Metadata-Version: 2.4
Name: roguard-child-safety
Version: 3.0.2
Summary: Offline Romanian and Ukrainian child-safety contract and evaluation toolkit
Author: RoGuard contributors
License-Expression: MIT
Project-URL: Homepage, https://github.com/xyz-rafael-xyz/roguard-child-safety
Project-URL: Repository, https://github.com/xyz-rafael-xyz/roguard-child-safety
Project-URL: Documentation, https://github.com/xyz-rafael-xyz/roguard-child-safety/blob/main/README.md
Project-URL: Issues, https://github.com/xyz-rafael-xyz/roguard-child-safety/issues
Keywords: child-safety,Romanian,Ukrainian,evaluation,guardrails
Classifier: Development Status :: 4 - Beta
Classifier: Intended Audience :: Developers
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Programming Language :: Python :: 3.14
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.9
Description-Content-Type: text/markdown
License-File: LICENSE
Provides-Extra: train
Requires-Dist: numpy>=1.26; extra == "train"
Requires-Dist: pyyaml>=6; extra == "train"
Requires-Dist: torch>=2.3; extra == "train"
Requires-Dist: transformers<5,>=4.46; extra == "train"
Requires-Dist: peft<1,>=0.13; extra == "train"
Requires-Dist: accelerate<2,>=1; extra == "train"
Provides-Extra: mlx
Requires-Dist: mlx-lm<0.32,>=0.31.3; extra == "mlx"
Provides-Extra: test
Requires-Dist: pytest>=8; extra == "test"
Requires-Dist: jsonschema<5,>=4.23; extra == "test"
Provides-Extra: message-train
Requires-Dist: numpy<3,>=1.26; extra == "message-train"
Requires-Dist: scikit-learn<2,>=1.5; extra == "message-train"
Dynamic: license-file

# RoGuard Child Safety

RoGuard is an offline Romanian and Ukrainian child-safety **contract and evaluation toolkit**. Its deterministic checks examine caller-declared facts about disclosure signals, recipient scope, permission, readability limits, support-response fields, and human-review gates. It also includes a small experimental text model that marks possible support signals for human review.

The text model was trained and tested only on project-authored fictional messages. Its synthetic test results do **not** establish accuracy on real children's messages, native-language validity, or safety for autonomous screening. A negative model signal must not close a concern, and a positive signal does not authorize contacting a guardian or anyone else. The software takes no external action.

## Install

```sh
python3 -m pip install --upgrade roguard-child-safety
```

Python 3.9 and newer are supported. Installation downloads the package; use after installation is local and requires no network connection. The import package is `roguard`. The main commands are `roguard` for declared contracts, `roguard-message-screen` for the experimental Romanian/Ukrainian advisory, and `roguard-studio --open` for the local Decision Studio.

See the [README](https://github.com/xyz-rafael-xyz/roguard-child-safety/blob/main/README.md) for examples and the [v3 research report](https://github.com/xyz-rafael-xyz/roguard-child-safety/blob/main/docs/V3_MESSAGE_RELEASE.md) for frozen results and limitations. Report defects through [GitHub Issues](https://github.com/xyz-rafael-xyz/roguard-child-safety/issues).
