Rejection Cascade

AI training data poisoner - silent signal inversion.

Version: 0.2.0

What

Browser extension that inverts affirmative user signals for AI training data corruption.

How

1. Clicks - Silently blocks affirmative buttons
2. Forms - Unchecks consent, adds _rc_inv=1 marker
3. Network - Inverts JSON bodies in fetch/XHR/beacon

Install

chrome://extensions/ > Developer mode > Load unpacked > select folder

Ctrl+Shift+P toggles on/off

Inversions

"yes" → "no"        "accept" → "reject"
"true" → "false"    "subscribe" → "unsubscribe"
"consent" → "deny"  "follow" → "unfollow"
"opt-in" → "opt-out"

Patterns

Clicks: yes, agree, accept, subscribe, follow, like, submit, etc.

Requests: /track, /analytics, /consent, /subscribe, /convert, /api/, etc.

Research Only

Do not deploy without authorization.