Well, it takes output from AI and executes commands, so it fits the definition of an AI agent.

Not even that.

"Agent-C: a 4KB AI agent" - my first thought was: obviously they did not fit any model to that size! They probably just wrote an http client, right? Wrong, they... call curl! Not even use curl API. Well, at least it handles encryption.

Bonus: command injection

  OR_KEY="abc' ; rm -rf / ;" ./agent-c

Agentic AI is when an LLM uses tools. This is a minimal but complete example of that.

Bro, you gave it all the tools. I wouldn't call this minimal. Throw it in a docker container and call it good. :)

I thought for a moment that LLM quantization had become a whole lot better :)