A small story to motivate you guys to work on cutting edge frontier tech:
You might have heard about Gemini CLI[1] release a few days back. I managed to integrate Apple Containers (also released recently) to execute the code generated by AI safely.
Gemini CLI supports code execution itself but via shell, which can be a security issue. CodeRunner solves this by running it in a safe isolated container.
Got appreciation from Google team:
cperry is from the Google team who commented on my reply with "<3 amazing". I will consider it a win.
Some asked me,
That is amazing. Could you tell me how you did it and how you even thought of it in the first place? I'm a student so please eli5
Context: Everyday something new is released, I read about it and in depth -- doesn't take a lot of time now as I have been doing this for a few years.
When Apple Container was released I had already read about WASM[1], docker[2] etc. The native apple containers are kind of a big thing I felt, so I and my friend created CodeRunner[4] for apple containers 4 days back.
I had also seen a trend of MCPs[3] -- initially I brushed it off but it seems there are a lot of tech built around it. So, converted my APIs (which eventually execute code) to MCP APIs which do the same thing but are now MCP compatible.
Cut to Gemini CLI release, I went to check if they have MCP support and since I had read about MCPs before I knew it would be a quick change to make it work for Gemini and indeed they did have support and after that it was a matter of config change!
Moral of the story is, be ready, opportunities will come and then you strike.
And if somehow we magically reach 500 stars on github -- who knows it might become part of official Gemini CLI. Lot of small wins can lead to a big win!
WASM: https://en.wikipedia.org/wiki/WebAssembly
docker: https://en.wikipedia.org/wiki/Docker_(software)
MCP: https://en.wikipedia.org/wiki/Model_Context_Protocol
CodeRunner (mine): https://github.com/BandarLabs/coderunner
Top comments (0)