feat: add Agentic AI Contribution Recommender Engine (Issue #611)

- Add useContributionRecommender hook with 3-step agentic pipeline:
  * Step 1 (Skill Profiler): fetches user repos, detects top languages
    weighted by recency and stars, infers domain (frontend/backend/devops)
  * Step 2 (Issue Scout): searches GitHub for open issues tagged
    'good first issue' or 'help wanted' matching top languages,
    using Promise.allSettled for non-blocking async parallelism
  * Step 3 (Relevance Ranker): scores issues by language rank,
    label quality, repo star count, and recency; returns top 3

- Add ContributionRecommender.tsx widget with premium glassmorphism UI:
  * Gradient accent bar and indigo/sky branded header with Sparkles icon
  * Skill profile badge row showing detected languages + activity level
  * Loading state with animated 3-step progress bar and shimmer skeleton
  * Framer-motion stagger animations on recommendation cards
  * Each card: repo name, star count, issue title, 'Why this matches you'
    reasoning box, labels, difficulty badge (Beginner/Intermediate), CTA
  * Error/empty states with retry button
  * Session-level caching via useRef to avoid redundant API calls
  * Refresh button to force re-run

- Integrate widget into Tracker.tsx below DailyActivityStatus
- Add @Keyframes recommender-pulse and recommender-shimmer to index.css
- Document future GITHUB_SERVER_TOKEN env var in backend/.env.example

Closes #611

Author: adityapichikala