Atlassian MCP Server
# Atlassian MCP Server
MCP server that lets AI assistants interact with Jira through the Atlassian CLI.
## Quick Start
See [QUICKSTART.md](QUICKSTART.md) for installation and setup.
## Features
✅ Search issues with JQL queries
✅ View full issue details
✅ Add comments
✅ Update status/transitions
✅ Assign to users
✅ Manage labels and custom fields
✅ Link related issues
## How It Works
The MCP server calls the Atlassian CLI (`acli`) commands:
1. AI assistant calls an MCP tool (e.g., `jira_search`)
2. MCP server executes corresponding `acli` command
3. Results are parsed and returned to the AI
## Prerequisites
- Node.js 18+
- Atlassian CLI: `winget install Atlassian.AtlassianCli`
- Authenticated Jira account: `acli jira auth login`
## Available Tools
| Tool | Description |
|------|-------------|
| `jira_search` | Search with JQL queries |
| `jira_get_issue` | Get issue details |
| `jira_comment` | Add comments |
| `jira_update_status` | Change status |
| `jira_assign` | Assign to user |
| `jira_set_field` | Update custom fields |
| `jira_add_label` | Add labels |
| `jira_remove_label` | Remove labels |
| `jira_link_issue` | Link issues |
## Usage Example
Ask your AI assistant:
- "Find my open work items"
- "Add a comment to PROJ-123"
- "Show high priority issues"
- "Move PROJ-456 to In Progress"
The AI will automatically use the appropriate MCP tools.
## Development
```powershell
npm install # Install dependencies
npm run build # Build TypeScript
npm run watch # Auto-rebuild on changes
npm start # Run the server
```
## Links
- [Quick Start Guide](QUICKSTART.md)
- [MCP Documentation](https://modelcontextprotocol.io)
- [Atlassian CLI](https://github.com/atlassian-labs/atlassian-cli)
TDQS
Scored across 9 tools
Every tool has a clearly distinct purpose targeting specific Jira operations with no overlap. For example, jira_add_label and jira_remove_label are complementary but distinct, while jira_get_issue, jira_search, and jira_update_status each handle different aspects of issue management. The descriptions reinforce these boundaries, making misselection unlikely.
All tools follow a consistent 'jira_verb_noun' pattern using snake_case throughout. The verbs are specific and descriptive (e.g., add, assign, comment, get, link, remove, search, set, update), creating a predictable and readable naming convention across the entire set.
With 9 tools, the server is well-scoped for Jira operations, covering core workflows without bloat. Each tool earns its place by addressing a specific need in issue management, from retrieval and search to updates and relationships, making the count appropriate for the domain.
The toolset provides strong coverage for common Jira workflows, including CRUD-like operations (get, search, update status, set fields), collaboration (comment, assign), and organization (labels, links). A minor gap exists in missing explicit create or delete issue tools, but agents can work around this using jira_set_field or other methods for creation, and the surface otherwise supports most agent tasks effectively.