Panic mode
Emergency vault destruction from the GUI.
Panic mode overwrites a vault's header with random bytes so the encrypted content becomes mathematically unrecoverable. Use this when you need to destroy a vault quickly. Open it from the sidebar with PANIC: destroy a vault.... There is no undo.

What panic mode does
Fill in the Target section: the vault path (or Browse...), and, if the vault uses a detached header, tick This vault uses a detached header sidecar and give the header's path.
| Choice | What it does | When to pick |
|---|---|---|
| Default | Overwrites the 8 KiB header (or the detached .hdr file) with random bytes, and overwrites the header's crash-recovery copy (.header-bak next to it) completely, so the header cannot be rebuilt from it. The encrypted data stays on disk but is unreadable. |
Fast destruction; takes the same time whatever the vault size |
| ALSO overwrite the entire vault data area with random (slow on large vaults) | The above, plus overwrites the whole vault file and its metadata copy (.meta-bak) with random bytes. The file keeps its size. |
Full destruction; takes time proportional to vault size |
On success the GUI says "vault destroyed", returns to the Welcome view and removes the vault from the recent-vaults list.
Why this is mathematically unrecoverable
Without the 8 KiB header you have:
- No magic bytes (
LUKSBOX2/LUKSBOX3) to confirm the file is a vault - No salt
- No keyslot table (no wrapped MVK ciphertexts to brute-force)
- No KDF parameters
- No header HMAC (so even if you guessed the MVK, no way to verify)
- No
cipher_suite/kdf_idfields
Even an attacker with infinite compute can't unlock a header-less vault because there's nothing to attack - the wrapped MVK ciphertexts that would normally take Argon2id-bounded time to brute-force are GONE.
The confirmation gate
The Destroy button stays disabled until you:
- Enter the vault path
- Type DESTROY, a space and the exact vault path in the
Confirm field, as the form shows it (for example
DESTROY /home/you/vault.lbx)
This isn't there to make you feel safer; it's there because panic
mode has no undo. If you accidentally panic the wrong vault, your
only recovery is a header backup made beforehand with
luksbox header-backup (which we strongly
recommend keeping on separate trusted storage), and only if you did
not overwrite the data area too.
Use cases
| Situation | Recommended panic mode |
|---|---|
| Selling / scrapping a hard drive | Tick ALSO overwrite the entire vault data area |
| Border crossing with sensitive data | Header only (fast); restore from a header backup later |
| Compromised device + you have time | Tick ALSO overwrite the entire vault data area |
| Compromised device + URGENT (panic-button scenario) | Header only - fastest path to "unrecoverable" |
| Suspected insider threat | Header only, with a header backup kept elsewhere; investigate then restore |
What survives panic
Panic does NOT touch:
- Other files next to the vault: the
.anchorsidecar, the.hybridsidecar, the.kyberseed file. Delete them yourself if they must not show that a vault existed. - Header backups you made with
luksbox header-backup. - OS-level file metadata (creation time, last-access time)
- Forensic recovery of the original header bytes from disk slack or filesystem journal if you only overwrite the header and the filesystem journals deletions (tick ALSO overwrite the entire vault data area on journaled FSes for stronger destruction). SSDs and copy-on-write filesystems may also keep old copies of overwritten blocks.
Next
- Back to GUI overview ->
- CLI equivalent:
luksbox panic-> (--header,--wipe-data; no GUI confirmations,-yskips the prompt)