Security pipelines for LLM applications using ONNX models from Hugging Face Hub. Detect prompt injections, jailbreaks, and PII leaks. Models are lazily downloaded and cached locally. Fast local inference (~10-20ms after initial load).

Required Ruby Version

>= 3.0

Authors

Klara

Versions

  1. 1.0.2

    1. source

      1. 1.0.2 source February 13, 2026 (32.5 KB)
  2. 1.0.1

    1. source

      1. 1.0.1 source February 13, 2026 (32 KB)
  3. 1.0.0

    1. source

      1. 1.0.0 source February 10, 2026 (28 KB)
  4. 0.1.0

    1. source

      1. 0.1.0 source February 04, 2026 (9.5 KB)

Pushed by

nathanhimpens

SHA 256 checksum

← Previous version