mcpField
G
gitscrum-core

gitscrum-core

Local
github.com/gitscrum-coreProductivity5Updated 3w ago

About gitscrum-core

gitscrum-core is an MCP server connecting AI assistants to GitScrum workspaces for conversational project management. It covers tasks, sprints, time tracking, collaboration, and client workflows, with hosted SSE and local stdio access.

Key features

  1. 01Manage tasks, sprints, and projects
  2. 02Track time and view analytics
  3. 03Manage discussions, wiki, and notes
  4. 04Handle clients, invoices, and proposals
  5. 05Review budgets, standups, and activity reports

Use cases

  1. 01Find tasks due today
  2. 02Create a sprint and assign backlog items
  3. 03Check project budget consumption and risks

MCP support

It exposes GitScrum workspace operations to MCP-compatible clients. It can run as a hosted SSE endpoint or locally over stdio, and supports OAuth 2.0 Device Authorization.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
"mcpServers": {
"gitscrum-core": {
"command": "npx",
"args": [
"-y",
"@gitscrum-studio/mcp-server"
],
"env": {
"GITSCRUM_API_URL": "<GITSCRUM_API_URL>",
"GITSCRUM_TOKEN": "<GITSCRUM_TOKEN>"
}
}
}
}
Run commandnpx -y @gitscrum-studio/mcp-server
Environment variables to set
GITSCRUM_API_URLGitScrum API base URL
GITSCRUM_TOKENSecretGitScrum API authentication token

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npm@gitscrum-studio/mcp-server1.0.1

Tools

29

Saved tool list · updated 2w ago

This tool takes no parameters.

Details

From the project README

GitScrum

GitScrum MCP Server: Your AI Agent as a Project Manager 🚀

npm version MIT License Tests

Turn your AI Agent into a full-scale Project Manager.
The Official Model Context Protocol (MCP) server for GitScrum provides AI assistants (Claude, Cursor, Windsurf, Copilot) with deep context and operational access to your entire project stack.

GitScrum MCP Server

⚡ Hosted (Recommended)

Zero install. Works everywhere.

https://mcp.gitscrum.com/sse

SSE-based server hosted on AWS. Just add URL + token to your AI client.

Best for: Claude Desktop, Cursor, Windsurf, any SSE-compatible client.

→ Setup Guide

📦 Local (npm)

Self-hosted via npx.

npx -y @gitscrum-studio/mcp-server

Runs locally via stdio transport. Requires Node.js 18+.

Best for: VS Code, GitHub Copilot, offline environments.

→ Local Setup


Overview

GitScrum Studio MCP Server connects AI assistants to your GitScrum workspace via the Model Context Protocol. It gives Claude, GitHub Copilot, Cursor, and any MCP-compatible client full operational access to your project management stack — tasks, sprints, time tracking, user stories, epics, kanban workflows, team discussions, wiki, notes, client CRM, invoicing, proposals, budget tracking, analytics dashboards, standup reports, and activity feeds.

Everything your team does in the GitScrum web app, your AI assistant can now do through conversation.

You:    "What's on my plate today?"
Assistant: Fetches your tasks due today across all projects.

You:    "Create a sprint for next week with the top 5 backlog items"
Assistant: Creates the sprint, assigns tasks, and sets the timeline.

You:    "Show me which projects are over budget"
Assistant: Returns burn-down data and flags at-risk projects.

You:    "Send the Q1 proposal to Acme Corp"
Assistant: Creates the proposal, attaches the client, and sends it.

You:    "What did the team ship this week?"
Assistant: Generates a standup digest with completed work and blockers.

29 tools. 160+ operations. Zero context switching.


Quick start

Install
npx -y @gitscrum-studio/mcp-server
Configure your client
Claude Desktop

Edit the configuration file:

  • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
  • Windows: %APPDATA%\Claude\claude_desktop_config.json
{
  "mcpServers": {
    "gitscrum": {
      "command": "npx",
      "args": ["-y", "@gitscrum-studio/mcp-server"]
    }
  }
}
VS Code / Cursor

Add to .vscode/mcp.json or your MCP settings:

{
  "servers": {
    "gitscrum": {
      "command": "npx",
      "args": ["-y", "@gitscrum-studio/mcp-server"]
    }
  }
}
Authenticate

Tell your AI assistant: "Login to GitScrum"

The server initiates an OAuth 2.0 Device Authorization Grant flow. You authorize in the browser — credentials are never shared with the MCP server.


Tools

Each tool uses a consolidated action parameter, reducing LLM context tokens by ~80% compared to individual tool definitions.

Core
ToolActionsDocs
taskmy today get create update complete subtasks filter by_code duplicate move notificationstasks
sprintlist all get kpis create update stats reports progress metricssprints
workspacelist getprojects
projectlist get stats tasks workflows types efforts labels membersprojects
timeactive start stop logs analytics team reports productivity timelinetime-tracking
Planning
ToolActionsDocs
user_storylist get create update alluser-stories
epiclist create updateepics
labellist create update attach detach togglelabels
task_typelist create update assigntask-types
workflowcreate updateworkflows
Collaboration
ToolActionsDocs
discussionall channels channel messages send search unread mark_read create_channel update_channeldiscussions
commentlist add updatecomments
wikilist get create update searchwiki
notelist get create update share revisionsnotevault
note_folderlist create update movenotevault
search—search
ClientFlow CRM
ToolActionsDocs
clientlist get create update contacts interactions add_interactionclientflow
invoicelist get stats create update issue send mark_paidclientflow
proposallist get stats create update send approve reject convertclientflow
clientflow_dashboard8 reportsclientflow
clientflow_cross_workspace4 reportsclientflow
Insights PRO
ToolActionsDocs
standupsummary completed blockers team stuck digest contributorsstandup
analytics10 reportsanalytics
activityfeed user_feed notifications activities task_workflowactivity
budgetprojects_at_risk overview consumption burn_down alerts eventsbudget
Authentication
ToolDescriptionDocs
auth_loginInitiate device code flowauth
auth_completeComplete authorizationauth
auth_statusCheck session statusauth
auth_logoutClear stored credentialsauth

Full reference: docs/TOOLS.md


Security

The server is designed around the principle of least privilege.

LayerProtection
OperationsOnly CREATE, READ, UPDATE. DELETE is blocked at MCP and API layers.
AuthenticationOAuth 2.0 Device Grant — credentials never touch the server.
Token storageLocal filesystem with restricted permissions.
Rate limitingAutomatic lockout after failed auth attempts.

Destructive operations must be performed in the GitScrum Studio.

Full details: docs/SECURITY.md

Found a vulnerability? Report privately to [email protected].


Documentation

Hosted ServerSSE setup for Claude, Cursor, Windsurf, and more
Usage GuidePractical examples and common workflows
Tools ReferenceAll 29 tools with parameters and response shapes
Per-Tool GuidesDeep-dive into each tool module
SecuritySecurity model, token handling, threat mitigations
DevelopmentLocal setup, architecture, testing, contribution
ChangelogVersion history and migration notes

Development

git clone https://github.com/gitscrum-core/mcp-server.git
cd mcp-server
npm install
npm run build
npm test          # 378 tests across 22 suites

Inspect locally with the MCP Inspector:

npx @modelcontextprotocol/inspector node dist/index.js
RequirementVersion
Node.js>= 18.0.0
npm>= 8.0.0

Full guide: docs/DEVELOPMENT.md


Contributing

See CONTRIBUTING.md for guidelines.

git checkout -b feature/my-feature
# make changes, add tests
npm test
git commit -m "feat: describe your change"

License

MIT — see LICENSE.


Website  ·  Docs  ·  Issues  ·  Changelog

Built by GitScrum

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.