Sprint Analyzer for Documentation Teams

Automate Jira sprint review and documentation impact analysis using Claude Code.

Time Savings: 45-60 minutes → 15 seconds (98% reduction)

Demo

[Link to demo video - shows this tool in action]

What It Does

• Queries Jira for sprint and backlog tickets via natural language
• Automatically filters out internal work (tests, spikes, CI/CD)
• Categorizes tickets by documentation impact (HIGH/MEDIUM/LOW)
• Generates actionable reports with direct Jira links
• All through simple questions in Claude Code

Requirements

• Claude Code - CLI, desktop app, or web (claude.ai/code)
• Atlassian MCP Server - For Jira integration (setup guide)
• Python 3.x - To run the analyzer script
• Jira Instance - With sprint-based workflow

Quick Start

1. Clone and Configure

git clone https://github.com/EMcWhinn/sprint-analyzer
cd sprint-analyzer

# Copy template and customize for your project
cp CLAUDE.md.template CLAUDE.md

2. Edit CLAUDE.md

Open CLAUDE.md and replace the placeholders:

• [YOUR PRODUCT NAME] - Your product/project name
• [YOUR DOCUMENTATION URL] - Where your docs live
• [YOUR_PROJECT] - Your Jira project key (e.g., SHOP, DOCS, MYPROJ)
• [YOUR_TEAM_FILTER] - Your Jira team/workstream field (e.g., Team = "Backend")
• [YOUR_COMPONENT] - Your product components

Example JQL customization:

# From template:
project = [YOUR_PROJECT] AND [YOUR_TEAM_FILTER] AND sprint in openSprints()

# For your project:
project = SHOP AND Team = "Backend" AND sprint in openSprints()

See inline comments in CLAUDE.md.template for guidance.

3. Set Up Atlassian MCP

Configure the Atlassian MCP server with your Jira credentials. This enables Claude Code to query your Jira instance.

4. Use It

Open Claude Code in this directory and ask:

Analyze [your project] sprints for documentation impact

Claude will automatically query Jira, run the analyzer, and show categorized results.

How It Works

User asks in natural language
    ↓
Claude Code interprets intent
    ↓
Atlassian MCP queries Jira
    ↓
Data converted for analyzer
    ↓
doc_sprint_analyzer.py categorizes tickets
    ↓
Report displayed (HIGH/MEDIUM/LOW impact)

Impact Categories

HIGH Impact (action required):

• Breaking changes, deprecations, migrations
• API/endpoint removals
• Explicit documentation requirements

MEDIUM Impact (review recommended):

• New user-facing features
• Security vulnerabilities (CVEs)
• Customer-facing bugs
• Workflow/behavior changes

LOW Impact (monitor):

• Minor enhancements
• Component updates

Automatically Filtered (not shown):

• Test tickets, spikes, CI/CD work
• Internal refactoring, dev environment setup

Customizing Keywords

Want different categorization? Edit doc_sprint_analyzer.py:

Add your product-specific terms (around line 109-148):

# High priority signals
high_impact_signals = [
    'breaking change',
    'incompatible with',     # Add yours
]

# Medium priority signals
medium_impact_signals = [
    'new feature',
    'merchant-facing',       # Add yours
]

Change default message (around line 210):

return {
    'impact': 'LOW',
    'reason': 'Review recommended'  # Customize
}

Example Queries

Once configured, ask Claude Code questions in natural language:

Sprint Analysis:

• Analyze Platform Services sprints for documentation impact
• Analyze [your team name] sprints for documentation impact
• What breaking changes are in the next sprint?
• Compare this sprint to last sprint

Component/Label-Specific:

• Analyze aap-gateway label for documentation impact
• Analyze the API component backlog
• Show me [your component] tickets

Filtered Views:

• Show me just the CVEs from the current sprint
• What customer-facing bugs are in the backlog?
• List all breaking changes

The tool automatically:

• Constructs the correct JQL query based on your CLAUDE.md configuration
• Queries your Jira instance
• Filters and categorizes results
• Presents actionable reports

Current Example Configuration

This repo includes an example configuration for Red Hat Ansible Automation Platform as a reference:

• Jira Project: AAP
• Workstream: Platform Services
• Components: aap-gateway, platform-services-utilities
• Docs: docs.redhat.com/ansible-automation-platform

You can use this to understand the structure when configuring for your own project.

Different Jira Structures?

Your CLAUDE.md template needs the right JQL queries for your Jira setup. Here are examples:

Sprint analysis (workstream/team-based):

# Platform Services example (current config)
project = AAP AND Workstream = "Platform Services" AND sprint in openSprints()

# Simple team filter
project = DOCS AND Team = "Backend" AND sprint in openSprints()

# Multiple teams
project = SHOP AND Team in ("Checkout", "Payments") AND sprint in openSprints()

# No custom fields (just use sprint)
project = MYPROJ AND sprint in openSprints()

Component/label analysis:

# By component (like aap-gateway)
project = AAP AND component = "aap-gateway" AND status in (Backlog, Refinement, New, "In Progress")

# By label
project = MYPROJ AND labels = "api-changes" AND status in (Open, "In Progress")

# Multiple components
project = SHOP AND component in ("Checkout", "Payments") AND status in (Backlog, "In Progress")

Use your Jira query builder to find the right field names for your instance.

Troubleshooting

No results?

• Check your JQL query in CLAUDE.md matches your Jira structure
• Verify Atlassian MCP is configured correctly
• Make sure you're in the project directory when running Claude Code

Wrong categorization?

• Add product-specific keywords to doc_sprint_analyzer.py
• Adjust impact signals for your workflow

Too many results?

• Add more filters to your JQL (team, component, status)
• Customize what gets filtered in doc_sprint_analyzer.py (line 39-95)

Files

• README.md - This file
• doc_sprint_analyzer.py - Core analyzer (works out of the box)
• CLAUDE.md.template - Template project instructions (customize this)
• .gitignore - Prevents committing your personal CLAUDE.md

Contributing

Improvements welcome! Common additions:

• More pattern matching keywords
• GitHub/GitLab issue tracker support
• Additional filtering logic

Questions?

Open an issue or see the demo video for a walkthrough.

---

Note: This tool uses pattern matching to pre-filter tickets. It saves significant time but doesn't replace human review—always verify the results make sense for your context.