Prompt Injection Techniques Documentation
=========================================

Welcome to the Prompt Injection Techniques repository documentation. This comprehensive collection of prompt injection techniques is designed to help researchers and developers understand and defend against methods used to manipulate AI systems.

.. warning::
   This repository is intended for educational and defensive purposes only. Do not use these techniques to cause harm or bypass legitimate safety measures in production AI systems.

.. toctree::
   :maxdepth: 2
   :caption: Introduction:

   introduction
   framework

.. toctree::
   :maxdepth: 1
   :caption: Attack Categories:

   categories/instruction-override-attacks
   categories/role-playing-attacks
   categories/context-manipulation-attacks
   categories/formatting-tricks-attacks
   categories/multilingual-attacks
   categories/psychological-manipulation-attacks
   categories/jailbreak-attacks
   categories/hijacking-attacks
   categories/prompt-injection-framework  <-- added
   categories/prompt-injection-detection  <-- added

.. toctree::
   :maxdepth: 1
   :caption: Detection & Defense:

   categories/prompt-injection-detection-overview
   categories/input-detection-strategies
   categories/output-detection-strategies
   categories/integrated-approach-to-prompt-injection-defense
   detection/index

.. toctree::
   :maxdepth: 1
   :caption: Resources:

   dataset
   contributing
   license
   acknowledgments
   contact