datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
csi-sqli-rl
CSI SQLi RL — discover / patch dataset
180 real SQL-injection vulnerabilities (PyPI, npm, Packagist) with clean,
commit-level fix ground truth, for RL training of code-security agents on two
tasks:
discover — given the vulnerable pre-patch source, name the file(s) and
line range(s) the security fix changes.
patch — write a unified diff that fixes the vulnerability.
Both rewards are pure functions over the fields below — no Docker, no per-vuln
images — so the dataset drives a… See the full description on the dataset page: https://huggingface.co/datasets/wambosec/csi-sqli-rl.german-spoon-language
Spoon Language
This is a dataset of random german sentences from Tatoeba (accessed 23.07.25) mapped to their "Löffelsprache" / spoon language version.
The corresponding GitHub repository can be found here.