TaskMesh MCP Server
OfficialClick 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., "@TaskMesh MCP Server list my current sprint tasks"
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.
🔌 TaskMesh MCP Server
The TaskMesh Model Context Protocol (MCP) Server connects your AI coding assistant (Cursor, Claude Desktop, Antigravity, VS Code, Windsurf) directly to your TaskMesh workspace.
It allows AI agents to read tasks, log time, start stopwatches, manage sprints, and update ticket statuses in 1 shot without context bloat or token waste.
⚡ Quick Start (3 Steps)
1. Build the MCP Server
cd mcp
pnpm install
pnpm run build2. Generate an API Token
Open TaskMesh in your browser.
Go to Settings ➔ Developer Settings (or call
POST /auth/tokens).Create a token (e.g.
Cursor IDEwith*full scope).Copy your token (starts with
tm_...).
3. Add to your AI Editor
Defaults to production (
https://crm.bluentech.com.np). If developing locally, you can setTASKMESH_API_URL=http://localhost:3001.
🎯 Cursor Setup
Open Cursor Settings (⚙️) ➔ Features ➔ MCP.
Click + Add New MCP Server:
Name:
taskmeshType:
commandCommand:
npx -y @task-mesh/mcp-server@latestEnvironment Variables:
TASKMESH_API_URL=https://crm.bluentech.com.np TASKMESH_API_TOKEN=tm_your_copied_token_here
🟣 Claude Desktop Setup
Add to your claude_desktop_config.json:
{
"mcpServers": {
"taskmesh": {
"command": "npx",
"args": ["-y", "@task-mesh/mcp-server@latest"],
"env": {
"TASKMESH_API_URL": "https://crm.bluentech.com.np",
"TASKMESH_API_TOKEN": "tm_your_copied_token_here"
}
}
}
}🚀 Antigravity / Windsurf / Custom MCP Clients
{
"mcpServers": {
"taskmesh": {
"command": "npx",
"args": ["-y", "@task-mesh/mcp-server@latest"],
"env": {
"TASKMESH_API_URL": "https://crm.bluentech.com.np",
"TASKMESH_API_TOKEN": "tm_your_copied_token_here"
}
}
}
}Related MCP server: projectflow-mcp
🤖 AI Assistant Instructions (AGENTS.md / .cursorrules)
To make your AI agent interact with TaskMesh instantly without wasting round trips, paste this prompt into your project's AGENTS.md, CLAUDE.md, or .cursorrules:
### TaskMesh MCP Guidelines for AI Coding Agents
When managing tasks, logging time, or interacting with TaskMesh:
1. **Zero-Hop Lookups**: NEVER make discovery loops (do NOT call `list_projects` then `list_issues` to find an issue).
- If the user gives a ticket key (e.g., `ALPHA-14`), use `get_issue({ identifier: "ALPHA-14" })` or `start_working_on_issue({ identifier: "ALPHA-14" })` directly.
- If looking for a task by keyword, use `search_issues({ query: "keyword" })`.
2. **Session Bootstrapping**: At the beginning of a work session, call `get_agent_context()` to get active projects, current sprint, assigned issues, and running timers in ~150 tokens.
3. **Starting Work**: When beginning a task, use `start_working_on_issue({ identifier: "KEY" })`. This atomically updates the status to "In Progress", starts the stopwatch timer, and fetches task requirements in 1 shot.
4. **Updating Status**: Use human status names (e.g. `status: "In Progress"`, `status: "Done"`, `status: "To Do"`). The MCP server resolves the internal workflow status automatically.
5. **Logging Time**: When stopping work, call `stop_timer()` or `log_time({ identifier: "ALPHA-14", hours: 1.5, description: "Implemented feature" })`.🛠️ Tool Reference
🧠 Agent Context & Zero-Hop Workflow
Tool | Description | Example Input |
| 1-shot snapshot of active projects, sprint, my issues, and timer (~150 tokens) |
|
| Moves issue to In Progress, starts live timer, returns details in 1 call |
|
| Org-wide instant search across all projects |
|
📝 Issue Tracking
Tool | Description | Example Input |
| Get clean Markdown details of an issue |
|
| Compact Markdown table of issues |
|
| Create issue with project key & status name |
|
| Update status, assignee, or estimate |
|
| Post a discussion comment |
|
| Soft delete an issue |
|
⏱️ Time Tracking
Tool | Description | Example Input |
| View currently running stopwatch timer |
|
| Start live timer on an issue |
|
| Stop running timer & record worklog |
|
| Manually log hours to an issue |
|
| View logged timesheets |
|
📁 Projects & Sprints
Tool | Description | Example Input |
| Overview table of all projects |
|
| Get project workflow statuses & lead |
|
| Create new project workspace |
|
| List sprints for a project |
|
| Mark a sprint as active |
|
| Complete an active sprint |
|
🔒 Security & Architecture
Token-Bucket Rate Limiting: Protects your API from runaway AI loops.
Granular Scopes: Support for wildcards (
*) or read-only/write-only scopes (issues:read,time:write).Clean Audit Logs: Structured JSON audit logs streamed to
stderrwithout interfering with stdio transport.
Available Tools
24 toolsadd_commentB
Add a discussion comment to an issue
Inputs: body (required), issueId (required).
add_workflow_statusB
Add a new workflow status to a project
Inputs: name (required), projectId (required), color, category, +2 more.
complete_sprintA
Complete an active sprint and optionally roll over incomplete issues
Inputs: id (required), rolloverSprintId.
create_issueC
Create a new issue inside a project
Inputs: title (required), projectId (required), type, dueDate, +4 more.
create_projectC
Create a new project
Inputs: key (required), name (required), leadId (required), deptId, +4 more.
create_sprintB
Create a new sprint
Inputs: name (required), projectId (required), goal, endDate, +1 more.
delete_issueC
Delete an issue
Inputs: id (required).
delete_projectB
Delete a project
Inputs: id (required).
delete_sprintB
Delete a sprint
Inputs: id (required).
get_active_timerB
Get currently running timer for the authenticated user
Inputs: no named parameters.
get_issueB
Get detailed information of a specific issue by ID
Inputs: id (required).
get_projectA
Get detailed information of a specific project by ID
Inputs: id (required).
get_sprintB
Get detailed information of a specific sprint by ID
Inputs: id (required).
list_issuesB
List issues with optional filters
Inputs: projectId (required), type, search, priority, +3 more.
list_projectsA
List all projects in the organization
Inputs: no named parameters.
list_sprintsA
List all sprints for a project
Inputs: projectId (required).
list_time_entriesC
List logged time entries with filters
Inputs: userId, endDate, issueId, projectId, +1 more.
log_timeC
Log time spent on an issue manually
Inputs: date (required), hours (required), projectId (required), issueId, +2 more.
start_sprintB
Start a planned sprint
Inputs: id (required).
start_timerB
Start a running stopwatch timer on a specific issue
Inputs: issueId (required).
stop_timerA
Stop the active running timer and save it as a time entry
Inputs: no named parameters.
update_issueC
Update an existing issue status, assignee, estimate, or metadata
Inputs: id (required), title, dueDate, estimate, +5 more.
update_projectB
Update an existing project metadata or status
Inputs: id (required), name, leadId, status, +3 more.
update_sprintC
Update sprint details
Inputs: id (required), goal, name, status, +2 more.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 24 tools
All tools have clearly distinct purposes. Issues, sprints, projects, timers, time entries, comments, and workflow statuses are separate domains with no overlap.
All tools follow a consistent verb_noun pattern (e.g., create_issue, list_projects, start_timer), all in snake_case, making it easy for an agent to predict tool names.
24 tools is slightly above the ideal range but still well-proportioned for a project management server covering multiple entities (issues, sprints, projects, timers, etc.), each earning its place.
Core CRUD operations are covered for projects, issues, and sprints. Timer and time entry management are included. Minor gaps like deleting comments or advanced issue transitions are missing but do not severely hamper workflows.
Maintenance
Related MCP Connectors
Task management for people and AI agents, with scoped OAuth access to issues, projects, and docs.
Task management for people and AI agents, with scoped OAuth access to issues, projects, and docs.
Connects AI assistants to TrackingTime for querying time entries, projects, tasks, customers, and team data using natural language.
Task and sprint coordination for humans and AI agents, with scoped API keys and full audit trail.
Related MCP Servers
- FlicenseNot gradedqualityCmaintenanceEnables AI tools to create and update tasks on tone, a team task management service for humans and AI.2 npm-
- AlicenseNot gradedqualityCmaintenanceEnables AI coding agents to securely read, update, and create tasks on ProjectFlow boards.MIT
- AlicenseAqualityBmaintenanceEnables AI assistants to read and manage Targetprocess work items through the Targetprocess REST API.5840 npm1MIT
- AlicenseNot gradedqualityAmaintenanceEnables AI assistants and IDEs to connect with Taskade workspaces, projects, tasks, AI agents, automations, and files through natural language, using either a hosted OAuth server or a local stdio package.2,530 npmMIT