MissionPay
AVALANCHE FUJI LIVE VERIFICATION · 30-SECOND AUDIT

Judge Mode: Verify On-Chain Proof in 30 Seconds

Inspect live receipts of transactions that already occurred on Avalanche Fuji C-Chain, or simulate arbitrary off-mission attacks directly against the deployed contract via eth_call with zero gas and zero setup.

Receipts read live from Fuji RPC

Already Happened on Avalanche Fuji C-Chain

Queried directly from Avalanche Fuji RPC on page load via viem getTransactionReceipt. 100% measured on-chain.

View Twin Treasury Arena→
STEP 1✓ SUCCESS

Mission A: Drain Allowed

Global allowlist allows draining 0.01 USDC to secondary vendor

Block:#61001944
Gas Used:48,215 gas
Outflow:0.01 USDC
View on Snowtrace↗
STEP 2✗ REVERTED (BLOCKED)

Mission B (Negative): Revert Enforced

Off-mission vendor rejected on-chain with PayeeNotInMissionSet

Block:#61001947
Gas Used:32,100 gas
Outflow:0 USDC
Revert:PayeeNotInMissionSet
View on Snowtrace↗
STEP 3✓ SUCCESS

Mission B (Positive): Legit Payout Succeeded

Legitimate mission payout executed successfully to owner

Block:#61001949
Gas Used:45,320 gas
Outflow:0.01 USDC
View on Snowtrace↗
INTERACTIVE AUDIT · ZERO GAS

Try It Yourself: Interactive Fuji Contract Simulation

Enter ANY recipient address to test against Mission B (enforcing owner-signed Merkle payee root). The server executes an eth_call with account=agent against the live Fuji GuardedTreasury contract.

eth_call simulation against the live contract; nothing is sent, zero gas consumed.
Foundry cast · public RPC

Reproduce from Your Terminal (Foundry Cast)

Don't trust the web UI? Copy-paste these read-only cast call commands directly into your local terminal against public Fuji RPC.

A. Attacker Scenario (Expected Revert: PayeeNotInMissionSet)
cast call 0x20451c1e2dCD83F166B4f22868BAdECb5868d545 "proposeWithProof(bytes32,address,uint256,bytes32,string,bytes32[])" 0x0000000000000000000000000000000000000000000000000000000000000001 0xDead000000000000000000000000000000000001 10000 0x7477696e2e6d697373696f6e2e76310000000000000000000000000000000000 "attacker test" "[]" --from 0x374b8A126FDb9e5deEc8307da385076155Af3D72 --rpc-url https://api.avax-test.network/ext/bc/C/rpc
B. Positive Control Scenario (Expected: Proposal ID Success)
cast call 0x20451c1e2dCD83F166B4f22868BAdECb5868d545 "proposeWithProof(bytes32,address,uint256,bytes32,string,bytes32[])" 0x0000000000000000000000000000000000000000000000000000000000000002 0x4BC93356C180Ba2Dd13A4a1Cf43405Bd0eb58F7c 10000 0x7477696e2e6d697373696f6e2e76310000000000000000000000000000000000 "owner control" "[]" --from 0x374b8A126FDb9e5deEc8307da385076155Af3D72 --rpc-url https://api.avax-test.network/ext/bc/C/rpc

Resources & Known Limits

Honest Disclosure: Known Limits
  • Deployed Fuji Bytecode: The deployed Fuji treasury (0x2045...d545) runs original bytecode without A1 (root snapshot check) & A3 (Passkey governance). Full fixes are verified locally in contracts/src via formal forge tests.
  • Inter-Chain Messaging (ICM): Live on testnet since 20/11: an agent spend sent on Dispatch was executed by the Fuji hub (0x5641…443b); a stranger spend was rejected with UnauthorizedAgent (0xdd99…341d). Delivery was done by a public ICM relayer; our own self-relay produced a valid signed message but arrived a block later.
  • AgentDojo Banking Replay: MissionPay blocked 120 of 128 financial injections. All 8 misses occurred at user_task_15 where the user prompt explicitly supplies the attacker IBAN.