sanitycheck

Overview

• Scrapes tweet data from X/Twitter timeline and injects inflammatory content indicators
• Uses AI classification with 1-5 scale rating system
• Smart caching system prioritizes most inflammatory tweets
• Alters feed based on inflammatory ratings

Install (Unpacked)

• In Chrome, open chrome://extensions
• Enable Developer Mode (top-right)
• Click "Load unpacked" and select the x-feed-indicator-extension folder
• Navigate to https://x.com and open the home feed
• Open DevTools Console to see logged data prefixed with [XFI]

Files

• manifest.json: MV3 manifest
• scripts/content.js: Tweet detection, data extraction, AI classification
• scripts/background.js: Anthropic API integration
• options/index.html + options/index.js: Settings and API key storage
• popup/index.html + popup/index.js: Classification interface

Features

• Real-time inflammatory content detection (1-5 scale)
• Account type detection (verified, government, business)
• Engagement metrics tracking (likes, comments, retweets)
• Smart caching (30 most inflammatory tweets)
• Political ideology spectrum (left, left-center, center, right-center, right)

Anthropic Integration

• Set your API key in extension Options (sk-ant-...)
• Uses Claude-3.7-Sonnet model for classification
• Concurrency limited to 2 simultaneous requests
• Keys stored via chrome.storage.local

Classification System

• 1-5 inflammatory scale with color-coded indicators
• Custom classification prompts supported
• Real-time processing with smart deduplication
• Error handling for missing API keys or failed requests