> For AI agents: the documentation index is at https://docs.umbraprivacy.com/llms.txt, and every page is available as Markdown at its URL with .md appended.

# Known risks (/docs/security/risks)

These are the risks the protocol carries today, at the level of the protocol rather than any one deployment.

## Funds

1. **The program upgrade authority can replace any program at once**, bypassing every timelock, including the keys the pool checks proofs against. See [governance and upgrades](https://docs.umbraprivacy.com/docs/security/governance).
2. **A token's issuer keeps its powers.** The pool accepts tokens whatever powers their issuers hold. An issuer able to freeze or move its token could freeze or drain the pool's holdings of that token, though never of any other.
3. **The circuit proofs rest on hand-written models**, which are assumed, not proven, to match the circuits. See [audits and verification](https://docs.umbraprivacy.com/docs/security/audits).

## Privacy

4. **The indexer can link spends to identities**, because a spend's request names the spender and the notes it spends.
5. **Deposits are attributable.** A deposit's public side can be traced to the wallet that registered.

## Liveness

6. **Any one of six attesters can halt value transactions.** The escape hatch is the way out meanwhile.
7. **Without a forester, spending stops** once the staging buffer fills.
8. **Encrypted notes live off chain.** A note whose encrypted copy is lost cannot be found or spent by its recipient.
9. **Expired notes cannot be spent** by any path.

## Related

- [Trust assumptions](https://docs.umbraprivacy.com/docs/security/assumptions): What each risk rests on.
- [Governance and upgrades](https://docs.umbraprivacy.com/docs/security/governance): The upgrade authority and the timelock.
- [Liveness](https://docs.umbraprivacy.com/docs/security/liveness): What stops, and the escape hatch.
- [Audits and verification](https://docs.umbraprivacy.com/docs/security/audits): What has been checked.


Last updated 12 October 2026