Brand Protection looks for things on the public internet that resemble a client's brand and might mislead its customers. It lives on the client's Reputation page, under the Brand Protection tab.
What it checks today
| Check | How | Signal |
|---|---|---|
| Look-alike domains | Common misspellings and variants of the client's domain are resolved through public DNS and registration records. | Possible impersonation when the domain was registered recently or has mail set up; otherwise Observed signal. |
| Packages | Public package registries (npm, PyPI) are searched for names that match the brand. | Observed signal; check whether the author is the brand owner. |
| Repositories | Public GitHub repositories whose name matches the brand. | Observed signal. |
Not checked: look-alike social media accounts, copied images or logos, and data leaks. Brand Protection is not a legal enforcement service and does not replace trademark advice.
How to run a scan
- Select the client.
- Open the Agent dock and choose Scan for brand impersonation signals.
- When the run finishes, open Reputation → Brand Protection.
Each finding shows its subject, severity, confidence and evidence links, and the date it was last seen. A finding is never an accusation until a person confirms it.
Reviewing and acting
- Review the evidence and decide whether the finding is a real impersonation or a coincidence.
- Ask the Agent to draft a takedown request for a finding. The draft is addressed to the right place (the domain registrar's abuse contact, npm support, GitHub support or the PyPI admins) and includes the evidence. GrowthPilot never sends it; your team reviews, edits and sends it.
- Everything you do with a finding is recorded in the client's activity.