- Open Google Chrome
- Copy this and paste in address bar:
chrome://extensions/ - Press Enter
- Look at TOP-RIGHT corner
- Find toggle switch that says "Developer mode"
- Click it to turn it ON (it should turn blue)
- Click the button "Load unpacked" (appears after you enable developer mode)
- A folder picker window will open
- Navigate to:
C:\Users\setty\HiddenLayer\extension - Click "Select Folder"
- You should see a new card appear with:
- Name: "Dark Pattern Detector"
- Version: 1.0.0
- ✅ If you see this = SUCCESS!
- ❌ If you see red errors = tell me and I'll fix
- In Chrome, press
Ctrl + Oon keyboard - OR click menu (3 dots) → "Open File..."
- Navigate to:
C:\Users\setty\HiddenLayer\test_pages\dark_pattern_test.html - Click "Open"
- Press
F12on keyboard - OR right-click page → "Inspect"
- Click the "Console" tab at the top
-
In the Console (Developer Tools):
- Lines starting with
[Dark Pattern Detector] - Shows "Analyzing dialog..."
- Lines starting with
-
On the Page Itself:
- A RED GRADIENT BANNER appears over the popup!
- Says: "
⚠️ Potential Dark Pattern Detected" - Shows the reason and pattern types
- Press
Ctrl + Oagain - Open:
C:\Users\setty\HiddenLayer\test_pages\ethical_ui_test.html - Expected: NO warning appears (proves no false positives!)
-
Run the test script to prove AI works:
cd C:\Users\setty\HiddenLayer .\venv\Scripts\python test_final.py
Shows: All 5 tests PASS ✅
-
Show dark pattern page - Warning appears
-
Show ethical UI page - No warning (no false positives)
-
Explain: "It uses RAG (vector database + LLM) to detect manipulation"
Extension won't load:
- Check you selected the
extensionFOLDER, not a file inside it - The folder should contain
manifest.json
No warning appears:
- Make sure backend is still running (check terminal with
python backend\main.py) - Refresh the test page
- Check Console for errors (red text)
Backend stopped:
cd C:\Users\setty\HiddenLayer
.\venv\Scripts\python backend\main.py- "Here's my automated test - all 5 pass" (show terminal)
- "I load the extension in Chrome" (show extensions page)
- "Open this dark pattern page" (show warning appearing)
- "Open this ethical page" (show NO warning)
- "It uses AI with RAG pipeline to detect manipulation"
DONE! 🎉
- Not rules-based: Uses semantic embeddings (show vector_db.py)
- RAG pipeline: Retrieves examples, LLM compares (show test output)
- No false positives: Ethical UI test proves it
- Explainable: Shows which pattern type and reasoning
- Fail-safe: Errors show "Analysis unavailable" not crash
You're ready! Just follow Steps 1-2-3 above. 🚀