Cencurity Engine – Security for AI-generated code (CAST)

Shipping AI tools is getting weird.

We tried to release a VSCode extension.

It kept hitting 500 errors. No explanation.

Maybe just a glitch.

Maybe something else.

Either way, we stopped waiting.

So we shipped the engine instead.

Cencurity Engine (CAST)

A streaming security engine for AI-generated code.

Instead of scanning code after it's written,

Cencurity runs inline with code generation

and enforces security decisions in real time.

What it does:

- blocks dangerous code (eval, subprocess, unsafe execution)

- redacts secrets (API keys, credentials)

- enforces policies during generation

Why this matters:

AI is now writing real production code.

But security workflows are still built for

post-generation analysis (SAST, reviews, etc).

That gap is where insecure code slips through.

CAST closes that gap.

Works across:

- OpenAI-compatible models (xAI, DeepSeek, LLaMA)

- GPT

- Claude

- Gemini

Same engine. Same enforcement.
No plugin required.

Just run it as a local proxy

between your IDE/agent and the LLM.

100% open source.

Open to direct feedback — what works, what doesn’t.
https://github.com/cencurity/cencurity-engine#