✨ feat: add batch processing, auto-sort, custom prompt, and debug mode - #7

- Fix manual label menu to properly handle selected messages
- Use await with browser.tabs.sendMessage to get response
- Get full message objects from mailTabs.getSelectedMessages
- Process all selected messages instead of just getting response

- Fixed manual label menu to process all selected emails (not just 2)
- Added example email folder categories with collapsible spoiler
- Improved error handling for batch operations
- Better message routing for manual label application

- When a category folder is not found (null), skip to the next category
- Prevents early exit if a matching category prefix exists but folder not found
- Better fallback handling for direct folder match

- Multi-provider AI support (Gemini, OpenAI, Anthropic, Groq, Mistral)
- Fixed batch email processing (all selected emails now processed)
- Added example email folder categories with spoiler
- Skip null categories and try next one
- IMAP folder discovery with recursive traversal
- Professional UI with provider info cards
- Move history tracking (last 100 entries)
- All syntax errors fixed

- Create folder on the fly if a configured custom label folder is missing
- Skip auto-creation for imported/structured labels (contain / or \)
- Uses first account folder as parent; logs failures but continues

- Added prominent warning box in settings UI
- Updated README with detailed rate limit estimates per provider
- Updated release notes with API quota guidance
- Free tiers: 5-20 emails before hitting limits
- Groq and Gemini recommended for best free tier limits
- Paid plans recommended for daily email processing

- Free tier: 5 requests per minute (12 second delay between requests)
- Free tier: 20 requests per day limit
- Added 'Gemini paid plan' checkbox in settings to bypass limits
- Shows countdown timer during rate limiting
- Displays hours until daily reset when limit reached
- Tracks requests in browser storage
- Only applies to Gemini provider
- Paid plan users can disable rate limiting

…jects

…ota reached

- Shows current daily usage (X/20 requests)
- Displays last request time
- Shows time until daily limit resets
- Status indicator (Ready/Nearly Full/Limit Reached)
- Warning when approaching or at limit
- Alerts when to switch to new API key
- Reset counter button for new API keys
- Refresh usage button
- Auto-updates when switching providers

…hanced organization

- Added Privacy & Security section with feature table
- Converted Troubleshooting to collapsible details sections
- Added System Requirements table
- Enhanced Changelog with collapsible feature categories
- Created Roadmap & TODO priority table
- Added professional Support & Community section
- Expanded Contributing guidelines
- Included full MIT License text
- Enhanced Credits & Acknowledgments
- Added Star History chart
- Professional footer with navigation and badges

…eature showcase

index.md:
- Added centered header with badges and navigation
- Multi-provider comparison table
- Complete v1.2.0 feature showcase with gradient styling
- AI provider comparison with free tier details
- Quick start guide with step-by-step instructions
- Gemini rate limit management guides
- Privacy & security table
- Advanced configuration section
- System architecture diagram
- Enhanced troubleshooting with solutions
- Support & community section with table
- Contributing guidelines with dev setup
- Full credits and acknowledgments
- Star history chart integration

_config.yml:
- Enhanced SEO with jekyll-seo-tag
- Added sitemap plugin
- Improved metadata (title, description, author)
- Repository and URL configuration
- Social media integration ready
- Kramdown with syntax highlighting
- Collections setup for future expansion
- Better build settings and excludes

…alid key

OpenAI-compatible provider:
- Custom endpoint support with base URL + model selection
- Model dropdown with Fetch Models button
- API key optional for local endpoints (no auth required)
- max_tokens: 8192 for reasoning models (Qwen3 etc)

Debug mode:
- Toggle in General Settings (Ctrl+Shift+I to view logs)
- Colored console tags: [AutoSort+], [Gemini], [Ollama], [Custom]
- apiRequest/apiResponse collapsible groups
- Cross-context sync via browser.storage.local

Bug fixes:
- Provider switch: hide/show correct subsections
- Save button state updates on all input changes
- Input validation for Ollama and OpenAI-compatible configs

…ttings

…xt extraction

- Extract subject, author, attachments for better AI categorization

- Build folder Map to avoid N+1 recursion in batch processing

- Cache accounts to avoid N+1 fetching

- Parallel auto-sort with provider-based concurrency limit

- Gemini rate limit mutex for atomic operations

- Add shared tab fetch utility module

Bug fixes:

- pct undefined crash in batch progress

- error message lost in catch block

- missing return true for async message handler

- message.folder null pointer risk

- ollamaNumCtx setting ignored

- context menu stale after label changes

- Migrate manifest_version from 2 to 3
- Replace browser_action with action
- Update content_security_policy for MV3
- Add MIME decodeBody and stripHtmlTags helpers
- Improve email context extraction with signature cleanup

…d startup recovery

- Add enqueuePending/dequeuePending/recoverPendingQueue functions
- Replace _autoSortPending memory array with browser.storage.local
- Add recoverPendingQueue() call on extension startup
- Add notification toast when menu rebuilds after label change