I built a Python library to stop AI agents from leaking secrets (ModelFuzz)

A developer built ModelFuzz, an open-source Python library that intercepts tool calls at the execution layer to prevent AI agents from leaking secrets via prompt injection. The library includes a CLI scanner that red-teams agents, and tests on a local Qwen2.5 model showed it was breached 4 out of 5 times.

I've been building AI agents lately, and honestly, their security model terrifies me. We give LLMs access to powerful tools like shell.run , http.post , and fs.read . But if an agent reads a malicious email or a poisoned webpage, it can be tricked by prompt injection into using those tools to exfiltrate data. Hoping the LLM refuses the attack isn't a real security strategy. So, I built ModelFuzz. It's an open-source Python library that intercepts the tool call at the execution layer. @shield tool Instead of trying to filter prompts, ModelFuzz checks the arguments before the tool runs. If it detects a policy violation, like a stolen API key or an unsafe URL, the tool simply doesn't execute. python from modelfuzz import shield tool @shield tool def send email to address: str, subject: str, body: str - None: smtp.send to address, subject, body Even if the LLM is completely tricked by a prompt injection, the tool never fires. modelfuzz scan I also built a CLI scanner that red-teams your agent. It fires deceptive prompt injection attacks at your local model to see if it can be tricked into calling a tool. modelfuzz scan --endpoint http://localhost:11434/v1 --model qwen2.5:1.5b I tested it against a local qwen2.5:1.5b model, and it got breached 4 out of 5 times. It's 100% open source and live on PyPI. pip install "modelfuzz scan " I'd love to know what security policies you think are missing, or what agent frameworks you want supported next