Future request:

[ksecurity45]

Hi Team,

We’re currently exploring options for building an automated security scanning workflow for AI models. We’re evaluating Osemdus, but we wanted to ask if you would recommend something better or more suitable for this use case.

Our goal is to create a workflow where we can integrate multiple open-source AI security scanners and run them automatically as part of a unified pipeline for model evaluation.

The scanners we are considering integrating include:

ModelScan

PickleScan

Promptfoo (model audit features)

We’d like to design an automated flow where:

A model is submitted

All configured scanners are triggered

Results are aggregated

The output feeds into our AI risk assessment process

Do you think Osemdus is the right tool for orchestrating this? Or would you recommend a different framework or workflow orchestration system that would better support this kind of automated multi-scanner pipeline?

[j3ssie]

Hi @ksecurity45,

If you're looking for an orchestration engine with YAML support and fully customizable conditional steps, Osmedeus would be a great fit. Since my main focus is security automation, some features may feel a bit redundant depending on your specific use case.

That said, Osmedeus includes native support for HTTP, LLM, and Agent steps out of the box. I’ve also recently added ACP support (see commit: 520977d), which makes it an even stronger option to consider.

If you’d like more details, feel free to reach out to me on X: https://x.com/j3ssie