Prepared by:
HALBORN
Last Updated 10/03/2025
Date of Engagement: September 16th, 2025 - September 17th, 2025
100% of all REPORTED Findings have been addressed
All findings
5
Critical
1
High
0
Medium
0
Low
1
Informational
3
Spiko
engaged Halborn to conduct a security assessment of the Stellar Smart Account contracts, beginning on September 16th, 2025, and ending on September 18th, 2025. This security assessment was scoped to the smart contracts located in the Spiko-Tech Stellar Contracts GitHub repository. Commit hashes and further details can be found in the Scope section of this report.
Spiko provides the on-chain backbone for tokenising real-world fund shares on Stellar. It couples a permissioned fungible-token contract (representing each share) with a central access-control registry and a redemption module that burns tokens when investors cash out in the off-chain fund. By enforcing roles, whitelisting and escrowed redemptions, the suite bridges traditional assets and blockchain while preserving regulatory and operational safeguards.
The team at Halborn assigned a full-time security engineer to verify the security of the smart contracts. The security engineer is a blockchain and smart-contract security expert with advanced penetration testing, smart-contract hacking, and deep knowledge of multiple blockchain protocols.
The purpose of this assessment is to:
Ensure that smart contract functions operate as intended
Identify potential security issues with the smart contracts
In summary, Halborn identified some improvements to reduce the likelihood and impact of risks, which have been completely addressed by the Spiko team
. The main ones were the following:
Correct the redemption execution logic so that token burns are performed from the Redemption contract balance rather than from the user account to prevent pending redemptions and locked escrowed funds.
Prevent the admin role from being renounced without an alternative recovery mechanism by overriding or replacing the default renounce_admin flow with an explicit rotation/proposal-accept process.
Ensure idempotency keys and events are only consumed/emitted when an operation produces a real state change (reject empty-batch and zero-amount operations).
Move fixed, non-parameterized initialization logic into the constructor to eliminate unnecessary public initialization entrypoints and reduce deployment sequencing risk.
Improve in-code documentation to reduce misconfiguration and improve auditability.
Halborn performed a combination of manual and automated security testing to balance efficiency, timeliness, practicality, and accuracy in regard to the scope of this assessment. Manual testing was emphasized to uncover flaws in logic, process, and contract interaction, while automated tools supported the detection of dependency vulnerabilities and unsafe coding patterns.
The following phases and associated tools were used during the assessment:
Research into the architecture, purpose, and operational model of the Stellar Smart Account system.
Manual code review and walk-through of all contracts.
Verification of initialization flows and prevention of double-init or bypass conditions.
Analysis of upgrade and migration flows.
Review of cross-contract interactions and fail-closed behavior on errors.
Scanning of Rust code for vulnerabilities and unsafe usage.
Review and improvement of integration tests.
Verification of integration test execution and addition of new ones where required.
EXPLOITABILITY METRIC () | METRIC VALUE | NUMERICAL VALUE |
---|---|---|
Attack Origin (AO) | Arbitrary (AO:A) Specific (AO:S) | 1 0.2 |
Attack Cost (AC) | Low (AC:L) Medium (AC:M) High (AC:H) | 1 0.67 0.33 |
Attack Complexity (AX) | Low (AX:L) Medium (AX:M) High (AX:H) | 1 0.67 0.33 |
IMPACT METRIC () | METRIC VALUE | NUMERICAL VALUE |
---|---|---|
Confidentiality (C) | None (C:N) Low (C:L) Medium (C:M) High (C:H) Critical (C:C) | 0 0.25 0.5 0.75 1 |
Integrity (I) | None (I:N) Low (I:L) Medium (I:M) High (I:H) Critical (I:C) | 0 0.25 0.5 0.75 1 |
Availability (A) | None (A:N) Low (A:L) Medium (A:M) High (A:H) Critical (A:C) | 0 0.25 0.5 0.75 1 |
Deposit (D) | None (D:N) Low (D:L) Medium (D:M) High (D:H) Critical (D:C) | 0 0.25 0.5 0.75 1 |
Yield (Y) | None (Y:N) Low (Y:L) Medium (Y:M) High (Y:H) Critical (Y:C) | 0 0.25 0.5 0.75 1 |
SEVERITY COEFFICIENT () | COEFFICIENT VALUE | NUMERICAL VALUE |
---|---|---|
Reversibility () | None (R:N) Partial (R:P) Full (R:F) | 1 0.5 0.25 |
Scope () | Changed (S:C) Unchanged (S:U) | 1.25 1 |
Severity | Score Value Range |
---|---|
Critical | 9 - 10 |
High | 7 - 8.9 |
Medium | 4.5 - 6.9 |
Low | 2 - 4.4 |
Informational | 0 - 1.9 |
Critical
1
High
0
Medium
0
Low
1
Informational
3
Security analysis | Risk level | Remediation Date |
---|---|---|
Redemption execution burns from user account instead of redemption contract | Critical | Solved - 09/21/2025 |
Admin can renounce and leave contract without admin | Low | Solved - 09/21/2025 |
Idempotency keys and zero-amount / empty-batch operations can be consumed without effective work | Informational | Solved - 09/21/2025 |
Initialize placed outside constructor without variable inputs | Informational | Solved - 09/21/2025 |
Insufficient in-code documentation | Informational | Solved - 09/21/2025 |
//
During redemption
execution, the contract attempts to burn the redeemed tokens from the user’s account (from
) even though those tokens were already transferred into the Redemption
contract during the initial redeem
call.
As a result, execution reverts due to insufficient user balance, leaving the redemption stuck in Pending
while tokens remain locked in the Redemption contract. If the user later holds funds, this could double-charge the user while keeping escrowed funds locked, causing supply/accounting inconsistencies.
Code of execute_redemptions
function from contracts/redemption/src/contract.rs file:
pub fn execute_redemptions(
e: &Env,
caller: Address,
operations: Vec<ExecuteRedemptionOperation>,
) {
caller.require_auth();
Self::assert_has_role(e, &caller, &REDEMPTION_EXECUTOR_ROLE);
let redemption_contract_address = e.current_contract_address();
for operation in &operations {
let token = operation.0;
let from = operation.1;
let amount = operation.2;
let salt = operation.3;
Self::assert_token_registered(e, &token);
let redemption_hash = Self::compute_redemption_hash(e, &token, &from, amount, &salt);
Self::assert_redemption_status(e, &redemption_hash, RedemptionStatus::Pending);
let client: TokenClient<'_> = TokenClient::new(e, &token);
client.burn(&from, &amount, &redemption_contract_address);
Self::set_redemption_status(e, &redemption_hash, RedemptionStatus::Executed);
e.events().publish(
(REDEMPTION_EVENT, REDEMPTION_EXECUTED_EVENT),
RedemptionEntry(token, from, amount, salt),
);
}
}
A whitelisted user redeems a positive token amount. The Token contract transfers that amount from the user to the Redemption contract (escrow) and records the entry as Pending
via Redemption.on_redeem
.
An authorized executor later calls Redemption.execute_redemptions
for that same entry.
The Redemption contract attempts to burn the redeemed amount from the user’s address instead of from the Redemption contract (where the tokens actually reside after redeem).
use token::contract::{Token, TokenClient};
fn deploy_real_token(e: &Env) -> (Address, Address, TokenClient<'_>) {
let owner: Address = Address::generate(e);
let name: String = String::from_str(e, "Token");
let symbol: String = String::from_str(e, "EUTBL");
let decimals: u32 = 6;
let token_address = e.register(Token, (owner.clone(), name, symbol, decimals));
let client = TokenClient::new(e, &token_address);
(owner, token_address, client)
}
#[test]
fn poc_execute_redemptions_fails_with_real_token() {
let e = setup_env();
// Deploy Redemption and PermissionManager
let (_owner_r, redemption_address, redemption_client) = deploy_redemption(&e);
let (admin, permission_manager_address, permission_manager_client) =
deploy_permission_manager(&e);
redemption_client.set_permission_manager(&permission_manager_address);
// Deploy real token
let (_owner_t, token_address, token_client) = deploy_real_token(&e);
token_client.set_permission_manager(&permission_manager_address);
token_client.set_redemption(&redemption_address);
// Roles
let user: Address = Address::generate(&e);
let executor: Address = Address::generate(&e);
// Whitelist user and Redemption (current design requires it)
permission_manager_client.grant_role(&admin, &user, &WHITELISTED_ROLE);
permission_manager_client.grant_role(&admin, &redemption_address, &WHITELISTED_ROLE);
// Allow minting and burning
let minter: Address = Address::generate(&e);
permission_manager_client.grant_role(&admin, &minter, &MINTER_ROLE);
permission_manager_client.grant_role(&admin, &redemption_address, &BURNER_ROLE);
// Register token on Redemption
redemption_client.add_token(&token_address);
// Mint to user and redeem (moves funds into Redemption)
let amount: i128 = 1000;
let salt: String = String::from_str(&e, "SALT");
token_client.mint(&user, &amount, &minter);
token_client.redeem(&amount, &user, &salt);
// Sanity check
assert_eq!(token_client.balance(&user), 0);
assert_eq!(token_client.balance(&redemption_address), amount);
// Execute redemptions → should revert due to burning from user instead of Redemption
let mut ops = Vec::new(&e);
ops.push_front(ExecuteRedemptionOperation(
token_address.clone(),
user.clone(),
amount,
salt.clone(),
));
permission_manager_client.grant_role(&admin, &executor, &REDEMPTION_EXECUTOR_ROLE);
let res = redemption_client.try_execute_redemptions(&executor, &ops);
assert!(res.is_err());
// Funds remain locked in Redemption
assert_eq!(token_client.balance(&redemption_address), amount);
}
Mint succeeded; redeem moved the funds to the Redemption contract (user balance = 0, Redemption balance = amount).
Calling execute_redemptions
returned an error (revert) because the burn targeted the user (insufficient balance).
Funds remained locked in the Redemption contract (Redemption balance unchanged), and the redemption status did not progress, demonstrating a DoS/lock on the redemption flow.
Burning of tokens should be performed from the Redemption contract's own balance (burn(account = Redemption, caller = Redemption)) rather than from the user's account.
SOLVED: The Spiko team has solved this issue by replacing the user's address with the contract address.
//
The contract exports AccessControl’s default renounce_admin
entrypoint. An authenticated admin
can invoke it to remove the Admin key from storage, leaving the contract without an admin.
This does not directly compromise funds but creates operational risk: future role administration (set_role_admin
, grant
/ revoke
) becomes impossible without a separate recovery path.
Code of renounce_admin
function from AccessControl crate:
/// Allows the current admin to renounce their role, making the contract
/// permanently admin-less. This is useful for decentralization purposes or when
/// the admin role is no longer needed. Once the admin is renounced, it cannot
/// be reinstated.
///
/// # Arguments
///
/// * `e` - Access to Soroban environment.
///
/// # Errors
///
/// * [`AccessControlError::AdminNotSet`] - If no admin account is set.
///
/// # Events
///
/// * topics - `["admin_renounced", admin: Address]`
/// * data - `[]`
///
/// # Notes
///
/// * Authorization for the current admin is required.
pub fn renounce_admin(e: &Env) {
let admin = enforce_admin_auth(e);
e.storage().instance().remove(&AccessControlStorageKey::Admin);
emit_admin_renounced(e, &admin);
}
It is recommended to override renounce_admin
to revert (disallow), or replace it with an explicit admin rotation flow (propose/accept) ensuring the contract cannot be left without an admin.
SOLVED: The Spiko team has solved this issue by overwriting the renounce admin function.
//
The Token contract spends idempotency keys and emits events even when the call produces no meaningful state change:
Empty batch: mint_batch
and burn_batch
consume the provided idempotency key when operations.len() == 0, leaving no balances modified but preventing re-use of that key for a genuine retry.
Zero amount: safe_transfer
, transfer
, mint
, burn
, and per-item loops in their batch variants allow amount == 0.
In safe_transfer
the key is burned, a transfer event is raised, yet no assets move.
In the other paths the call is simply a no-op, producing confusing audit logs.
While the issue is non-exploitable, it degrades UX for legitimate operators, wastes storage slots for idempotency tracking, and may break client-side retry logic that assumes keys are consumed only when the underlying effect succeeds.
It is recommended to skip idempotency-key consumption and event emission when the assertion would fail, thereby ensuring keys are burned only after real token state changes. To implement this, explicit guards should be added:
assert!(operations.len() > 0, "Empty batch")
at the start of mint_batch
and burn_batch
.
assert!(amount > 0, "Zero-amount transfer")
in safe_transfer
, transfer
, mint
, burn
, and within each batch loop.
SOLVED: The Spiko team has solved this issue by implementing the recommended fixes.
//
The initialize
function, from PermissionManager contract, only sets a fixed role relationship (WHITELISTER_ROLE
as admin of WHITELISTED_ROLE
) and does not depend on dynamic inputs.
Keeping it as a separate public entry point adds operational risk (missed initialization, sequencing mistakes) without functional benefits. Although it is admin-gated via set_role_admin
(admin.require_auth
), its separation is unnecessary and can lead to avoidable deployment complexity.
Code of initialize
function from contracts/permission-manager/src/contracts.rs file:
pub fn initialize(e: &Env) {
access_control::set_role_admin(e, &WHITELISTED_ROLE, &WHITELISTER_ROLE);
}
It is recommended to move the role-admin setup into the constructor immediately after setting the admin.
SOLVED: The Spiko team has solved this issue by relocating the set_role_admin
call into the contract constructor.
//
Multiple high-impact entrypoints lack doc comments that clearly explain each function's purpose along with required auth, invariants, storage/TTL effects, events, and idempotency semantics.
This is not an immediate exploit, but it increases misconfiguration risk and hurts auditability. Auditors depend on clear, consistent docs to verify boundaries and state changes.
It is recommended to add concise doc comments to each function covering: purpose and inputs, required authorization/roles, preconditions and invariants (e.g., amount > 0
, registered token), storage writes and TTL behavior, emitted events, expected error messages, etc.
SOLVED: The Spiko team has solved this issue by adding documentation to all critical function.
Halborn strongly recommends conducting a follow-up assessment of the project either within six months or immediately following any material changes to the codebase, whichever comes first. This approach is crucial for maintaining the project’s integrity and addressing potential vulnerabilities introduced by code modifications.
// Download the full report
Stellar Contracts
* Use Google Chrome for best results
** Check "Background Graphics" in the print settings if needed