Legal
Incident and Breach Notification Policy
Version 1.0 · Last updated: October 26, 2026 · Next review: by October 26, 2027
Owner: Andrew Miller, Rebel Ops
This policy explains how RebelSites handles a security incident that affects a hosted client site, and how we tell you about it.
This policy describes our practice. It adds to our Privacy Policy and Terms of Service and does not replace them. If a law or your contract requires faster notice, the faster deadline applies.
1. Scope
A security incident is any event where client data, a client site, or the accounts and domain settings that control it may have been accessed, changed or exposed without authorization. Examples are a compromised sign-in, defaced or changed site content, malicious code served from a site, a hijacked domain or DNS record, and exposed form submissions.
A service outage is a site that is down or slow with no sign of unauthorized access. We handle an outage as an operations matter and tell affected clients by email. This policy sets no uptime commitment.
What we hold for a client site
The Privacy Policy also describes what we keep about visitors to a client site.
- Form submissions. We keep every field the visitor sends, with the visitor’s internet address, browser type and referring page. We store them on the platform until the site is deleted. Messages held as spam are deleted after 30 days. Members of the workspace can view and export the submissions. Contact-form submissions are also emailed to the site owner.
- Owner account data, as listed in section 2 of the Privacy Policy.
- The site itself: its source, content and media.
- Visit counts. We keep a short one-day code made from the visitor’s address and browser. We never keep the address itself. Our analytics provider deletes these records after three months.
How a site is served and edited
- A published site serves pages built in advance. Loading a page runs no database query. A form post is stored in the platform database.
- Editing happens in the RebelSites builder and content editor, behind a sign-in. An AI assistant that the owner connects edits with a token the owner granted. Authorized staff can edit a site for support.
- Each site also has a review address on rebelsites.ai that ends in the site’s id. Anyone with that address can see the last built copy without signing in. This is true even for a site that was never published or that we took offline. The review address cannot be used to edit. When we take a site offline during an incident, our response covers its review address too.
2. Security contact
Report a suspected incident or security problem to hello@rebelops.ai. Clients, site visitors and security researchers can all use this address.
3. Detection and triage
We find out about problems through these sources:
- Checks on every build and publish. A failed build or a failed publish check is caught, and a broken publish is blocked.
- A check every 15 minutes for saved content that has not reached the live site.
- A check every 15 minutes that each live site answers. We record the result internally. It does not yet send an automatic message to a person.
- Security events, such as refused requests and exceeded rate limits. These are written to the platform’s request log.
- Notices from Cloudflare and our other providers.
- Reports from clients, site visitors and researchers.
We use no outside error-reporting service. None of these checks detects a break-in by itself. Our staff watch for problems and act on every report.
We give every report a severity level:
| Level | Meaning | Examples |
|---|---|---|
| 1 – Critical | Confirmed unauthorized access to or exposure of client data, or someone other than the client controls the site, its content or its domain. | Leaked form submissions, defacement, hijacked DNS, malicious code served |
| 2 – High | Credible risk with no confirmed harm, or a confirmed issue contained to a single account. | A compromised credential, or a weakness we have no evidence anyone used |
| 3 – Low | Contained, with no effect on client data or site integrity. | Blocked attacks, a flaw fixed before anyone exploited it |
4. Notification commitment
- Who: the workspace owner, any member whose own account is affected, and the incident contact you named, where you have given us one.
- Channel: email. Where useful, we also post a notice on the service.
- When: within 72 hours. The clock starts when we become aware of an incident that affects, or likely affects, your data or site.
- Severity 2: we email affected clients within the same 72 hours, counted from when we become aware that their site or data is, or is likely, affected.
- Severity 3: we log it, and share the details on request.
- We send what we know, even if the picture is incomplete, and we say what we do not know yet.
Where the law requires notice to authorities or to the visitors of your site, we tell you what we know so you can meet that duty. We notify authorities where the law places that duty on us.
5. What a notice contains
- What happened
- What was affected (data, sites, time range)
- What we did and are doing
- What you should do, if anything
- When you will hear from us next
6. Upstream providers
Our sites, data and builder run on Cloudflare. Under its data processing terms, Cloudflare notifies us “without undue delay” of a security breach. It sets no fixed number of hours.
Two other providers touch custom domains. Vercel redirects a domain without “www” to the “www” address. A domain-setup provider writes the DNS records when you connect a domain. A domain hijack can involve either of them.
When a notice from any provider touches your site or data, we pass it on to you under section 4.
7. After the incident
We send a written summary within 14 business days of resolution. It covers a timeline, the impact, the root cause and the fix. We record the lessons and apply them to the platform and to this policy.
8. Review
Andrew Miller reviews this policy once a year, and after any Severity 1 incident.