file for MCP (Model Context Protocol) servers. It follows the schema defined by the MCP Registry working group and enables AI agents and MCP clients to automatically discover, install, and configure servers.

Before generating the file:

Check the latest schema version at https://github.com/modelcontextprotocol/static/tree/main/schemas to find the most recent dated directory (e.g., 2025-12-11).

Download the schema to a temp file so it can be used for validation later:

curl -sL "https://raw.githubusercontent.com/modelcontextprotocol/static/main/schemas/<latest-date>/server.schema.json" -o /tmp/server.schema.json

Read /tmp/server.schema.json to understand the exact fields, types, and constraints — use this as the authoritative reference for what fields are valid. Do NOT invent fields that are not in the schema.
Use that schema URL as the $schema value in the generated file.
Pull https://github.com/modelcontextprotocol/registry/blob/main/docs/reference/server-json/generic-server-json.md into context for examples of what a variety of server.json files might look like.

Instructions

When the user asks you to create a server.json, start by doing the following:

Check for a server.json file corresponding to the server at hand on registry.modelcontextprotocol.io
Check for a server.json file on pulsemcp.com/servers
Fetch any context provided by the user that could be useful for putting together the ensuing information

In some cases, you might be able to just call a finding from (1) or (2) the final version. In other cases, you might only get a partial server.json that either (1) isn't up to date, or (2) doesn't actually contain comprehensive remotes and/or packages entries for how we can actually run the server.

Using the context you collected, you should be able to construct a comprehensive server.json that describes a meaningful amount of top level metadata + structured, comprehensive metadata for how to run/connect to a server embedded into packages and/or remotes.

Validation

After writing the server.json file, validate it against the downloaded schema using a JSON Schema validator. First ensure a validator is available, then run it:

# Install validator if needed (check first)
which ajv || npm install -g ajv-cli

# Validate the generated file against the schema
# The schema uses draft-07 and non-standard keywords like "example", so use --spec=draft7 --strict=false
ajv validate -s /tmp/server.schema.json -d <path-to-generated-server.json> --spec=draft7 --strict=false

If validation fails, read the errors, fix the generated file, and re-validate until it passes. Do not consider the task complete until validation succeeds.

Mais deste repositório

mesmo repositório

adjust-claude-mcp-json

pulsemcp/agentic-engineering

Generate or update a Claude Code .mcp.json file from the project's mcp-servers/mcp.json. Use when the user wants to configure Claude Code's MCP servers from the trusted server set.

2026-03-100

create-mcp-json

pulsemcp/agentic-engineering

Create an mcp.json client configuration from a server.json. Use when the user wants to generate a ready-to-use MCP client config for connecting to an MCP server.

2026-03-100

perform-release

pulsemcp/agentic-engineering

Trigger a production release via GitHub Actions, monitor the ECS deployment, and verify the release is healthy. Use this skill when it's time to ship — after the PR is merged, CI is green, and staging has been verified. Also use it when asked to deploy, release, or ship to production.

2026-02-090

wait-for-ci

pulsemcp/agentic-engineering

Block until CI passes or fails on the current PR. Use this skill every time you push a commit and intend to tell the user the work is done — never present a pushed commit as complete without first confirming CI is green. Also use it when waiting on CI before merging, deploying, or moving on to the next task in a multi-step plan.

2026-02-090

initial-pr-review

pulsemcp/agentic-engineering

Run a structured first-pass code review on a PR. Use this skill when asked to review a PR, when opening a PR that should be sanity-checked before requesting human review, or when picking up someone else's PR to understand what changed and whether it's ready to merge.

2026-02-090

test-in-staging

pulsemcp/agentic-engineering

Deploy the current branch to staging and verify it works end-to-end. Use this skill after CI passes when you need to confirm changes actually work in a real environment before merging — especially for user-facing changes, API modifications, or anything that interacts with external services. Do not consider a PR "ready to merge" for non-trivial changes without a staging check.

2026-02-090

create-server-json

Sequencing Checklist

Fetch the latest schema from GitHub and save to /tmp/server.schema.json

Read the downloaded schema to understand valid fields and constraints

Pull the generic-server-json.md reference into context for examples

Check for an existing server.json on registry.modelcontextprotocol.io

Check for an existing server.json on pulsemcp.com/servers

Fetch any additional context provided by the user

Construct the server.json (set $schema to the schema URL, add top-level metadata, packages, and/or remotes)

Validate the generated file against the schema using ajv

Fix any validation errors and re-validate until it passes

Create a valid MCP Registry server.json file for an MCP server project. The user provided this context about the server:

$ARGUMENTS

What is server.json?

Before generating the file:

Check the latest schema version at https://github.com/modelcontextprotocol/static/tree/main/schemas to find the most recent dated directory (e.g., 2025-12-11).

Download the schema to a temp file so it can be used for validation later:

curl -sL "https://raw.githubusercontent.com/modelcontextprotocol/static/main/schemas/<latest-date>/server.schema.json" -o /tmp/server.schema.json

Read /tmp/server.schema.json to understand the exact fields, types, and constraints — use this as the authoritative reference for what fields are valid. Do NOT invent fields that are not in the schema.

Use that schema URL as the $schema value in the generated file.

Pull https://github.com/modelcontextprotocol/registry/blob/main/docs/reference/server-json/generic-server-json.md into context for examples of what a variety of server.json files might look like.

Instructions

When the user asks you to create a server.json, start by doing the following:

Check for a server.json file corresponding to the server at hand on registry.modelcontextprotocol.io

Check for a server.json file on pulsemcp.com/servers

Fetch any context provided by the user that could be useful for putting together the ensuing information

Validation

After writing the server.json file, validate it against the downloaded schema using a JSON Schema validator. First ensure a validator is available, then run it:

# Install validator if needed (check first) which ajv || npm install -g ajv-cli # Validate the generated file against the schema # The schema uses draft-07 and non-standard keywords like "example", so use --spec=draft7 --strict=false ajv validate -s /tmp/server.schema.json -d <path-to-generated-server.json> --spec=draft7 --strict=false

If validation fails, read the errors, fix the generated file, and re-validate until it passes. Do not consider the task complete until validation succeeds.