Skip to content
github-actions[bot] edited this page Nov 21, 2026 · 1 revision

When personal data leaks

A personal data breach is any loss, leak or change of personal data you didn't intend: a stolen key, a database dump in the wrong place, a server someone else got into. The GDPR gives you 72 hours from when you know of it to tell your supervisory authority, unless it is unlikely to put anyone at risk (Art. 33). This page is a checklist, not legal advice.

In the first hour

  1. Stop it. Rotate what leaked: in the server's .env, AGENT_TOKEN, AUTH_SECRET (it signs everyone out), database passwords, the backup bucket's key, Cloudflare tokens. Redeploy Spotcraft. Turn off a deploy hook whose token leaked, in the app's settings.
  2. Keep what you need to understand it. Don't delete the server or its logs yet. A Hetzner snapshot keeps the disk as it is.
  3. Write down what you know, with times: what happened, when you noticed, what you did.

Within 72 hours

Work out, as well as you can:

  • What data: which apps, which tables, how many people. The Settings page's "What this instance does with personal data" lists what the platform holds; each app's own tables hold the rest.
  • Who could have it, and whether it was protected. Backups in the bucket are encrypted only by the provider, and a leaked auth schema holds session tokens, which let someone act as those guests.
  • What could happen to the people: for guests without an email, often little.

Then, unless the risk is unlikely, report to the supervisory authority set on the Settings page, through its form. What you don't know yet can follow later. If the risk to people is high, tell them too (Art. 34), plainly, in the app.

Afterwards

Keep a record of every breach, reported or not: what happened, its effects, what you did (Art. 33(5)). Fix what let it happen, and check the fix.

Clone this wiki locally