How pankajtanwarbanna/cringe-guard works
Project type and purpose
• This is a client-side Chrome extension (“cringe-guard”) whose purpose is to filter LinkedIn feed posts in real time using an external AI model. It injects a content script into LinkedIn pages, analyzes posts via an API, and then either blurs or removes “cringe” content. It also provides a popup/settings UI for users to input their API key and choose filtering modes.
File structure
• Top-level files:
• manifest.json – Chrome extension manifest defining content scripts and UI pages
• README.md, LICENSE, Privacy-Policy.md – docs
• pages/ directory (popup.html, settings.html) – extension UI for user settings
• scripts/ directory
• content.js – injected into LinkedIn pages, detects new posts, sends content to AI, manipulates DOM
• popup.js & settings.js – handle user input in popup.html and settings.html, persist to Chrome Storage
README for additional insights
• Architecture described in prose: real-time detection, API analysis, DOM filtering, user control via settings, persisting settings in Chrome Storage API.
Final advice
• Be very detailed: break out each script and UI page as its own component.
• Settings persistence and retrieval explicitly.
• Real-time nature of the content script’s interaction with the LinkedIn DOM.