MCP Gestão de Tarefas
This MCP server connects a development agent to the Gestão de Tarefas system so it can read and manage projects, demands, subtasks, and sprints, including offline queuing and sync.
Detect the active project and whether the MCP is enabled for it (via
.gestaotarefas.jsonor Git), blocking unidentified or ignored repositories.List projects (optionally filtered by status) and list sprints, newest first.
Create demands with HTML description, priority, status, responsible, dates, sprint, ITIL classification, and story points.
List active demands by project/status/responsible and fetch full details of a demand including its subtasks and assignees.
Update existing demand fields and associate a demand with a sprint (rejected if already in another sprint).
Create subtasks linked to a demand (accepts an offline client_id) and update one or many subtasks (up to 500) — status, title, description, due date.
Bulk-complete or change status of subtasks by ID list, or all pending subtasks of a demand.
Renew the authentication session using saved or newly supplied credentials.
Check connectivity, token validity, and the pending offline queue size, and force-sync queued demands/subtasks.
Reads local Git repositories (scanned under configured directories) to collect the author's commits and files modified within a given time window, which are used to automatically build and register daily stand-up reports through the Gestão de Tarefas tools.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@MCP Gestão de Tarefas Listar projetos ativos vinculados a este repositório"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP Task Management
Standalone server based on the Model Context Protocol (MCP) to integrate development agents with the Task Management system.
Requirements
Node.js 22.5 or higher;
access to the Task Management server;
valid credentials or an authentication token.
Related MCP server: Cherry Studio Task Board
Installation
npm install
npm run buildConfiguration and authentication
The assistant configures the server URL and authentication locally. The URL
can be provided via the GESTAO_TAREFAS_API_URL variable, the --api-url
argument, or during the assistant's execution.
npm run setupWhen the URL is not configured, the assistant will prompt:
URL do Gestão de Tarefas:It is also possible to provide access data via arguments:
npm run setup -- \
--api-url=https://seu-servidor-de-gestao.exemplo \
--email=seu.email@empresa.gov.br \
--password=suasenhaTo use an existing web session or token:
npm run setup -- --api-url=https://seu-servidor-de-gestao.exemplo --token=SEU_TOKENThe assistant attempts to authenticate via a JSON endpoint and, if not available,
uses the web form with CSRF. Credentials are not versioned. The session or
token is saved in the local .env and in ~/.gestao-tarefas-mcp/config.json.
An existing web session or token can be provided to the assistant or
configured in GESTAO_TAREFAS_API_TOKEN.
Environment variables
The .env.example file contains the reference configuration:
GESTAO_TAREFAS_API_URL=https://seu-servidor-de-gestao.exemplo
GESTAO_TAREFAS_API_TOKEN=cole_a_sessao_web_ou_token_aqui
OFFLINE_QUEUE_PATH=~/.gestao-tarefas-mcp/queue.sqlite
REQUEST_TIMEOUT_MS=5000
IGNORE_PREFEITURA=true
IGNORED_PROJECT_PATTERNS=prefeitura,pmsp,pref-,pm-,pm_Do not place tokens, passwords, or cookies in versioned files.
Features
The server provides tools for:
detecting the current project and its link to the system;
listing projects, demands, and sprints;
creating demands and subtasks;
updating subtasks and querying demand details;
associating demands with sprints;
operating with an offline queue and syncing items later;
checking connectivity and authentication status.
Projects marked as external or belonging to the City Hall can be
automatically ignored. This behavior is controlled by
IGNORE_PREFEITURA, IGNORED_PROJECT_PATTERNS, and the
.gestaotarefas.json file.
Execution
The server uses stdio and must be run from the build:
npm run build
node dist/index.jsExample configuration for an MCP client:
{
"mcpServers": {
"gestao-tarefas": {
"command": "node",
"args": ["/caminho/para/mcp-gestao-tarefas/dist/index.js"],
"env": {
"GESTAO_TAREFAS_API_URL": "https://seu-servidor-de-gestao.exemplo",
"GESTAO_TAREFAS_API_TOKEN": "CONFIGURADO_LOCALMENTE",
"OFFLINE_QUEUE_PATH": "~/.gestao-tarefas-mcp/queue.sqlite",
"IGNORE_PREFEITURA": "true"
}
}
}
}After changing the code, run npm run build and restart the MCP client.
The component and flow overview is in ARCHITECTURE.md.
Development agent configuration
An agent with workspace access can configure the server following this sequence:
Configure o servidor MCP deste projeto.
1. Leia o README.md e identifique o cliente MCP em uso.
2. Use o diretório atual como diretório do projeto.
3. Instale as dependências apenas se necessário.
4. Execute npm run build.
5. Registre o servidor usando node dist/index.js e preserve as demais
configurações existentes do cliente.
6. Solicite a URL do servidor e a autenticação caso ainda não estejam
configuradas. Não exiba nem grave tokens ou senhas no chat.
7. Inicie o servidor ou informe que o cliente precisa ser reiniciado.The agent must ask for confirmation before overwriting existing configurations or installing dependencies. If it does not have permission to change the client configuration, it must provide instructions for manual application.
Project configuration
The .gestaotarefas.json file must be created at the root of the repository
that will be integrated with Task Management, at the same level as the .git
directory:
meu-projeto/
├── .git/
├── .gestaotarefas.json
├── package.json
└── src/The MCP looks for this file starting from the current directory and continues
going up through parent folders. Thus, a configuration placed in a common
folder can also be shared by multiple repositories. The alternative name
.gestao-tarefas.json is also accepted.
To link the repository to a project, create meu-projeto/.gestaotarefas.json
with the corresponding identifier:
{
"projeto_id": 1,
"nome": "Gestão de Tarefas",
"departamento": "TI"
}To disable the MCP only in this repository, use the same file:
{
"nome": "Projeto externo",
"tipo": "externo",
"ignorado": true,
"motivo": "Projeto fora do escopo"
}Tests
npm testThe suite covers authentication, context detection, bypass of ignored projects, HTTP communication, offline queue, synchronization, and MCP tools.
Available Tools
14 toolsassociar_demanda_sprintA
Associa uma demanda existente a uma sprint. A API rejeita a operação se a demanda já estiver em outra sprint.
Inputs: sprint_id (required), demanda_id (required).
atualizar_demandaA
Atualiza campos de uma demanda existente (projeto, título, descrição, prioridade, status, responsável, sprint, datas, classificação ITIL, estimativa, solicitante).
Inputs: demanda_id (required), projeto_id (required), status, titulo, +12 more.
atualizar_subtarefaC
Atualiza campos de uma subtarefa (título, descrição, status ou data limite) ou atualiza múltiplos IDs em lote.
Inputs: status, titulo, descricao, data_limite, +2 more.
concluir_subtarefasA
Conclui ou altera o status de múltiplas subtarefas em uma única operação (passando uma lista de subtarefa_ids ou o demanda_id para concluir todas as pendentes daquela demanda).
Inputs: status, demanda_id, subtarefa_ids.
criar_demandaA
Registra uma nova demanda no sistema de Gestão de Tarefas. A operação é bloqueada para projetos não identificados ou explicitamente ignorados. Se a rede estiver offline, salva na fila local.
Inputs: titulo (required), descricao (required), prioridade (required), projeto_id (required), +11 more.
criar_subtarefaA
Cria uma subtarefa técnica vinculada a uma demanda existente. A operação é bloqueada para projetos não identificados ou explicitamente ignorados. Se offline, armazena na fila local para posterior envio.
Inputs: titulo (required), demanda_id (required), descricao, data_limite, +2 more.
listar_demandas_ativasB
Lista as demandas abertas ou em andamento de um projeto para análise de contexto e vínculo de subtarefas.
Inputs: projeto_id (required), status, responsavel_id.
listar_projetosC
Lista todos os projetos cadastrados no sistema de Gestão de Tarefas.
Inputs: status.
listar_sprintsA
Lista as sprints visíveis, em ordem da mais recente para a mais antiga.
Inputs: no named parameters.
obter_contexto_projetoA
Identifica o projeto ativo com base no diretório de trabalho atual (via .gestaotarefas.json ou Git), verificando se o MCP deve operar ou se está desativado para projetos externos ou não identificados.
Inputs: diretorio_path.
obter_detalhes_demandaB
Consulta os detalhes completos de uma demanda específica, incluindo suas subtarefas e responsáveis.
Inputs: demanda_id (required).
renovar_sessaoB
Renova a sessão de autenticação do Gestão de Tarefas utilizando as credenciais salvas ou novas credenciais informadas.
Inputs: email, password.
sincronizar_fila_offlineB
Força o envio de todas as demandas e subtarefas que foram geradas localmente enquanto desconectado da VPN/intranet.
Inputs: no named parameters.
verificar_status_conexaoA
Informa o estado da conexão com a API da intranet, a validade do token de autenticação e a quantidade de itens na fila offline pendente.
Inputs: no named parameters.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 14 tools
Most tools target distinct resources and actions, e.g. project context vs project listing, demand details vs active demand list. Some boundary ambiguity exists between atualizar_subtarefa and concluir_subtarefas, since both can change subtask statuses in bulk, but their primary intents are distinguishable.
All tools follow a consistent verb_noun pattern in Portuguese snake_case (obter_, listar_, criar_, atualizar_, concluir_, etc.). There are no mixed conventions or inconsistent verb styles.
14 tools is well within the appropriate range for a task-management MCP server. Each tool addresses a distinct workflow area: project context, demand lifecycle, subtask management, sprints, and offline/session handling.
The core demand/subtask lifecycle is covered: create, read, update, list, bulk conclude, and sprint association. Minor gaps exist such as no delete/archive operations and no all-demand listing beyond active demands, but these are workable.
Maintenance
Related MCP Connectors
Tasks, sprints, releases and docs for engineering teams with AI coding agents.
Task manager your agent can fully operate: boards, tasks, sprints, roles, worklogs, day planner.
Task and sprint coordination for humans and AI agents, with scoped API keys and full audit trail.
Track stories, organize sprints, and manage project workflows across your team
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceEnables AI agents to manage projects, epics, and tasks with atomic locking, real-time dashboard, and multi-agent coordination.MIT
- AlicenseNot gradedqualityDmaintenanceProvides Kanban, Gantt, list views, multi-project management, and archiving for task management via MCP protocol, enabling Cherry Studio Agent to create, update, query, and organize tasks.1 npmISC
- AlicenseCqualityDmaintenanceEnables AI agents to manage backlogs via a local API, supporting task and project operations with fail-fast resilience.226 npmMIT
- AlicenseBqualityCmaintenanceEnables AI agents to manage tasks through a board with backlog, ready, in-progress, and done columns, plan daily work, track velocity, forecast backlog completion, and leverage rule-based or DeepSeek-powered recommendations.221MIT