> ## Documentation Index
> Fetch the complete documentation index at: https://docs.x0protocol.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Audit Status

> Current audit status, known findings, and security checklist for the x0 protocol.

## Audit Status

<Warning>
  The x0 protocol has **not yet undergone a third-party security audit**. The protocol is currently on devnet and should not be used with real assets on mainnet until audited.
</Warning>

## Internal Security Review

The codebase has been through an internal security review that identified and addressed findings across multiple severity levels. The following categories of findings have been resolved:

### HIGH Severity — Resolved

| ID     | Finding                               | Resolution                                                                                    |
| ------ | ------------------------------------- | --------------------------------------------------------------------------------------------- |
| HIGH-1 | Clock manipulation via `Clock::get()` | Added slot-based dual checks with buffers (`ROLLING_WINDOW_SLOTS`, `TIME_CHECK_BUFFER_SLOTS`) |

### MEDIUM Severity — Resolved

| ID        | Finding                                 | Resolution                                                              |
| --------- | --------------------------------------- | ----------------------------------------------------------------------- |
| MEDIUM-2  | Policy update spam                      | Added `POLICY_UPDATE_COOLDOWN_SLOTS` (750 slots, \~5 min) rate limiting |
| MEDIUM-6  | Arbiter rush resolution                 | Added `ARBITER_RESOLUTION_DELAY_SLOTS` (216,000 slots, \~24h) delay     |
| MEDIUM-8  | No per-transaction limit                | Added `maxSingleTransaction` field to `AgentPolicy`                     |
| MEDIUM-9  | Reputation gaming with few transactions | Added `MIN_TRANSACTIONS_FOR_REPUTATION` (10) threshold                  |
| MEDIUM-10 | Extra account metas re-initialization   | Added `ExtraMetasAlreadyInitialized` check with authority validation    |
| MEDIUM-12 | Dust transfer spam                      | Added `MIN_TRANSFER_AMOUNT` (100 micro-units) threshold                 |

### LOW Severity — Resolved

| ID    | Finding                           | Resolution                                              |
| ----- | --------------------------------- | ------------------------------------------------------- |
| LOW-2 | Unbounded string fields in events | Added max length constraints on all string event fields |
| LOW-3 | Error code numbering scheme       | Established structured numbering (0x1100–0x1700 ranges) |

## Security Checklist

### On-Chain Programs

* [x] **Signer verification** — All instructions verify that the expected authority has signed
* [x] **Account ownership checks** — All accounts verified against expected owning programs
* [x] **PDA seed validation** — All PDAs derived with canonical seeds and bumps
* [x] **Integer overflow protection** — All arithmetic uses checked operations or `BN`
* [x] **Re-initialization prevention** — All `initialize_*` instructions check for existing accounts
* [x] **Access control** — Owner-only, buyer-only, seller-only, admin-only checks on all sensitive operations
* [x] **State machine enforcement** — Escrow state transitions validated
* [x] **Rate limiting** — Policy updates, Blink generation, and daily redemptions are rate-limited
* [x] **Timelock governance** — All admin actions go through 48-hour timelock
* [x] **Reserve invariant** — Checked on every deposit and redemption
* [x] **Transfer hook enforcement** — Guard validates every transfer through the Token-2022 hook
* [x] **ZK proof validation** — Proof contexts checked for freshness, ownership, and parameter matching
* [x] **Delegation model** — Agent must be delegate; self-delegation prevention; bound account enforcement

### SDK

* [x] **Input validation** — All public functions validate parameters before building instructions
* [x] **Error handling** — All errors have descriptive messages and error codes
* [x] **Type safety** — Full TypeScript type coverage
* [x] **Deterministic key derivation** — ElGamal and AES keys derived deterministically from owner keypair + mint
* [x] **WASM proof generation** — ZK proofs generated client-side, never exposing secret keys to the network

### Testing

* [x] **47 native unit tests** for x0-zk-proofs (Rust)
* [x] **14 WASM integration tests** for browser/Node.js proof generation
* [x] **SDK-to-core consistency** — All discriminators, constants, and type definitions verified against on-chain programs

## Planned Audit Scope

When the protocol proceeds to audit, the following areas are in scope:

1. **x0-guard** — Transfer hook logic, spend limit enforcement, whitelist verification, delegation model
2. **x0-escrow** — State machine transitions, timeout handling, CPI to reputation
3. **x0-wrapper** — Reserve invariant, timelock governance, fee calculation, emergency operations
4. **x0-token** — Mint initialization, Token-2022 extension configuration, confidential transfer setup
5. **x0-zk-verifier** — Proof verification logic, context freshness, parameter binding
6. **x0-registry** — Registration, deregistration, capability metadata
7. **x0-reputation** — Score calculation, temporal decay, CPI authorization
8. **x0-zk-proofs** — Rust cryptographic primitives, WASM bindings
9. **Cross-program interactions** — CPI call graph, account sharing, PDA ownership

## Responsible Disclosure

If you discover a security vulnerability in the x0 protocol, please report it responsibly:

* **Do not** open a public GitHub issue
* Contact the team directly via the channels listed in the project repository
* Allow reasonable time for the team to investigate and patch before public disclosure