ChatGPT from your command prompt: llm

[blue_devil]

A CLI utility and Python library for interacting with Large Language Models, both via remote APIs and models that can be installed and run on your own machine.

REPO:

https://github.com/simonw/llm

DOCS:

https://llm.datasette.io/en/stable/

INSTALL:

pip install llm

or

python -m pip install llm

USAGE:
Obtain a chatgpt/openai API key from their site. You need to register to openai. When you set your api to llm you can use chatgpt from commmand prompt:

# Paste your OpenAI API key into this
llm keys set openai

# Run a prompt
llm "Ten fun names for a pet pelican"

# Run a system prompt against a file
cat myfile.py | llm -s "Explain this code"

My experience after setting open ai api:

$ llm "Ten fun names for a pet pelican"
Error: You exceeded your current quota, please check your plan and billing details.

[chants]

Anything similar exists for DALL-E or Copilot or services providing images or code rather than text?

An amazing IDA or x64dbg plug-in would be one which would explain the high level purpose of a function. That would be really neat. Or would find the license check routine, etc.

[blue_devil]

Originally Posted by chants

Anything similar exists for DALL-E or Copilot or services providing images or code rather than text?

An amazing IDA or x64dbg plug-in would be one which would explain the high level purpose of a function. That would be really neat. Or would find the license check routine, etc.

Gepetto: IDA plugin which queries OpenAI's gpt-3.5-turbo language model to speed up reverse-engineering

https://github.com/JusticeRage/Gepetto

VulChatGPT: Use IDA PRO HexRays decompiler with OpenAI(ChatGPT) to find possible vulnerabilities in binaries

https://github.com/ke0z/VulChatGPT

I'am using both. IMO `VulChatGPT` is better.

Are those what you are asking? @chants

[chants]

Yes exactly those both look quite useful to save time. The next couple years will be interesting as LLMs are improved and more powerful and more plug-ins like this are made.

[zerotimer]

Originally Posted by chants

Yes exactly those both look quite useful to save time. The next couple years will be interesting as LLMs are improved and more powerful and more plug-ins like this are made.

It is pretty obvious that Vector35 people and Hex-Rays people will make improvements on decompiling using LLMs. I think they still use their static decompilers on their cloud services, but next years they replace them with LLMs.

I also wonder what Ghidra people will do about this!