Author: Jason Tang
A plugin that provides LLM helpers to explain code and assist in RE.
This is a LLM plugin aimed at enabling the use of local LLM's (ollama, text-generation-webui, lm-studio, etc) for assisting with binary exploration and reverse engineering. It supports any OpenAI v1-compatible API. Recommended models are LLaMA-based models such as llama3.1:8b, but others should work as well.
Current features include:
- Explain the current function - Works at all IL levels.
- Explain the current instruction - Works at all IL levels.
- General query - Query the LLM directly from the UI.
- Propose actions - Provide a list of proposed actions to apply.
- Function calling - Allow agent to call functions to navigate the binary, rename functions and variables.
- RLHF dataset generation - To enable model fine tuning.
- RAG augmentation - Supports adding contextual documents to refine query effectiveness.
- Settings to modify API host, key, model name and max tokens.
Future Roadmap:
- Agentic assistant - Use Autogen or similar framework for self-guided binary RE.
- Model fine tuning - Leverage the RLHF dataset to fine tune the model.
- If necessary,
pip install -r requirements.txt
from the plugin directory. - Open Settings -> BinAssist.
- Ensure the RLHF and RAG database paths are appropriate for your environment.
- Point the API host to your prefered API provider and set the API key.
- Load a binary.
- Open BinAssist with the 'BA' sidebar icon and start exploring.
https://github.com/jtang613/BinAssist
An OpenAI compatible API is required. For local LLM support, use Ollama, LMStudio, Open-WebUI, Text-Generation-WebUI, etc.
pip install -r requirements.txt
Untested but should work. Submit an Issue or Pull Request for support.
Untested but should work. Submit an Issue or Pull Request for support.
This plugin requires the following minimum version of Binary Ninja:
- 4000
The following dependencies are required for this plugin:
- pip - openai, pysqlite3, markdown, httpx, chromadb, sentence-transformers
This plugin is released under a MIT license.
2