Force include a transaction through the L1 inbox
The canonical liveness escape on an L2 is to route the transaction through L1, where the sequencer has
no choice but to include it after a timeout — but each rollup exposes a different contract, and the
path only works if you build the exact message the inbox expects.
Procedure
Identify the rollup family and its inbox contract before writing anything.
- Arbitrum: the delayed inbox, address
0x4Dbd4fc535Ac27206064B68FfCf827b0A60BAB3f.
- OP Stack: the
OptimismPortal deposit path (L1 to L2 always works); it is a deposit, not an
arbitrary forced call, so encode accordingly.
- zkSync Era:
requestL2Transaction on the mailbox plus the priority queue.
- Starknet: an
L1Handler message consumed on L2.
For Arbitrum, submit an unsigned L2 transaction to the delayed inbox. The sequencer must include
the delayed inbox within its grace period (24 h on Arbitrum One), after which anyone can call
forceInclusion on the SequencerInbox:
cast send 0x4Dbd4fc535Ac27206064B68FfCf827b0A60BAB3f \
"sendUnsignedTransaction(uint256,uint256,uint256,address,uint256,uint256,bytes)" \
800000 0 0 $TARGET 0 10000000000000000 0x<calldata> \
--value 0 --rpc-url $L1RPC --private-key $L1KEY
For OP Stack, use the L1 standard bridge or portal deposit; the deposit is enqueued and executed by
the sequencer on the L2, and remains executable even if the regular mempool is uncooperative:
cast send $L1_STANDARD_BRIDGE \
"depositETH(uint32,bytes)" 200000 0x --value 1ether --rpc-url $L1RPC
Record the L1 receipt and the message/leaf that was enqueued. For Arbitrum note the InboxMessageDelivered
event and the ticket index; for OP note the TransactionDeposited event's deposit hash.
Poll for inclusion on the L2. If the sequencer honours the inbox, it appears within its normal
latency; if not, wait out the grace period and trigger forced inclusion.
Budget the cost honestly: forced inclusion pays full L1 gas plus the L2 execution cost, and the
calldata is not compressed the way the batcher's is, so a forced call is often an order of
magnitude more expensive than a normal L2 transaction.
Log the transaction hashes on both L1 and L2 and the wall-clock latency so the escape's real
speed is known for the next incident.
Pitfalls
- Sending to the regular bridge when the sequencer is censoring your address — the deposit path is
processed by the same sequencer, so verify the inbox contract is the force-inclusion path.
- Getting the gas limit or refund fields wrong on an Arbitrum unsigned transaction; the ticket can be
created but fail to redeem, leaving funds stuck in the retryable.
- Assuming forced inclusion bypasses the L2 fee — it does not; the L2 execution still needs gas.
- Using an L1 account with insufficient ETH for L1 gas; the escape is gas-expensive exactly when L1 is
congested.
- Forgetting that some chains (OP Stack historically) do not offer arbitrary forced L2 calls at all —
only deposits — so a "forced transfer" for an existing L2 balance may be impossible.
Verification
cast receipt $L1_TXHASH --rpc-url $L1RPC --field status
cast tx $L2_SEARCH_HASH --rpc-url $L2RPC # once included, or read the queued message on L1
# Arbitrum: the delayed-inbox message index advanced and, if needed, forceInclusion succeeded
Report the inbox contract used, the L1 transaction hash, the grace-period timer, and whether the
transaction executed on the L2 within it.
1---2name: force-include-a-transaction-through-l1-inbox3description: Use when a sequencer is censoring or halting your transaction and you need to force it onto the L2 by submitting through the L1 inbox or deposit contract.4metadata:5 source: https://skillmd.com/skills/jiayaoqijia/force-include-a-transaction-through-l1-inbox6---78# Force include a transaction through the L1 inbox910The canonical liveness escape on an L2 is to route the transaction through L1, where the sequencer has11no choice but to include it after a timeout — but each rollup exposes a different contract, and the12path only works if you build the exact message the inbox expects.1314## Procedure15161. Identify the rollup family and its inbox contract before writing anything.17 - Arbitrum: the delayed inbox, address `0x4Dbd4fc535Ac27206064B68FfCf827b0A60BAB3f`.18 - OP Stack: the `OptimismPortal` deposit path (L1 to L2 always works); it is a deposit, not an19 arbitrary forced call, so encode accordingly.20 - zkSync Era: `requestL2Transaction` on the mailbox plus the priority queue.21 - Starknet: an `L1Handler` message consumed on L2.22232. For Arbitrum, submit an unsigned L2 transaction to the delayed inbox. The sequencer must include24 the delayed inbox within its grace period (24 h on Arbitrum One), after which anyone can call25 `forceInclusion` on the `SequencerInbox`:2627 cast send 0x4Dbd4fc535Ac27206064B68FfCf827b0A60BAB3f \28 "sendUnsignedTransaction(uint256,uint256,uint256,address,uint256,uint256,bytes)" \29 800000 0 0 $TARGET 0 10000000000000000 0x<calldata> \30 --value 0 --rpc-url $L1RPC --private-key $L1KEY31323. For OP Stack, use the L1 standard bridge or portal deposit; the deposit is enqueued and executed by33 the sequencer on the L2, and remains executable even if the regular mempool is uncooperative:3435 cast send $L1_STANDARD_BRIDGE \36 "depositETH(uint32,bytes)" 200000 0x --value 1ether --rpc-url $L1RPC37384. Record the L1 receipt and the message/leaf that was enqueued. For Arbitrum note the `InboxMessageDelivered`39 event and the ticket index; for OP note the `TransactionDeposited` event's deposit hash.40415. Poll for inclusion on the L2. If the sequencer honours the inbox, it appears within its normal42 latency; if not, wait out the grace period and trigger forced inclusion.43446. Budget the cost honestly: forced inclusion pays full L1 gas plus the L2 execution cost, and the45 calldata is not compressed the way the batcher's is, so a forced call is often an order of46 magnitude more expensive than a normal L2 transaction.47487. Log the transaction hashes on both L1 and L2 and the wall-clock latency so the escape's real49 speed is known for the next incident.5051## Pitfalls5253- Sending to the regular bridge when the sequencer is censoring *your address* — the deposit path is54 processed by the same sequencer, so verify the inbox contract is the force-inclusion path.55- Getting the gas limit or refund fields wrong on an Arbitrum unsigned transaction; the ticket can be56 created but fail to redeem, leaving funds stuck in the retryable.57- Assuming forced inclusion bypasses the L2 fee — it does not; the L2 execution still needs gas.58- Using an L1 account with insufficient ETH for L1 gas; the escape is gas-expensive exactly when L1 is59 congested.60- Forgetting that some chains (OP Stack historically) do not offer arbitrary forced L2 calls at all —61 only deposits — so a "forced transfer" for an existing L2 balance may be impossible.6263## Verification6465 cast receipt $L1_TXHASH --rpc-url $L1RPC --field status66 cast tx $L2_SEARCH_HASH --rpc-url $L2RPC # once included, or read the queued message on L167 # Arbitrum: the delayed-inbox message index advanced and, if needed, forceInclusion succeeded6869Report the inbox contract used, the L1 transaction hash, the grace-period timer, and whether the70transaction executed on the L2 within it.