Safety & privacy

Understand project trust, incomplete dry runs, stopping, masking and the limits of the evidence you share.

Trust a project deliberately

Opening an unfamiliar folder runs nothing from it until you agree. Review the trust prompt and setup checklist, including what the project would run on your Mac. BareTold asks again when the trusted tools change.

Check is incomplete

Check shows what Ansible predicts. Some steps are skipped, and others depend on earlier changes that a dry run has not made. Read the notes about what cannot be predicted.

Some tasks can make real changes even during a dry run. Review your playbooks before pressing Check. A prediction is not a guarantee of what will happen on the host.

Apply is a new run

Apply re-runs the playbook. It cannot apply a saved plan, and its result may differ from Check. Review both the predicted changes and the limits before choosing Apply. Changes not predicted are flagged as Ansible reports them.

Stop is an interrupt

Stop now interrupts the run. A step may still have finished on the host. Check again to find out its current state; stopping does not undo changes already made.

Masked views and outside logs

BareTold masks secrets in its change views, raw log and history. Some masked values cannot be revealed. Missing or hidden details remain limits on what you can learn from a run.

Ansible may also write logs outside BareTold. The app cannot hide secrets in those copies. Review your playbooks and output as you normally would.

Review before sharing

Masked does not mean anonymous. Reports and copied output can still identify projects, hosts or people. Review anything you export or send for help, and never send credentials or secret values.

If you choose Ask a model, review the text with sensitive details removed before sharing it. Read the privacy note for the app’s connections and data practices. For a practical starting point, use the quick-start help or Guide.

Beta opens soon