badger-builder

badger-builder is an AI-assisted tool for generating dynamic Brute Ratel C4 profiles. Simply provide badger-builder a flavor for your desired profile and it will prompt OpenAI for fitting configurations.

Listener/payload profile configs that are AI generated:

C2 URIs
Request/response HTTP headers
HTTP body data prepended and appended to C2 requests/responses
The server's empty-response HTTP body

HTTP body data can be generated in XML, JSON or JavaScript (and request format can differ from response format).

I'd recommended always reviewing the configs generated by OpenAI, especially for cases where it failed to avoid usage of test/sample/placeholder values. The AI generated configs are generally a good starting point, but you'll likely want to make some modifications.

Supports Brute Ratel 1.5.2

Currently only HTTP listener profile configs are supported (no DoH support)

Note on OpenAI Usage

The prompts badger-builder sends to the OpenAI API consume tokens. Usage generally consumes a few thousand tokens (< $0.05 USD). Your OpenAI account must either be on the free trial or have setup API billing. (Check your usage and billing here)

Install

badger-builder can be installed by cloning this repository and running pip3 install . and subsequently executed from PATH with badger-builder

For those that hate poetry projects, there is a requirements.txt you can use with badger-builder.py to call the typer app's entrypoint.

Usage

Move .env.sample to .env, add your OpenAI API key and run badger-builder create

After a profle has been generated, use badger-builder parse to display sample HTTP requests using the profile configs.

Development

badger-builder uses Poetry to manage dependencies. Install from source and setup for development with:

git clone https://github.com/tw1sm/badger-builder
cd badger-builder
poetry install
poetry run badger-builder --help

requirements.txt can be generated with

poetry export --without-hashes --format=requirements.txt > requirements.txt

Improvements

Option (subcommand?) to provide flags via yaml file
Better/more token efficent OpenAI prompts
DoH/SMB/TCP listener integration
More prepend/append data formats
autorun actions such as setting kill date, chid, malloc/threadex settings

vysecurity / badger-builder