jira（Jira 项目管理）

JIRA Management Skill

A comprehensive Claude Code skill for managing JIRA issues, projects, and workflows using the Atlassian MCP server.

Table of Contents

• Overview
• Prerequisites
• Skill Workflow

- Issue Creation - Issue Search and Management - Workflow and Transitions - Agile/Scrum Operations - Linking and Relationships - Comments and Collaboration - Batch Operations - Project and Version Management

• Best Practices
• Common Use Cases
• References
• Troubleshooting

Overview

This skill provides intelligent JIRA management capabilities including:

• Creating and managing issues with proper field validation
• Searching and filtering issues using JQL
• Managing workflows and transitions
• Working with epics, sprints, and agile boards
• Adding comments, attachments, and links
• Batch operations for efficiency
• Project and version management

Prerequisites

Required MCP Server

• Atlassian MCP (mcp__atlassian) must be configured in Claude Code
• JIRA credentials (API token or OAuth) must be set up
• Appropriate JIRA permissions for the operations you need

Environment Configuration

The Atlassian MCP may use environment variables:

• JIRA_URL: Your JIRA instance URL
• JIRA_API_TOKEN: API token for authentication
• JIRA_EMAIL: Email associated with API token
• JIRA_PROJECTS_FILTER: (Optional) Comma-separated project keys to filter

Skill Workflow

1. Issue Creation Workflow

When creating JIRA issues, follow this sequence:

Step 1: Gather Requirements

Ask the user for:

• Project key (e.g., "PROJ", "DEV", "SUPPORT") - NEVER ASSUME
• Issue type (Task, Bug, Story, Epic, Subtask)
• Summary (title/description)
• Priority (if applicable)
• Assignee (optional - email, display name, or account ID)
• Additional fields (components, labels, custom fields)

Step 2: Validate Project

Use mcp__atlassian__jira_get_all_projects to:

• Verify project exists
• Get available projects if user unsure
• Confirm project key matches exactly

Step 3: Search Available Fields (if needed)

Use mcp__atlassian__jira_search_fields to find custom field names and IDs. See Custom Field Discovery for detailed methodology.

Step 4: Create Issue

Use mcp__atlassian__jira_create_issue with:

{
  "project_key": "PROJ",
  "summary": "Implement user authentication",
  "issue_type": "Task",
  "description": "Detailed description in Markdown format",
  "assignee": "user@example.com",
  "components": "Frontend,API",
  "additional_fields": {
    "priority": {"name": "High"},
    "labels": ["security", "authentication"],
    "parent": "PROJ-123"
  }
}

Step 5: Follow-up Actions (if needed)

• Link to epic: mcp__atlassian__jira_link_to_epic
• Add attachments: Include in update or create
• Add comments: mcp__atlassian__jira_add_comment
• Create issue links: mcp__atlassian__jira_create_issue_link

2. Issue Search and Management

Searching Issues

Use mcp__atlassian__jira_search with JQL. For comprehensive JQL documentation, see JQL Guide.

Essential JQL Patterns:

# Open issues in project
project = PROJ AND status = Open

# Your assigned issues
assignee = currentUser() AND status != Done

# Issues in current sprint
sprint IN openSprints()

# Recently updated
updated >= -7d AND project = PROJ

Parameters:

• jql: JQL query string
• fields: "summary,status,assignee,priority" or "*all"
• limit: Max results (1-50, default 10)
• start_at: Pagination offset

Getting Issue Details

Use mcp__atlassian__jira_get_issue:

• issue_key: "PROJ-123"
• fields: Comma-separated or "*all"
• expand: "renderedFields", "transitions", "changelog"
• comment_limit: Number of comments to include

Updating Issues

Use mcp__atlassian__jira_update_issue:

{
  "issue_key": "PROJ-123",
  "fields": {
    "summary": "Updated summary",
    "assignee": "user@example.com",
    "priority": {"name": "Critical"}
  },
  "additional_fields": {
    "labels": ["urgent", "hotfix"]
  },
  "attachments": "/path/to/file1.txt,/path/to/file2.pdf"
}

3. Workflow and Transitions

Get Available Transitions

Use mcp__atlassian__jira_get_transitions:

• Returns list of valid transitions for current issue state
• Each transition has an ID and name

Transition Issue

Use mcp__atlassian__jira_transition_issue:

{
  "issue_key": "PROJ-123",
  "transition_id": "31",
  "fields": {
    "resolution": {"name": "Fixed"}
  },
  "comment": "Moving to Done after completing implementation"
}

4. Agile/Scrum Operations

Working with Boards

1. Get boards: mcp__atlassian__jira_get_agile_boards

- Filter by: board_name, project_key, board_type (scrum/kanban)

1. Get board issues: mcp__atlassian__jira_get_board_issues

- Requires board_id and jql filter

Working with Sprints

1. Get sprints: mcp__atlassian__jira_get_sprints_from_board

- Filter by state: "active", "future", "closed"

1. Get sprint issues: mcp__atlassian__jira_get_sprint_issues

- Returns all issues in specified sprint

1. Create sprint: mcp__atlassian__jira_create_sprint

{
  "board_id": "1000",
  "sprint_name": "Sprint 15",
  "start_date": "2025-01-21T09:00:00.000+0000",
  "end_date": "2025-02-04T17:00:00.000+0000",
  "goal": "Complete authentication feature"
}

1. Update sprint: mcp__atlassian__jira_update_sprint

Working with Epics

1. Find epics: Use JQL: issuetype = Epic AND project = PROJ
2. Link to epic: mcp__atlassian__jira_link_to_epic
3. Find issues in epic: Use JQL: parent = EPIC-KEY

5. Linking and Relationships

Issue Links

Use mcp__atlassian__jira_create_issue_link:

{
  "link_type": "Blocks",
  "inward_issue_key": "PROJ-123",
  "outward_issue_key": "PROJ-456",
  "comment": "This issue blocks the other"
}

Get link types: mcp__atlassian__jira_get_link_types

Remote Links (Web/Confluence)

Use mcp__atlassian__jira_create_remote_issue_link:

{
  "issue_key": "PROJ-123",
  "url": "https://confluence.example.com/pages/123456",
  "title": "Technical Design Doc",
  "summary": "Detailed architecture documentation",
  "relationship": "documentation"
}

6. Comments and Collaboration

Add Comment

Use mcp__atlassian__jira_add_comment:

• Supports Markdown format
• Visible in issue activity

Get Comments

Use mcp__atlassian__jira_get_issue with appropriate fields

7. Batch Operations

Batch Create Issues

Use mcp__atlassian__jira_batch_create_issues:

{
  "issues": "[{\"project_key\":\"PROJ\",\"summary\":\"Task 1\",\"issue_type\":\"Task\"},{\"project_key\":\"PROJ\",\"summary\":\"Task 2\",\"issue_type\":\"Bug\"}]",
  "validate_only": false
}

Batch Get Changelogs

Use mcp__atlassian__jira_batch_get_changelogs:

• Get history for multiple issues
• Filter by specific fields
• Cloud only feature

8. Project and Version Management

List Projects

Use mcp__atlassian__jira_get_all_projects:

• Returns accessible projects
• Respects JIRA_PROJECTS_FILTER if configured

Get Project Issues

Use mcp__atlassian__jira_get_project_issues:

• Returns all issues for specific project
• Supports pagination

Version Management

1. Get versions: mcp__atlassian__jira_get_project_versions
2. Create version: mcp__atlassian__jira_create_version
3. Batch create: mcp__atlassian__jira_batch_create_versions

Best Practices

1. Always Validate Input

• Never assume project keys - always verify
• Use jira_search_fields to find custom field IDs
• Check available transitions before transitioning

2. Use Appropriate Field Selection

• Request only needed fields to reduce token usage
• Use "*all" sparingly, only when necessary
• Specify fields explicitly for better performance

3. JQL Query Construction

See JQL Guide for comprehensive documentation.

4. Error Handling

• Check for required fields before creating issues
• Validate transition IDs before executing
• Handle permission errors gracefully

5. Efficiency

• Use batch operations for multiple issues
• Paginate large result sets
• Use JQL filters to reduce result size

Common Use Cases

Create Story with Subtasks

1. Create Epic (if needed)
2. Create Story and link to Epic
3. Create multiple Subtasks with parent = Story key
4. Optionally add to sprint

Bug Triage Workflow

1. Search for bugs: issuetype = Bug AND status = Open
2. For each bug: Update priority, assign, add to sprint, transition

Sprint Planning

1. Get active sprint
2. Search backlog issues
3. Move selected issues to sprint
4. Update estimates and assign

Release Management

1. Create version for release
2. Search issues: fixVersion = "1.0.0"
3. Verify all are Done
4. Create release notes

References

See references/ directory for detailed documentation:

• jql_guide.md - Comprehensive JQL query guide
• custom_field_discovery.md - Custom field discovery methodology

See templates/ directory for:

• issue_creation.json - Standard issue creation templates

Troubleshooting

Common Issues

Issue: "Project not found"

• Verify project key is correct (case-sensitive)
• Use jira_get_all_projects to list available projects
• Check JIRA_PROJECTS_FILTER environment variable

Issue: "Field required but not provided"

• Use jira_search_fields to find required fields
• Check project configuration for mandatory fields
• Some fields may be required by workflow

Issue: "Invalid transition"

• Use jira_get_transitions to see available transitions
• Check current issue status
• Verify permissions for transition

Issue: "Custom field not found"

• See Custom Field Discovery for discovery methodology
• Format: customfield_10010
• Check if field applies to issue type

Skill Invocation

This skill is automatically invoked when users:

• Ask to "create a JIRA ticket/issue"
• Request "search JIRA for..."
• Say "update JIRA issue..."
• Request "move issue to..." or "transition..."
• Ask about "sprint", "epic", or "board" operations
• Request batch operations on JIRA issues

Notes

• The Atlassian MCP (mcp__atlassian) prefix is used for all tools
• All date/time values use ISO 8601 format
• JQL syntax is similar to SQL but has specific JIRA operators
• Cloud vs Server/Data Center may have feature differences