convez.ai
MCP ServerDevOps & InfraestruturaIntermediário
kawehtaher

DeployCheck

kawehtaher/deploycheck

Verifica se seu deployment está pronto para produção em um comando

Ver no GitHub
InicialPluginScore · de 0 a 100

Servidor MCP que verifica a prontidão de deployments para Claude Code. Executa health checks, valida status de CI, compara tags Docker e gera diffs de release em uma única varredura, permitindo que Claude Code identifique se ambientes estão seguros para produção.

Instalação

$npx convez-ai kawehtaher/deploycheck

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
1
Forks
0
Acompanhando
0
Saúde
42%
Último commit
há 7 meses
Versão
—
Issues abertas
0
Contribuidores
1

Casos de uso

  • Validar saúde de ambientes antes de deployments críticos
  • Comparar versões Docker entre staging e produção
  • Verificar status de pipelines CI/CD automaticamente
  • Gerar relatório consolidado go/no-go de readiness

Veredito

Ferramenta sólida e bem-estruturada para automação de verificações pré-deploy. Excelente integração com Claude Code via MCP, cobrindo casos reais de validação de pipeline com suporte a GitHub e GitLab.

README

A tradução para o português deste README ainda está na fila.

deploycheck

Claude Code Plugin License: MIT TypeScript

Deploy readiness checker for Claude Code. An MCP server that lets Claude verify if your environments are safe to ship — health checks, CI status, Docker tag comparison, and release diffs in one sweep.

Prerequisites

  • Node.js >= 18.0.0
  • npm
  • A GitHub or GitLab personal access token (see Token scopes)

Quick start

# Clone
git clone https://github.com/kawehtaher/deploycheck.git
cd deploycheck

# Install and build
cd server
npm install
npm run build
cd ..

# Configure
cp deploycheck.example.yml deploycheck.yml
# Edit deploycheck.yml with your project details

# Set tokens
export GITHUB_TOKEN=ghp_...      # or GITLAB_TOKEN for GitLab
export DOCKER_HUB_TOKEN=...      # optional, for private images

# Install the plugin in Claude Code
claude plugin add ./

How it works

deploycheck runs as a Model Context Protocol (MCP) server over stdio. When installed as a Claude Code plugin, it exposes five tools that Claude can call to inspect your deployment pipeline:

  1. Pings health endpoints on your configured environments
  2. Queries GitHub Actions or GitLab CI for the latest pipeline status
  3. Compares Docker image tags between environments via Docker Hub
  4. Diffs commits between staging and production refs via the GitHub API
  5. Combines all of the above into a single go/no-go readiness report

Configuration

Place a deploycheck.yml in your project root, or set the DEPLOYCHECK_CONFIG environment variable to point to it.

project: my-app

environments:
  staging:
    url: https://staging.myapp.com
    health_endpoint: /api/health
    docker_image: myapp/api
    docker_tag_source: git       # "git" or "docker-hub"

  production:
    url: https://myapp.com
    health_endpoint: /api/health
    docker_image: myapp/api
    docker_tag_source: git

ci:
  provider: github-actions       # or "gitlab-ci"
  repo: owner/repo
  branch: main

git:
  staging_ref: staging
  production_ref: production

Config reference

FieldRequiredDescription
projectYesProject name
environments.<name>.urlYesBase URL for the environment
environments.<name>.health_endpointYesPath to health endpoint (must start with /)
environments.<name>.docker_imageYesDocker image in [namespace/]repository format
environments.<name>.docker_tag_sourceNo"git" (default) or "docker-hub"
ci.providerYes"github-actions" or "gitlab-ci"
ci.repoYesRepository identifier (owner/repo for GitHub, project path or numeric ID for GitLab)
ci.branchNoBranch to check CI status for (default: main)
git.staging_refNoBranch or tag representing staging (default: staging)
git.production_refNoBranch or tag representing production (default: production)

Environment variables

VariableRequiredDescription
GITHUB_TOKENYes (GitHub)GitHub personal access token
GITLAB_TOKENYes (GitLab)GitLab personal access token
DOCKER_HUB_TOKENNoDocker Hub token for private images (public images work without it)
DEPLOYCHECK_CONFIGNoPath to config file (default: ./deploycheck.yml)

Token scopes

ProviderRequired scope
GitHubactions:read (for CI status), repo or contents:read (for release diff)
GitLabread_api
Docker HubRead access (only needed for private images)

Available tools

check_health

Ping one or all environment health endpoints. Returns status code, response time, and healthy/unhealthy verdict. Health checks timeout after 5 seconds.

ParameterRequiredDescription
environmentYesEnvironment name from config, or "all" to check every environment
> Check if staging is healthy
> Check health on all environments

get_ci_status

Get the latest CI pipeline status for a branch. Supports GitHub Actions and GitLab CI.

ParameterRequiredDescription
branchNoBranch to check (defaults to ci.branch from config)
> What's the CI status on main?
> Is the latest build passing?

compare_envs

Compare two environments side-by-side — Docker image tags, health status, and differences.

ParameterRequiredDescription
sourceYesSource environment name (e.g. "staging")
targetYesTarget environment name (e.g. "production")
> Compare staging and production
> Are staging and production running the same version?

get_release_diff

Show what would be deployed — commits between staging and production refs. GitHub only — uses the GitHub compare API.

ParameterRequiredDescription
source_refNoSource branch/tag (defaults to git.staging_ref from config)
target_refNoTarget branch/tag (defaults to git.production_ref from config)
> What commits are waiting to be deployed?
> Show me the diff between staging and production

run_all_checks

Full readiness sweep. Runs health checks on all environments, CI status, environment comparison, and release diff concurrently. Returns a go/no-go recommendation with blockers and warnings.

ParameterRequiredDescription
sourceNoSource environment for comparison (default: "staging")
targetNoTarget environment for comparison (default: "production")

Readiness verdict:

  • READY — all checks pass, no blockers or warnings
  • WARNINGS — no blockers, but issues worth reviewing (tag mismatches, cancelled CI, etc.)
  • NOT READY — blockers found (unhealthy environments, failed CI)
> Are we ready to deploy?
> Run a full deploy check
> Is it safe to ship?

Example conversations

Quick health check:

You: Is staging healthy?
Claude: [runs check_health for staging]
> staging: HEALTHY — HTTP 200, 142ms

Pre-deploy review:

You: Are we ready to deploy to production?
Claude: [runs run_all_checks]
> Deploy Readiness: WARNINGS
> Health: staging OK, production OK
> CI: SUCCESS on main (abc1234)
> Tags: staging=v1.3.0, production=v1.2.9 (mismatch)
> 4 commits ahead, 12 files changed
> Warning: Tag mismatch between staging and production

Investigating a failure:

You: Why isn't production healthy?
Claude: [runs check_health for production]
> production: UNHEALTHY — HTTP 503, 2341ms
> The production health endpoint is returning 503. This usually indicates
> the service is overloaded or in maintenance mode.

Project structure

deploycheck/
├── .claude-plugin/          # Claude Code plugin manifest
│   └── plugin.json
├── .mcp.json                # MCP server configuration
├── skills/
│   └── deploy-check/
│       └── SKILL.md         # Skill prompt for Claude
├── server/
│   └── src/
│       ├── index.ts         # MCP server entry point (stdio transport)
│       ├── types.ts         # Shared type definitions
│       ├── config/
│       │   ├── schema.ts    # Zod config schema + validation
│       │   └── loader.ts    # YAML config file loader
│       ├── tools/           # MCP tool handlers
│       │   ├── check-health.ts
│       │   ├── get-ci-status.ts
│       │   ├── compare-envs.ts
│       │   ├── get-release-diff.ts
│       │   └── run-all-checks.ts
│       └── providers/       # External service integrations
│           ├── ci/
│           │   ├── github-actions.ts
│           │   └── gitlab-ci.ts
│           └── registry/
│               └── docker-hub.ts
├── deploycheck.example.yml  # Example configuration
└── tsconfig.json

Development

cd server
npm install
npm run dev          # watch mode
npm test             # run tests
npm run build        # production build
npm run lint         # check linting
npm run typecheck    # type check without emitting

Contributing

  1. Fork the repository
  2. Create a feature branch (git checkout -b feature/my-feature)
  3. Make your changes with tests
  4. Ensure npm test passes
  5. Submit a pull request

License

MIT - Kaveh Taher

Ver o original no GitHub

Bom para

Relacionados