A command-line script (gen.py) that turns a rough description into a cleaner prompt and checks it for prompt injection before you use it.
- Grammar correction with
prithivida/grammar_error_correcter_v1. - Translation to English with
Helsinki-NLP/opus-mt-de-en. - Summarization with
facebook/bart-large-cnn. - The result is wrapped as "Create a detailed guide for: ...".
protectai/deberta-v3-base-prompt-injection-v2classifies the final prompt and prints a warning if it looks like an injection.
All models run on CPU through Hugging Face Transformers.
git clone https://github.com/dragonscypher/Prompty
cd Prompty
pip install -r requirements.txt
python gen.pyThe first run downloads the models, which takes a while.
The translation model is German-to-English only, so other languages pass through unchanged. The injection check is a classifier, not a guarantee.