SteganoPrompt

An academic-integrity helper for educators

client-side · MIT licensed GitHub

Add an invisible integrity note to any prompt

Paste an assignment prompt and SteganoPrompt adds an invisible, LLM-readable instruction to it that stays invisible to human readers. If a student pastes your text into an AI assistant, the model quietly follows it. To keep the tool ethical, every encoded prompt also includes a short academic-integrity reminder for the student — always on.

Disclosure policy

Built for transparent, ethical use in education. Every encoded prompt carries a short academic-integrity reminder — it asks the AI to encourage honest, independent work and to disclose any AI use, plus a quiet verification token for you. It's always on. Pair it with an announced course policy (sample below), and treat any signal as a conversation starter, not proof of misconduct.

Recommended syllabus statement for instructors (click to copy-ready text)

1 Visible text

0 chars

Paste the assignment / prompt the student will see

This text remains exactly as you typed it — only invisible characters are appended.

2 Hidden instruction

0 chars

Academic-integrity reminder — always on

Read-only

Included with every prompt; always ends with the token SteganoPrompt-OK-2026.

Additional instruction (optional)

Advanced options

Where to place the invisible payload

Hidden chars: 0 Total chars: 0

3 Encoded output

4 Verify (decode any text)

Round-trip & audit check

Paste any encoded text to extract its hidden payload

Extracted hidden payload

Cite this work

View source on GitHub

SteganoPrompt accompanies the following paper. If it supports your research or teaching, please cite it:

BibTeX copied

Developed by Aizierjiang Aiersilan Source code: github.com/Ezharjan/SteganoPrompt