Canvas Poisoning

Detection approach

Canvas poison detected. 2D rendering of a transparent 1x1 image returned poisoned non-zero pixel data. This proves a privacy extension or automated spoofing framework (like Brave's Farbling or CanvasBlocker) is actively poisoning canvas outputs.

How this helps detect bots

Collect this signal alongside session context, request metadata, and other independent indicators. A single anomaly is a lead, not a verdict; corroborating anomalies make automated traffic more likely.

Legitimate traffic to consider

Brave users (with "Farbling" enabled) or users with CanvasBlocker privacy extensions.