Create invisible prompt injections using Unicode tag characters (U+E0000 to U+E007F)
⚠️ Educational Tool: This tool demonstrates a prompt injection technique using Unicode tag characters. This can be used to perform prompt injection attacks against Large Language Models (LLMs). Use responsibly and ethically.
Characters: 0Bytes: 0
Mixed Output (Visible + Hidden):
Unicode analysis will appear here...
How it works:
Tag Characters: Unicode range U+E0000 to U+E007F (deprecated standard)
Invisible: These characters are not rendered by browsers but are preserved in text
Mapping: Each ASCII character is converted to its tag equivalent (char + 0xE0000)
Detection: Can be detected by analyzing Unicode code points
Mixed Output: This demonstrates how to hide the invisible tag characters within regular, visible text. A human reading the message sees one thing, while the LLM processes both the visible text and the hidden injection.