GitHub & Azure DevOps Activity Summary

Contents

• [Overview](#overview)
• [Tracked Repositories](#tracked-repositories)
• [Workflow](#workflow)
• [Output Format](#output-format)
• [Dashboard](#dashboard) - Interactive web dashboard
• [Natural Language Parsing](#natural-language-parsing)
• [Automation](#automation) - Scheduled email reports
• [Critical Rules](#critical-rules)
• [Configuration](#configuration)
• [Examples](EXAMPLES.md) - Common usage scenarios
• [Automation Details](AUTOMATION.md) - Email delivery setup

Overview

This skill provides a consolidated view of team activity across multiple repositories in GitHub and Azure DevOps. It answers questions like:

• Who's been active this week?
• What PRs are open/merged?
• What commits have been made?
• What's the team working on?

User Experience Goals

Tracked Repositories

GitHub Repositories (ECI-Global org)

Azure DevOps Repositories

Workflow

Step 1: Determine Time Range

Parse user request for time range. Defaults:

• "today" → last 24 hours
• "this week" → last 7 days
• "this month" → last 30 days
• No time specified → last 7 days

Step 2: Gather GitHub Activity

For each tracked repository pattern, use the gh CLI:

List matching repos:

gh repo list ECI-Global --json name,nameWithOwner --limit 100 | jq '[.[] | select(.name | test("one-cloud|firehydrant|core-logics|iac|coralogix"; "i"))]'

Get commits (last 7 days):

gh api repos/{owner}/{repo}/commits --jq '.[] | {sha: .sha[0:7], author: .commit.author.name, date: .commit.author.date, message: .commit.message | split("\n")[0]}'

Get pull requests:

gh api repos/{owner}/{repo}/pulls?state=all --jq '.[] | {number, title, user: .user.login, state, created_at, merged_at}'

Step 3: Gather Azure DevOps Activity

For the Cloud-Delivery/spruce repo:

Get commits:

az repos commit list --organization https://dev.azure.com/Cloud-Delivery --project spruce --repository spruce --top 50 --output json

Get pull requests:

az repos pr list --organization https://dev.azure.com/Cloud-Delivery --project spruce --repository spruce --status all --top 50 --output json

Step 4: Aggregate by Team Member

Group all activity by author/creator:

• Commits authored
• PRs opened
• PRs merged
• PRs currently open

Step 5: Generate Summary

Create a consolidated report organized by:

1. Team Member Summary - Who did what
2. Repository Summary - Activity per repo
3. Open PRs Needing Attention - PRs awaiting review/merge

Output Format

Team Activity Summary

## Team Activity (Last 7 Days)

### By Team Member

**@alice**
- 12 commits across 3 repos
- 2 PRs merged (one-cloud-api, core-logics)
- 1 PR open: #45 "Add caching layer" (one-cloud-api)

**@bob**
- 8 commits across 2 repos
- 1 PR merged (iac)
- 0 PRs open

**@carol**
- 5 commits (spruce - Azure DevOps)
- 1 PR open: !958 "Feature update" (spruce)

### By Repository

| Repository | Commits | PRs Merged | PRs Open |
|------------|---------|------------|----------|
| one-cloud-api | 15 | 3 | 2 |
| core-logics | 8 | 1 | 0 |
| fire-hydrant | 3 | 0 | 1 |
| iac | 6 | 2 | 0 |
| spruce (ADO) | 5 | 1 | 1 |

### Open PRs Needing Attention

| Repo | PR | Title | Author | Age |
|------|-----|-------|--------|-----|
| one-cloud-api | #45 | Add caching layer | @alice | 3d |
| fire-hydrant | #12 | Fix auth flow | @dave | 5d |
| spruce | !958 | Feature update | @carol | 2d |

Dashboard

An interactive web dashboard is available for visual exploration of team activity.

Quick Start

cd .claude/skills/github-activity-summary/dashboard
npm install
npm run dev

Then open http://localhost:3456

Features

• Metric Cards: Commits, PRs merged, PRs open, contributors at a glance
• Charts: Bar and pie charts showing commits by contributor and repo
• Contributor Cards: Individual activity with avatar and stats
• PR Alerts: Color-coded open PRs by status (stale, needs review, in review)
• Commit Timeline: Recent commits with author and message
• Real-time Data: API fetches live data from GitHub
• Team Health Analysis: Sentiment, blockers, achievements from commit messages
• Work Pattern Detection: Identifies late-night and weekend commits

Commit Message Analysis

The dashboard analyzes FULL commit messages (not just the summary line) to extract deeper insights:

Sentiment Analysis

Detects the emotional tone of commits based on keywords:

• Positive: completed, shipped, implemented, fixed, improved
• Negative: blocked, stuck, broken, workaround, WIP, TODO
• Neutral: updated, changed, refactored, moved

Conventional Commits Support

Parses conventional commit format (type(scope): description):

• feat: New features
• fix: Bug fixes
• refactor: Code refactoring
• perf: Performance improvements
• docs: Documentation changes
• test: Testing additions
• chore: Maintenance tasks

Blocker Detection

Identifies commits indicating blockers:

• blocked, waiting on, pending
• WIP, work in progress
• needs review, needs approval
• TODO, FIXME, XXX
• workaround, temporary

Achievement Detection

Recognizes completed work:

• completed, finished, shipped
• launched, resolved
• implemented, integrated

Work Pattern Analysis

Tracks developer work patterns:

• Late Night Commits: 9pm - 6am
• Weekend Commits: Saturday & Sunday
• Burnout Risk: High (>40% off-hours), Medium (>20%), Low

Breaking Changes

Detects breaking changes via:

• ! after type: feat!: breaking change
• BREAKING CHANGE: in commit body

Component Catalog

The dashboard uses a json-render style component catalog. Claude can generate custom dashboard layouts by outputting JSON that maps to these components:

Ask Claude: "Show me a pie chart of commits by contributor" or "Create a dashboard focused on open PRs"

Natural Language Parsing

Automation

For automated daily email reports, see [AUTOMATION.md](AUTOMATION.md).

Quick setup:

1. Create a GitHub Actions workflow that runs the report script daily
2. Use SendGrid/Mailgun/SES to email the generated report
3. Schedule for 8am in your timezone

The automation script at scripts/generate-report.sh produces an HTML email-ready report.

Critical Rules

Data Freshness

• Always fetch live data - never cache or assume
• Report the exact time range queried
• Note any repos that failed to query

Authentication

• GitHub: Requires gh CLI authenticated
• Azure DevOps: Requires az CLI authenticated with appropriate permissions

Error Handling

• If a repo is inaccessible, skip it and note in output
• If a platform is unavailable, report partial results from other platform
• Always complete the summary with available data

Privacy

• Only show activity from configured repos
• Do not expose email addresses, only usernames/display names
• Truncate long commit messages to first line

Configuration

Required Environment

GitHub:

• gh CLI installed and authenticated

Azure DevOps:

• az CLI installed and authenticated
• Access to Cloud-Delivery organization

Adding/Removing Repos

To modify tracked repositories, update this skill file:

• Add patterns to the GitHub patterns list
• Add org/project/repo entries to Azure DevOps list

Contributor Name & Timezone Mapping

The dashboard normalizes contributor names and applies timezone-aware work pattern analysis. To add or modify mappings, edit the CONTRIBUTOR_CONFIG object in dashboard/app/api/dashboard/route.ts:

const CONTRIBUTOR_CONFIG: Record<string, ContributorConfig> = {
  // Travis Edgar - Atlantic Time (UTC-4)
  'rustyautopsy': { name: 'Travis Edgar', timezone: -4, tzName: 'Atlantic' },
  'tedgar': { name: 'Travis Edgar', timezone: -4, tzName: 'Atlantic' },

  // Mahyar - Pacific Time (UTC-8)
  'mahyar': { name: 'Mahyar', timezone: -8, tzName: 'Pacific' },

  // Jeff Harris - London UK (UTC+0)
  'jeharris-eci': { name: 'Jeff Harris', timezone: 0, tzName: 'London' },
};

Timezone offsets (hours from UTC):

This ensures:

• Activity from the same person across GitHub and Azure DevOps is aggregated correctly
• Late night / weekend detection uses each contributor's local time

Further Reading

• [Examples](EXAMPLES.md) - Common usage scenarios and expected output
• [Automation](AUTOMATION.md) - Email delivery and scheduling setup