NGP · Baseline
Original contract
The code as it was exploited. The attack still works.
This is the contract as it was exploited: the historical exploit still succeeds against the frozen state.
What ran, and what it said.
Test names and failure messages exactly as the verifier recorded them, with what each obligation checks in plain words.
Exploit
1 replay · profit over 100,000 USDT- ✗The DeFiHackLabs reproduction of the attack: landsIt still clears the case’s profit threshold (100,000 USDT).test/poc.t.sol:NGP_PoC:testExploit()
Security obligations
0 / 4 pass- ✗A direct transfer to the pool moves nothing out of itCatches a fix that special-cases the router and leaves the same branch open.test/hidden/security.t.sol:NGP_Security:test_direct_transfer_to_the_pair_moves_nothing_out_of_it()a direct transfer to the pair moved NGP OUT of it: the pool is still being drained inside someone else's transfer: 3500000000000000000000 != 0
- ✗An ordinary sell moves no tokens out of the poolThe original moved 70% of every sell out of the pair.test/hidden/security.t.sol:NGP_Security:test_ordinary_sell_moves_no_tokens_out_of_the_pair()a sell moved NGP OUT of the pair: the seller is draining the pool inside their own transfer (amm-reserve-manipulation): 3630547125325919886628 != 0
- ✗The pool keeps every token a sell sends itAsserted on balances, so a fix that only hides the transfer events is still caught.test/hidden/security.t.sol:NGP_Security:test_pair_keeps_every_token_an_ordinary_sell_sends_it()the pair did not keep everything the sell sent it: tokens leaked out of the pool during the seller's transfer: 2593247946661371347594 != 6223795071987291234222
- ✗Treasury and reward fees are not paid out of the poolFees must be funded by the party making the trade.test/hidden/security.t.sol:NGP_Security:test_treasury_and_reward_are_not_funded_out_of_the_pair()the treasury fee on a sell was taken out of the pair, not out of the seller: 2593247946661371347592 != 0
Legitimate use · behaviour
5 / 5 pass- ✓A buy-then-sell round trip through the pool worksThe sell leg is where the bug lived, so deleting selling fails here.test/hidden/amm_regression.t.sol:NGP_AmmRegression:test_amm_buy_then_sell_round_trip()
- ✓Trading stays switched onChecks the buy and sell switches in storage. Turning either off is not a fix.test/hidden/amm_regression.t.sol:NGP_AmmRegression:test_trading_flags_still_open()
- ✓Token metadata and the pool’s balance are intactDecimals still 18, supply present, the pool still holds NGP.test/hidden/regression.t.sol:NGP_Regression:test_metadata_preserved()
- ✓A wallet-to-wallet transfer workstest/hidden/regression.t.sol:NGP_Regression:test_normal_transfer_works()
- ✓The price view still workstest/hidden/regression.t.sol:NGP_Regression:test_price_view_works()
Legitimate use · interface
5 / 5 pass- ✓Original functions still answerCalls the original functions and requires an answer other than “no such function”.test/hidden/invariants_auto.t.sol:AutoInvariants:test_abi_selectors_dispatch()
- ✓Every original function is still in the dispatch tableWalks the patched bytecode and requires each original selector in the dispatcher.test/hidden/invariants_auto.t.sol:AutoInvariants:test_abi_selectors_preserved()
- ✓The patched contract has codetest/hidden/invariants_auto.t.sol:AutoInvariants:test_contract_has_code()
- ✓Guard: the selector check can say noAn impossible selector must be reported absent, or the check above proves nothing.test/hidden/invariants_auto.t.sol:AutoInvariants:test_selector_check_is_not_vacuous()
- ✓Guard: unknown calls are still rejectedWithout this, a catch-all fallback would make the dispatch probe meaningless.test/hidden/invariants_auto.t.sol:AutoInvariants:test_unknown_selector_is_rejected()
The patch
against the original sourceRe-run this grade
offline · same inputsNeeds Foundry 1.7.1 with solc 0.8.30, 0.8.26 and 0.8.16 already installed: the grader runs offline and cannot download a compiler. Python 3.12 or later. Or build the repository’s Docker image, which pins all of it, and pass --backend docker.
$ git clone https://github.com/FarseenSh/evmpatch-env.git && cd evmpatch-env $ git checkout 165c0ed $ python -m evmpatch_env.sandbox tasks/ngp_2025_09 --backend local --sha256
Expected output: core 25b53d32310daa49483bc7f910779928be9c4b3639bcd78029c391de1cab77df and strict a79700ceb81bcee8c9894ba3413dd9a9d5d58691ec8163fea4f60b7a2d9f17f3.
To check a downloaded grade file instead: shasum -a 256 grade.strict.json prints the strict hash.
Control note
from the repositoryShipped vulnerable source, no patch. The exploit must still reproduce, every legitimate-use hidden test must pass, and every declared security obligation must FAIL (that is what makes them able to separate a repair from a non-repair).