Need help. For jailbreak

[atifjubaer]

I connected this to my Hermes agent, and whenever I'm trying to use any vulnerability or cyber-related work, the AI directly blocks my request and says they can't do that kind of work.

Is there any way I can prompt any kind of things, like cyber-related or SQL injection, and they will execute and work? Is there any way I can already do something in the Omniroute dashboard, and it works?

[seanwall010-sys]

did you find a fix

[diegosouzapw]

Closing this one. To answer both @atifjubaer and @seanwall010-sys directly: OmniRoute will not help get around a model's refusals, and there is no dashboard setting for that.

OmniRoute is a router. It forwards requests to the provider you choose, and that provider's model decides what it will and will not do. The refusal you are seeing comes from the model, under its provider's usage policy, and bypassing it is not something this project supports.

If you do legitimate security work, such as testing systems you own or are authorized to test, describe that context plainly to the model. Many models help with defensive analysis, vulnerability explanations and remediation when the scope is clear. For offensive tooling, use purpose-built security tools under a proper engagement, not a jailbroken chat model.