Changkun Dotfiles Dotfiles

GitHub Activity Analyzer Skill

Analyzes comprehensive GitHub activity for any specified time period with deep technical insights. Supports arbitrary time frames from days to full years.

When to Use This Skill

Use this skill when the user:

• Asks "what have I been working on?"
• Requests a summary of recent contributions (any time period)
• Wants to review their GitHub activity
• Needs to prepare a status update or report
• Says "show me my recent work" or "what did I do last week/month/year"
• Asks about contribution statistics for a specific period
• Wants insights into their coding patterns over time
• Requests activity analysis with various time formats:
  • Relative: "last 7 days", "past month", "last year"
  • Full years: "2025", "year 2024", "all of 2023"
  • Date ranges: "2024-10-01 to 2024-12-31", "Q4 2024", "October 2024"
  • From date: "since 2024-10-01", "from October"

Capabilities

1. Event Analysis: Fetches and categorizes all GitHub events (pushes, PRs, reviews, comments, releases)
2. Content Deep Dive: Extracts PR titles, descriptions, commit messages, and review comments
3. Theme Identification: Groups contributions by technical domain (e.g., performance, infrastructure, features)
4. Impact Assessment: Highlights most significant contributions and their business value
5. Statistics: Provides counts and trends across repositories

How It Works

Time Period Parsing

Extract and parse time period from user's request in multiple formats:

A. Relative time (days back from today):

• "last week", "7 days", "1w" → 7 days
• "last 2 weeks", "14 days", "2w" → 14 days (default if not specified)
• "last month", "30 days", "1m" → 30 days
• "last 6 months" → 180 days
• "last year", "365 days", "1y" → 365 days
• Custom: "45 days" → 45 days

B. Full year:

• "2025", "year 2025", "all of 2024" → Date range: YYYY-01-01 to YYYY-12-31
• Parse the year and set as full calendar year

C. Specific date range:

• "2024-10-01 to 2024-12-31" → Exact date range
• "from October 1st to December 31st 2024" → Parse natural language to YYYY-MM-DD
• "Q4 2024" → 2024-10-01 to 2024-12-31
• "October 2024" → 2024-10-01 to 2024-10-31
• "Q1 2025" → 2025-01-01 to 2025-03-31

D. From date to today:

• "since 2024-10-01" → From that date to today
• "from October 2024" → From 2024-10-01 to today

Default to 14 days (2 weeks) if no time period mentioned

Data Collection

• Uses gh api users/{username}/events to fetch activity
• For relative time: Dynamically calculates date N days ago from today
• For date ranges: Uses the parsed start and end dates directly
• Filters events within the date range (inclusive)
• Fetches detailed PR information for opened/merged PRs
• Retrieves commit messages from most active repositories
• Samples code review comments to understand review style

Important Note: GitHub's events API only returns the last 90 days of events. For time periods beyond 90 days:

• Use GitHub Search API for issues and PRs: gh api search/issues
• Query specific repository commits: gh api repos/{owner}/{repo}/commits
• Fetch data from multiple endpoints and aggregate
• Focus on major contributions and highlights for longer periods

Analysis

• Groups work by repository and technical theme
• Identifies patterns (e.g., VAT integration, performance optimization)
• Extracts technical details (formulas, benchmarks, metrics)
• Summarizes code review leadership

Output Format

• Executive summary
• Activity breakdown by theme
• Technical highlights with specifics
• Statistics table
• Impact assessment

Technical Requirements

• GitHub CLI (gh) must be installed and authenticated
• Access to user's GitHub events via API
• Ability to fetch PR and commit details from repositories

Implementation Guidelines

• Time Period Handling:

  • Always parse the user's requested time period from natural language
  • Support all formats: relative days, full years, specific ranges, from-date-to-today
  • Default to 14 days (2 weeks) if no period specified
  • Calculate dates dynamically from today's date for relative periods
  • Be flexible: "last week", "7 days", "1w" should all work the same
  • Convert natural language dates: "October 2024" → "2024-10-01 to 2024-10-31"
  • Handle quarters: "Q1 2025" → "2025-01-01 to 2025-03-31"
  • For periods >90 days, acknowledge API limitations and use alternative approaches

• API Strategy for Different Time Periods:

  • 0-90 days: Use events API directly (gh api users/{username}/events)
  • >90 days or specific old dates: Combine multiple approaches:
    • Search API: gh api search/issues?q=author:{username}+created:>=YYYY-MM-DD
    • Search API for PRs: gh api search/issues?q=is:pr+author:{username}+created:>=YYYY-MM-DD
    • Repo commits: Identify active repos and query commits
    • Focus on major milestones and highlights

• Content Adaptation by Period Length:

  • <30 days: Full detailed analysis with all PRs, commits, reviews
  • 30-90 days: Detailed with slight summarization of minor items
  • 3-6 months: Focus on major contributions, monthly themes
  • Full year: Quarterly breakdown, key highlights, statistics, major impact

• Technical Details:

  • Include ticket numbers (e.g., RMC-XXX, YIELD-XXXX) when present in commits/PRs
  • Provide specific technical details with formulas, benchmarks, and metrics - not generic summaries
  • Balance breadth (number of repos) with depth (technical implementation details)
  • Preserve context like performance gains, schema changes, integration specifics

• Performance:

  • Use parallel API calls where possible for better performance
  • For very long periods, paginate and aggregate strategically
  • Cache intermediate results if processing multiple endpoints

Source: changkun/dotfiles — distributed by TomeVault.