Origin Dollar (OUSD): `CompoundStrategy.liquidate()` ignores Compound's redeem error code

Source & credit. Reproduction of a public audit finding curated by AuditVault — the original finding: 18210-lack-of-return-value-checks-can-lead-to-unexpected-results-t. Standalone Foundry PoC and full write-up: 18210-lack-of-return-value-checks-can-lead-to-unexpected-results-t_exp in the evm-hack-registry mirror.

Trail of Bits finding TOB-OUSD-019 ("Lack of return value checks can lead to unexpected results"). This PoC deploys the real audited Origin Dollar source and drives the real exploit path end-to-end in an empty local genesis.

Root cause#

CompoundStrategy.liquidate() redeems the strategy's entire cToken balance back to the underlying asset, but discards the return value of cToken.redeem:

// src/strategies/CompoundStrategy.sol  (audited commit 06ed1650, liquidate L73-L87)
function liquidate() external onlyVaultOrGovernor {
    for (uint256 i = 0; i < assetsMapped.length; i++) {
        ICERC20 cToken = _getCTokenFor(assetsMapped[i]);
        if (cToken.balanceOf(address(this)) > 0) {
            cToken.redeem(cToken.balanceOf(address(this)));   // <-- return value IGNORED (L78)
            IERC20 asset = IERC20(assetsMapped[i]);
            asset.safeTransfer(vaultAddress, asset.balanceOf(address(this)));
        }
    }
}

Compound's cToken.redeem is documented to return 0 on success and a non-zero error code otherwise (see ICompound.ICERC20.redeem), and it does not revert when a redemption fails (e.g. TOKEN_INSUFFICIENT_CASH = 9 when a highly-utilised market has no cash to pay out). Because the strategy never checks the code, a failed redemption is silently treated as success: liquidate() runs to completion without reverting while 0 underlying is actually withdrawn, then transfers the strategy's (unchanged, 0) underlying balance to the Vault. The funds stay stranded in Compound while the caller is told the strategy was emptied.

The exploit (real numbers)#

1. The strategy invests 100 USDC into Compound → it holds 5,000 cUSDC (100e6 * 1e18 / 2e14), Compound holds 100 USDC of cash.
2. The Compound market becomes illiquid — a borrower draws out all 100 USDC of cash (a normal, common state for a 100%-utilised money market).
3. The Governor calls liquidate() to pull all funds back to the Vault.
4. cToken.redeem(5000e8) finds insufficient cash and returns error code 9, without reverting and without transferring anything.
5. The strategy ignores the code and reports success.

Harm asserted (with numbers):

A fixed liquidate() (require(cToken.redeem(...) == 0, "Redeem failed"), shipped in commit ed83f5d6) would instead revert and surface the failure, preventing the protocol from treating the illiquid strategy as emptied.

Sequence#

Reproduce#

_shared/run-poc/run_poc.sh 18210-lack-of-return-value-checks-can-lead-to-unexpected-results-t_exp -vvvvv

The real audited contracts deployed by the test: src/strategies/CompoundStrategy.sol, src/utils/InitializableAbstractStrategy.sol, src/governance/Governable.sol, src/strategies/ICompound.sol — copied verbatim from Origin Dollar commit 06ed1650 (the ToB-audited version; the fix is ed83f5d6). Only the external Compound cToken (faithful error-code semantics) and the opaque USDC ERC20 are minimal stand-ins.

Sources: Origin Dollar review (Trail of Bits), AuditVault finding #18210.

Sources & further analysis#

Reproductions & code

• Standalone PoC + full trace: 18210-lack-of-return-value-checks-can-lead-to-unexpected-results-t_exp (evm-hack-registry mirror).
• AuditVault finding: 18210-lack-of-return-value-checks-can-lead-to-unexpected-results-t.
• Upstream DeFiHackLabs PoC directory: src/test.

Alerts & third-party analyses

• DeFiHackLabs incident explorer: search "Origin Dollar (OUSD): CompoundStrategy.liquidate() ignores Compound's redeem error code".
• Web3Sec X hacked database: search.
• Rekt leaderboard: search.
• Solodit incident search: search.

These dashboards index community alerts tweets, post-mortems, and independent write-ups. Reach them through the protocol name above to cross-check this reproduction against other analyses.