Defend your box
Read your service checks, patch without breaking your services, and restore your box from its clean snapshot.
Defending in attack/defence means two things at once: stop other teams from stealing your flags, and keep every service working for the checker. A service that’s down scores nothing, however secure it is.
Read your service checks
My box lists every service and flag store with its result for the last completed tick:
| Status | Meaning |
|---|---|
| Up | The service passed its checks. |
| Down | The service didn’t respond successfully before the deadline. |
| Faulty | The service failed a functional check. |
| Flag missing | A flag from its validity window couldn’t be retrieved. |
| Recovering | Current checks pass, but a check for an earlier tick failed. |
Stores are scored separately, so a service can be Up in one store and Flag missing in another. You see these summaries only; the checker’s own messages stay with the organizers.
Patch safely
- The checker uses your service as a normal client would: registering, logging in, storing, listing, searching. A patch that removes or breaks any of that makes the service Faulty, even if it also stops an exploit.
- The checker reaches your box from changing addresses in the same range players use. Blocking by source address blocks it too.
- Firewalling is allowed: rate limits, payload filters and blocking specific exploits are part of the game. Refusing every other player while answering only the checker isn’t, and is detected per port: a service caught doing it on one of its ports scores Down in every flag store for that tick.
- Keep flags where the service puts them. Deleting or moving stored data makes the checker’s reads fail.
- Your box has no internet access. Bring any tool you need with you.
Restore from the snapshot
If you break a service beyond repair, click Restore from snapshot on My box and confirm. Your box goes back to how it was when the network opened:
- every change you made since then is lost, including your patches;
- every flag store on your box earns no SLA and no defence points in any tick the restore overlaps, from the moment you ask until the box is healthy again, even when the restore spans several ticks;
- if the restore fails, that lasts until your box is healthy again;
- flags stored before the restore are gone, so checks reading them back fail until they expire;
- flags other teams already stole stay stolen, captures you made keep scoring, and earlier scores don’t change.
Your box keeps its address and its SSH host key, and your team’s current SSH keys are put back after the restore. My box shows the restore’s progress and its result. One restore runs at a time.
A restore is refused before the network opens (there’s no snapshot yet), while the event host is unreachable, while another restore is running, and once the event has ended. The message says which.