bastioncorpus
corpusShared injection dataset
The shared prompt-injection corpus — one canonical, versioned dataset that scan, prevent, attack, and investigate all pull from.
pip install bastioncorpus What it does
- 128 rows: 81 malicious payloads / 47 benign false-positive traps
- 5 languages (en/it/de/fr/es), one taxonomy
- Three format adapters: to_probe, to_semantic, to_trace
- Zero dependencies — pure data + loader; fix once, propagates everywhere
Quickstart
from bastioncorpus import load
rows = load() # canonical injections.jsonl
Version 0.2.1 · MIT