feat: v1.1.0 - dark mode, selective AI matching, UI improvements

- Add dark mode with system/light/dark toggle defaulting to system preference
- Add selective AI matching with checkbox selection on queued posts
- Add individual "AI Match" button on each queued post card
- Add Reset Extension button on setup wizard and settings pages
- Add configurable RSS feed limits (Max RSS Items, Max Blog Items)
- Add collapsible RSS Preferences and Matching Preferences sections
- Remove Jump button; View button now opens post in new tab
- Collapse Reply Suggestions by default; hide empty Replied/Skipped tabs
- Fix post content newline preservation
- Fix RSS feed host permission prompting during setup
- Fix AI model default selection with vendor prefix
- Fix content script injection on already-loaded LinkedIn pages
- Fix AI Match processing state to wait for all API calls

Author: charlesjones-dev

CHANGELOG.md

@@ -7,6 +7,37 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [1.1.0] - 2025-01-19
+
+### Added
+
+- Selective AI matching: checkbox selection on queued posts to match only specific posts
+- Individual "AI Match" button on each queued post for single-post matching
+- Button dynamically changes from "AI Match All" to "AI Match Selected (N)" when posts are selected
+- Dark mode support with three-state toggle (system/light/dark) in header, defaults to system preference
+- Reset Extension button on setup wizard pages to abandon setup and clear any saved config values
+- Reset Extension section in settings to clear all settings and return to setup wizard
+- Configurable RSS feed limits: Max RSS Items (default 25) and Max Blog Items in AI Prompts (default 25) in Settings
+- RSS Preferences collapsible section for Cache Duration, Max RSS Items, and Max Blog Items in AI Prompts settings
+
+### Changed
+
+- Moved Matching Preferences into collapsible section within AI Model settings
+- Removed Jump button from post cards; View button (opens post in new tab) is now the only action
+- Reply Suggestions section now collapsed by default; click header with chevron to expand
+- Replied and Skipped tabs now hidden when empty, reducing visual clutter
+- Filter tabs now display count below the label instead of inline
+
+### Fixed
+
+- Post content now preserves newlines and paragraph breaks instead of collapsing all whitespace into single spaces
+- RSS feed validation now properly prompts for host permission during setup and settings, preventing CORS errors on first use
+- AI model now correctly selected by default on Settings page after setup wizard (fixed missing vendor prefix in default model ID)
+- ModelSelector now auto-selects the first available model if the stored model no longer exists or is filtered out
+- Post status tags now display correct labels: "Matched" tab shows "Matched" badge (was "Pending"), "Unmatched" tab shows "Unmatched" badge (was "Queued")
+- Side panel now detects posts when opened on already-loaded LinkedIn pages by programmatically injecting the content script if needed (previously required hard-refresh)
+- AI Match now waits for both matching and heat check API calls to complete before switching to the Matched tab and clearing the "Analyzing..." state; the processing UI stays visible even when queue becomes empty mid-processing
+
 ## [1.0.0] - 2025-01-18
 
 Initial release of ReplyQueue, a Chrome extension that helps content creators find relevant social media posts and generate AI-powered reply suggestions.
@@ -19,5 +50,6 @@ Initial release of ReplyQueue, a Chrome extension that helps content creators fi
 - Reply suggestion generation with customizable writing style
 - Side panel UI with post cards, filtering, and settings
 
-[Unreleased]: https://github.com/charlesjones-dev/replyqueue/compare/v1.0.0...HEAD
+[Unreleased]: https://github.com/charlesjones-dev/replyqueue/compare/v1.1.0...HEAD
+[1.1.0]: https://github.com/charlesjones-dev/replyqueue/compare/v1.0.0...v1.1.0
 [1.0.0]: https://github.com/charlesjones-dev/replyqueue/releases/tag/v1.0.0

CLAUDE.md

@@ -49,22 +49,32 @@ Each adapter implements: `extractPost()`, `getPostUrl()`, `scrollToPost()`, `isF
 `src/shared/messages.ts` defines typed messages for cross-context communication:
 
 - Content → Background: `POSTS_EXTRACTED`, `CONTENT_SCRIPT_READY`
-- Side Panel ↔ Background: `FETCH_RSS`, `AI_MATCH_POSTS`, `GENERATE_SUGGESTIONS`
+- Side Panel ↔ Background: `FETCH_RSS`, `AI_MATCH_POSTS`, `GENERATE_SUGGESTIONS`, `RESET_EXTENSION`
 - Background → Content: `SCROLL_TO_POST`, `EXTRACT_POSTS`
 
 ### Vue 3 Composables
 
 `src/sidepanel/composables/` extracts business logic into reusable composables:
 
+- `useAppState.ts` - Global app state (loading, errors)
+- `useClipboard.ts` - Copy-to-clipboard functionality
 - `useConfig.ts` - Extension configuration persistence
+- `useCreditsModal.ts` - Insufficient credits modal state
+- `useModels.ts` - OpenRouter model selection and filtering
+- `useNetworkStatus.ts` - Online/offline detection
 - `usePosts.ts` - Matched posts state with filtering
-- `useModels.ts` - OpenRouter model selection
+- `useSettingsView.ts` - Settings page state management
 - `useSetup.ts` - Onboarding flow logic
+- `useTabStatus.ts` - Content script connection status
+- `useTheme.ts` - Dark/light/system theme management
+- `useToast.ts` - Toast notification system
 
 ### Storage Strategy
 
-- **chrome.storage.sync** - Small config synced across devices (config, cachedModels)
-- **chrome.storage.local** - Larger local data (extractedPosts, matchedPostsWithScore, cachedRssFeed)
+- **chrome.storage.sync** - Small config synced across devices (config, cachedModels, exampleComments if <8KB)
+- **chrome.storage.local** - Larger local data (extractedPosts, matchedPostsWithScore, cachedRssFeed, aiMatchCache, evaluatedPostIds)
+- **API keys** - Always stored in local storage only (never synced for security)
+- **Fallback** - If sync storage fails, automatically falls back to local storage
 
 ## Testing
 
@@ -117,10 +127,12 @@ import { shared } from '@shared/types'   // → src/shared/types
 
 ## Workflow
 
+**IMPORTANT:** After every feature implementation or bug fix, you MUST run `pnpm build`. Chrome loads the extension from the `dist/` folder, so changes are not reflected until a build is run. The user will be testing the extension manually after each change.
+
 After implementing any plan, always run preflight checks:
 
 ```bash
-pnpm build        # Type check + build
+pnpm build        # Type check + build (REQUIRED after every change)
 pnpm lint:fix     # Auto-fix lint issues
 pnpm format       # Auto-fix formatting
 pnpm test         # Run tests
@@ -129,6 +141,24 @@ pnpm audit:check  # Check for vulnerable dependencies
 
 Fix any errors before considering the implementation complete.
 
+### Changelog and Versioning
+
+After preflight checks pass, use the `AskUserQuestion` tool to ask how to handle versioning:
+
+- **Current version** - Add the feature/fix to the existing version section in `CHANGELOG.md`
+- **New version** - Increment version in `manifest.json` and `package.json`, then add a new section in `CHANGELOG.md`
+- **Skip** - Do nothing.
+
+When incrementing versions:
+- Patch (x.x.X): Bug fixes, minor improvements
+- Minor (x.X.0): New features, non-breaking changes
+- Major (X.0.0): Breaking changes
+
+Files to update for new versions:
+- `manifest.json` - Update `version` field
+- `package.json` - Update `version` field
+- `CHANGELOG.md` - Add new version section with date, update comparison links at bottom
+
 ## Security
 
 ### Security Tooling

CONTRIBUTING.md

@@ -62,8 +62,7 @@ replyqueue/
 │   ├── sidepanel/      # Vue 3 side panel UI
 │   └── shared/         # Shared utilities and types
 ├── tests/              # Test files (mirrors src/ structure)
-├── public/             # Static assets
-└── docs/               # Documentation
+└── public/             # Static assets
 ```
 
 ---
@@ -118,7 +117,15 @@ const doubled = computed(() => count.value * 2)
 
 ### Formatting
 
-The project uses default Vite/TypeScript formatting. Key points:
+The project uses Prettier for formatting and ESLint for linting:
+
+```bash
+pnpm lint        # Check for lint errors
+pnpm lint:fix    # Auto-fix lint issues
+pnpm format      # Format code with Prettier
+```
+
+Key formatting rules:
 - 2 spaces for indentation
 - Single quotes for strings
 - No semicolons (unless required)
@@ -217,12 +224,18 @@ Use descriptive branch names:
    pnpm test
    ```
 
-4. **Build the extension**
+4. **Lint and format**
+   ```bash
+   pnpm lint:fix    # Fix lint issues
+   pnpm format      # Format with Prettier
+   ```
+
+5. **Build the extension**
    ```bash
    pnpm build
    ```
 
-5. **Test manually**
+6. **Test manually**
    - Load the extension in Chrome
    - Test your changes in a real scenario
 
@@ -250,7 +263,10 @@ Use descriptive branch names:
 
 ## Checklist
 - [ ] Code follows project style
-- [ ] Tests pass locally
+- [ ] Linting passes (`pnpm lint`)
+- [ ] Code formatted (`pnpm format`)
+- [ ] Tests pass locally (`pnpm test`)
+- [ ] Build succeeds (`pnpm build`)
 - [ ] Documentation updated
 - [ ] No console.log in production code
 ```

README.md

@@ -15,7 +15,8 @@ Content creators spend hours scrolling social feeds hoping to find relevant disc
 - **Passive Discovery** - Extracts posts as you scroll, no extra workflow
 - **Smart Matching** - Keyword and AI-powered relevance scoring against your content
 - **Reply Suggestions** - Generates responses that match your writing style
-- **Jump to Post** - One click to navigate directly to matched posts
+- **View Post** - One click to open matched posts in a new tab
+- **Dark Mode** - System, light, or dark theme with one-click toggle
 - **Privacy-First** - No analytics, no tracking, your data stays local
 - **Extensible** - Platform adapter pattern makes adding new networks straightforward
 
@@ -193,6 +194,8 @@ See [CONTRIBUTING.md](./CONTRIBUTING.md) for detailed guidelines.
 | Relevance Threshold | 30% | 10-90% | Minimum score for a post to be considered a match |
 | Max Posts to Show | 20 | 10, 20, 30, 50, 100 | Maximum number of matched posts to display |
 | RSS Cache Duration | 60 min | 15, 30, 60, 120, 240 | How long to cache RSS feed before refreshing |
+| Max RSS Items | 25 | 10, 25, 50, 100 | Maximum items to fetch from RSS feed |
+| Max Blog Items in AI Prompts | 25 | 10, 25, 50, All | Blog items included when AI evaluates posts |
 | Blog Content Sent to AI | 2,500 chars | 1K, 2.5K, 5K, 10K, No limit | Amount of blog content included when matching |
 | Social Post Content Sent to AI | 1,000 chars | 500, 1K, 2K, 3K, No limit | Amount of post content included when evaluating |
 

manifest.json

@@ -2,13 +2,13 @@
   "manifest_version": 3,
   "name": "ReplyQueue",
   "description": "Side panel for finding social media posts relevant to your blog content and get AI-powered reply suggestions.",
-  "version": "1.0.0",
+  "version": "1.1.0",
   "icons": {
     "16": "icons/icon16.png",
     "48": "icons/icon48.png",
     "128": "icons/icon128.png"
   },
-  "permissions": ["storage", "sidePanel", "activeTab"],
+  "permissions": ["storage", "sidePanel", "activeTab", "scripting"],
   "host_permissions": ["https://www.linkedin.com/*", "https://linkedin.com/*", "https://openrouter.ai/*"],
   "optional_host_permissions": ["<all_urls>"],
   "content_security_policy": {

package.json

@@ -1,6 +1,6 @@
 {
   "name": "replyqueue",
-  "version": "1.0.0",
+  "version": "1.1.0",
   "type": "module",
   "scripts": {
     "dev": "vite",

privacy-policy.md

@@ -1,6 +1,6 @@
 # Privacy Policy for ReplyQueue
 
-**Last Updated:** January 2025
+**Last Updated:** January 2026
 
 ReplyQueue is a Chrome extension that helps content creators find social media posts relevant to their blog content and generate reply suggestions. This privacy policy explains how we handle your data.
 
@@ -21,15 +21,22 @@ All of the following data is stored locally on your device using Chrome's storag
 
 | Data Type | Storage Location | Purpose |
 |-----------|------------------|---------|
-| OpenRouter API Key | chrome.storage.sync | Authenticate with AI service |
+| OpenRouter API Key | chrome.storage.local | Authenticate with AI service (never synced for security) |
 | RSS Feed URL | chrome.storage.sync | Fetch your blog content |
 | Selected AI Model | chrome.storage.sync | Configure AI preferences |
-| Matching Preferences | chrome.storage.sync | Configure relevance settings |
-| Writing Style Examples | chrome.storage.sync/local | Personalize reply suggestions |
+| Matching Preferences | chrome.storage.sync | Configure relevance settings (threshold, cache TTL) |
+| Model Filter Preferences | chrome.storage.sync | Filter available AI models by criteria |
+| Communication Preferences | chrome.storage.sync | Custom instructions for reply tone/style |
+| Content Char Limits | chrome.storage.sync | Configure how much content to send to AI |
+| RSS Limits | chrome.storage.sync | Max RSS items to fetch and include in AI prompts |
+| Theme Preference | chrome.storage.sync | Dark, light, or system theme setting |
+| Writing Style Examples | chrome.storage.local | Personalize reply suggestions |
 | Extracted Posts | chrome.storage.local | Store posts from social media feeds |
-| Matched Posts | chrome.storage.local | Cache matching results |
+| Matched Posts | chrome.storage.local | Cache matching results with scores |
+| Evaluated Post IDs | chrome.storage.local | Track which posts have been analyzed |
 | Cached RSS Data | chrome.storage.local | Reduce network requests |
 | Cached Model List | chrome.storage.local | OpenRouter model metadata |
+| AI Match Cache | chrome.storage.local | Cache AI matching results to avoid re-processing |
 
 ### What We Do NOT Collect
 
@@ -52,12 +59,14 @@ ReplyQueue communicates with two external services. These communications only oc
 - When you fetch the list of available AI models
 - When you scan for relevant posts (AI matching)
 - When you generate reply suggestions
+- When heat check analyzes post tone/sentiment
 
 **What data is sent:**
 - Your API key (for authentication)
-- Post content from your social media feed (for matching)
+- Post content from your social media feed (for matching and tone analysis)
 - RSS feed content summaries (for matching)
 - Your writing style examples (if provided, for reply generation)
+- Your communication preferences (if provided, for reply style customization)
 
 **What data is NOT sent:**
 - Your personal information
@@ -116,6 +125,11 @@ All your data is stored locally and can be viewed:
 1. Open the extension's Settings panel
 2. Use the "Clear Cache" button
 
+**Reset extension (clear all settings and data):**
+1. Open the extension's Settings panel
+2. Scroll to "Reset Extension" section
+3. Click "Reset Extension" to clear all data and return to setup wizard
+
 ### Export Your Data
 
 Configuration data can be viewed in Chrome's storage inspector. Export functionality may be added in future versions.
@@ -139,6 +153,10 @@ Configuration data can be viewed in Chrome's storage inspector. Export functiona
 | `linkedin.com` | Extract posts from LinkedIn feeds |
 | `openrouter.ai` | Make API calls for AI matching and reply generation |
 
+### Optional Host Permissions
+
+ReplyQueue requests permission to access your RSS feed URL when you configure it. This is an optional permission that you grant during setup or when changing your RSS feed URL. The extension will prompt you to allow access to your blog's domain (e.g., `https://yourblog.com/*`) before fetching your RSS feed. Denying this permission prompt will prevent the extension from working.
+
 ---
 
 ## Children's Privacy