Jira MCP Server

Jira MCP Server

Enables AI assistants to manage Jira issues with full CRUD, transitions, linking, commenting, and file attachments via the Jira REST API v3.

Category
访问服务器

README

Jira MCP Server

MCP (Model Context Protocol) server for Jira Cloud. Provides AI assistants with tools to read, search, create, update, transition, link, comment, and attach files to Jira issues via the REST API v3. Text fields (descriptions, comments) support Markdown formatting.

Prerequisites

  • Node.js 18+
  • Jira Cloud instance
  • API Token for your Atlassian account (create one here)

Setup

npm install
npm run build

Create a .env file (see .env.example):

JIRA_BASE_URL=https://jira.example.com
JIRA_EMAIL=user@example.com
JIRA_API_TOKEN=your-api-token

Running as MCP Server

Add to your Cursor mcp.json:

{
  "mcpServers": {
    "jira": {
      "command": "node",
      "args": ["C:\\repos\\Custom-MCP\\Jira\\dist\\index.js"]
    }
  }
}

Or run directly for development (without building):

npm run dev

Tools

get_my_tasks

Returns active tasks assigned to the current user.

Parameter Type Default Description
max_results number 10 Max tasks to return (1-50)

JQL: assignee = currentUser() AND status NOT IN (Done, Cancelled) ORDER BY priority DESC

get_task_details

Returns full context for a single issue: summary, description, status, assignee, comments, attachments, and issue links.

Parameter Type Description
issue_key string Issue key, e.g. PROJ-123

Response includes issueLinks[] with id, linkType, direction, relation, and linked issue details. Use the link id with unlink_issues to remove a link.

search_tasks

Searches issues by text query or raw JQL. Provide either query or jql, not both.

Parameter Type Default Description
query string Text to search for (uses text ~ "...")
jql string Raw JQL query
max_results number 5 Max results (1-50)

update_task_status

Transitions an issue to a new status. Fetches available transitions first, matches by name (case-insensitive), then executes.

Parameter Type Description
issue_key string Issue key, e.g. PROJ-123
transition_name string Target transition, e.g. In Progress

create_task

Creates a new Jira issue. Description supports Markdown formatting.

Parameter Type Default Description
project string Project key (e.g. PROJ)
summary string Issue title
issue_type string Task Issue type (Task, Bug, Story, Epic ...)
description string Description (supports Markdown)
parent string Parent issue key (e.g. epic key)

update_task

Updates fields on an existing Jira issue. Only provided fields are changed. Description supports Markdown formatting.

Parameter Type Description
issue_key string Issue key, e.g. PROJ-123
summary string New issue title
description string New description (supports Markdown)
issue_type string New issue type (Story, Bug, Task ...)
parent string New parent issue key (e.g. epic key)

link_issues

Creates a link between two Jira issues. Use get_task_details to see existing links.

Parameter Type Description
link_type string Link type name (e.g. Blocks, Relates, Duplicate)
outward_issue string Issue key that gets the outward relation (e.g. the blocker)
inward_issue string Issue key that gets the inward relation (e.g. the blocked)

Common link types: Blocks (blocks / is blocked by), Relates (relates to), Duplicate (duplicates / is duplicated by), Cloners (clones / is cloned by).

unlink_issues

Removes a link between two issues by link ID. Get the link ID from get_task_details → issueLinks[].id.

Parameter Type Description
link_id string Issue link ID to delete

add_comment

Adds a comment to an issue. Supports Markdown formatting.

Parameter Type Description
issue_key string Issue key, e.g. PROJ-123
comment string Comment text (supports Markdown)

attach_file

Uploads a file from disk and attaches it to an issue. Optionally adds a comment in the same call.

Parameter Type Required Description
issue_key string yes Issue key, e.g. PROJ-123
file_path string yes Absolute path to the file on disk
comment string no Optional comment to add after attaching

CLI

A standalone CLI runner is included for manual testing outside of MCP. It uses the same .env config and the same Jira client.

npm run cli -- <tool_name> [json_args]

Examples (key=value — works everywhere)

The recommended way to pass arguments. No quoting issues on any OS or shell.

npm run cli -- --help
npm run cli -- get_my_tasks
npm run cli -- get_my_tasks max_results=5
npm run cli -- get_task_details issue_key=PROJ-123
npm run cli -- search_tasks query="oauth bug" max_results=5
npm run cli -- search_tasks jql="project = PROJ AND status = Open"
npm run cli -- create_task project=PROJ summary="Fix login bug" issue_type=Bug description="**Steps:** ..." parent=PROJ-100
npm run cli -- update_task issue_key=PROJ-123 summary="Updated title" description="New **description**"
npm run cli -- update_task_status issue_key=PROJ-123 transition_name="In Progress"
npm run cli -- add_comment issue_key=PROJ-123 comment="Fixed in PR #42"
npm run cli -- attach_file issue_key=PROJ-123 file_path=C:\path\to\report.pdf
npm run cli -- attach_file issue_key=PROJ-123 file_path=C:\path\to\report.pdf comment="See attached report"
npm run cli -- link_issues link_type=Blocks outward_issue=PROJ-1 inward_issue=PROJ-2
npm run cli -- unlink_issues link_id=12345

Examples (JSON — Bash / macOS / Linux)

JSON is also accepted as a single argument when wrapped in single quotes:

npm run cli -- get_task_details '{"issue_key": "PROJ-123"}'
npm run cli -- search_tasks '{"query": "oauth bug", "max_results": 5}'

Windows note: npm.cmd passes arguments through cmd.exe, which strips double quotes. Use the key=value format on Windows instead of JSON.

Output is formatted JSON printed to stdout. Errors print to stderr with a non-zero exit code.

Project Structure

src/
  index.ts           — MCP server entry point (stdio transport)
  cli.ts             — CLI entry point for manual testing
  config.ts          — .env loader, validates JIRA_BASE_URL / EMAIL / API_TOKEN
  adf.ts             — ADF ↔ Markdown/plain-text conversion
  jira-client.ts     — HTTP wrapper over native fetch with Basic Auth
  response.ts        — successResponse / errorResponse helpers
  tools/
    getMyTasks.ts
    getTaskDetails.ts
    searchTasks.ts
    updateTaskStatus.ts
    addComment.ts
    attachFile.ts
    createTask.ts
    updateTask.ts
    linkIssues.ts
    unlinkIssues.ts
docs/
  markdown-formatting.md  — Supported Markdown syntax reference

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选