Permission Modes And Safety

Autonomy is powerful, but only when you decide how much of it to hand over. XHack AI puts that choice firmly in your hands. As a graphical desktop security agent, it lets you set exactly how hands-off it should be, tuned per chat, so you can supervise closely on a fresh target and loosen the reins once you trust the run. There is no command line to configure, just clear controls and a genuine sense of being in charge.

Four Modes, Your Call

XHack AI offers four permission modes so the level of control always fits the moment. Manual asks you to approve each action before it happens, keeping a human firmly in the loop. Allow Edits auto-approves file edits only, smoothing out routine changes while still checking the riskier stuff. Auto approves everything for that chat, ideal once you are confident. And Plan is fully read-only, letting the agent think through and lay out an approach without touching anything. You pick the posture that matches your comfort with the target.

Switch In A Single Click

Changing your safety posture never interrupts your flow. You can move between modes with one click, or use the Shift+Tab shortcut to cycle through them instantly. Because the setting is per chat, one conversation can run wide open in Auto while another stays locked down in Manual. That controlled automation means you tailor the guardrails to each engagement rather than settling for a single global setting.

Approve Risky Actions On The Spot

Even when you are supervising closely, XHack AI keeps the approval workflow simple. Any risky action can raise a clear prompt offering Allow, Always Allow, or Deny. You approve a one-off action, permanently trust a repeated one, or block it outright, all in the moment it comes up. This human-in-the-loop checkpoint keeps you aware of what the agent is about to do without slowing you to a crawl.

A Hard Safety Net That Never Turns Off

Here is the reassurance that matters most: even in fully autonomous mode, XHack AI keeps a hard safety net running underneath everything. Truly destructive commands, things like wiping disks, dropping databases, or shutting machines down, are blocked outright regardless of your permission mode. Autonomous safety is not left to chance or to a single approval click. These guardrails are always on, so full speed never means full risk.

Example: Ease In, Then Let It Run

Imagine starting on an unfamiliar target. You keep the chat in Manual and approve each step, watching how XHack AI behaves and confirming it is on the right track. Once you are satisfied, you switch that same chat to Auto with a click, and the agent runs the rest of the engagement on its own, still protected by the destructive-command safety net the whole time.

Confidence At Every Speed

This blend of flexible permission control and an unremovable safety net is what makes XHack AI comfortable to use on real systems. You decide how much to delegate, adjust it whenever you like, and know that the most dangerous actions are off the table no matter what. Whether you want careful oversight or hands-free automation, XHack AI lets you work at exactly the speed you trust.