Automata
Badges
About
Automata Network is an OP stack based Layer 2 Optimium acting as a modular attestation layer that extends machine-level trust to Ethereum with TEE Coprocessors and an EigenLayer AVS.
Badges
About
Automata Network is an OP stack based Layer 2 Optimium acting as a modular attestation layer that extends machine-level trust to Ethereum with TEE Coprocessors and an EigenLayer AVS.
Recategorisation
The project will be classified as "Other" due to its specific risks that set it apart from the standard classifications.
The project will move to Others because:
Consequence: projects without a proper proof system fully rely on single entities to safely update the state. A malicious proposer can finalize an invalid state, which can cause loss of funds.
Consequence: projects without a data availability bridge fully rely on single entities (the sequencer) to honestly rely available data roots on Ethereum. A malicious sequencer can collude with the proposer to finalize an unavailable state, which can cause loss of funds.
Learn more about the recategorisation here.
Funds can be stolen if
- an invalid state root is submitted to the system (CRITICAL),
- the sequencer is malicious and is able to economically outspend the altruistic challengers,
- there is no challenger willing to challenge unavailable data commitments,
- a contract receives a malicious code upgrade. There is no delay on code upgrades (CRITICAL).
Funds can be frozen if
MEV can be extracted if
Sequencer failure
Self sequenceState validation
NoneCurrently the system permits invalid state roots. More details in project overview.
Data availability
ExternalProof construction and state derivation rely fully on data that is NOT published onchain. A custom data availability (DA) provider without attestations is used, but data unavailability can be challenged.
Exit window
NoneThere is no window for users to exit in case of an unwanted regular upgrade since contracts are instantly upgradable.
Proposer failure
Cannot withdrawOnly the whitelisted proposers can publish state roots on L1, so in the event of failure the withdrawals are frozen.
Fraud proofs are not enabled
OP Stack projects can use the OP fault proof system, already being deployed on some. This project though is not using fault proofs yet and is relying on the honesty of the permissioned Proposer and Challengers to ensure state correctness. The smart contract system permits invalid state roots.
Funds can be stolen if an invalid state root is submitted to the system (CRITICAL).
Data required to compute fraud proof is published offchain without onchain attestations
The project relies on DA challenges for data availability. If a DA challenger finds that the data behind a tx data commitment is not available, they can submit a challenge which requires locking a bond within 12h. A challenge can be resolved by publishing the preimage data within an additional 12h. In such a case, a portion of the challenger bond is burned, with the exact amount estimated as the cost incurred by the resolver to publish the full data, meaning that the resolver and challenger will approximately lose the same amount of funds. The system is not secure if the malicious sequencer is able to outspend the altruistic challengers. If instead, after a challenge, the preimage data is not published, the chain reorgs to the last fully derivable state. This mechanism fully depends on the derivation rule of the L2 node and can only be verified in its source code, which can be reviewed here.
Funds can be stolen if the sequencer is malicious and is able to economically outspend the altruistic challengers.
Funds can be stolen if there is no challenger willing to challenge unavailable data commitments.
Automata DA is a data availability solution using data availability challenges (DA Challenges).
There are no onchain assets at risk of being slashed in case of a data withholding attack. However, there is a mechanism that allows users to challenge unavailability of data. The system is not secure if the malicious sequencer is able to outspend the altruistic challengers, and there is no pool of funds onchain to incentivize challengers.
There is no fraud detection mechanism in place. A data withholding attack can only be detected by nodes downloading the full data from the DA layer.
The committee does not meet basic security standards, either due to insufficient size, lack of member diversity, or poorly defined threshold parameters. The system lacks an effective DA bridge and it is reliant on the assumption of an honest sequencer, creating significant risks to data integrity and availability.
There is no delay in the upgradeability of the bridge. Users have no time to exit the system before the bridge implementation update is completed.
The relayer role is permissioned, and the DA bridge does not have a Security Council or a governance mechanism to propose new relayers. In case of relayer failure, the DA bridge will halt and be unable to recover without the intervention of a centralized entity.
Architecture
## Data Availability Challenges
Automata relies on DA challenges for data availability.
The DA Provider submits an input commitment on Ethereum, and users can request the data behind the commitment off-chain from the DA Provider.
If a DA challenger finds that the data behind a tx data commitment is not available, they can submit a challenge which requires locking a bond within 12h.
A challenge can be resolved by publishing the preimage data within an additional 12h.
In such case, a portion of the challenger bond is burned, with the exact amount estimated as the cost incurred by the resolver to publish the full data, meaning that the resolver and challenger will approximately lose the same amount of funds.
The system is not secure if the malicious sequencer is able to outspend the altruistic challengers.
If instead, after a challenge, the preimage data is not published, the chain reorgs to the last fully derivable state.
Only hashes of data batches are posted as DA commitments to an EOA on Ethereum.
However, there is a mechanism that allows users to challenge unavailability of data.
Funds can be lost if the sequencer posts an invalid data availability certificate and there are no challengers.
Funds can be lost if the sequencer posts an invalid data availability certificate, and he is able to outspend the challengers.
The system has a centralized operator
The operator is the only entity that can propose blocks. A live and trustworthy operator is vital to the health of the system.
MEV can be extracted if the operator exploits their centralized position and frontruns user transactions.
Users can force any transaction
Because the state of the system is based on transactions submitted on the underlying host chain and anyone can submit their transactions there it allows the users to circumvent censorship by interacting with the smart contract on the host chain directly.
Regular exit
The user initiates the withdrawal by submitting a regular transaction on this chain. When the block containing that transaction is finalized the funds become available for withdrawal on L1. The process of block finalization takes a challenge period of 7d to complete. Finally the user submits an L1 transaction to claim the funds. This transaction requires a merkle proof.
Funds can be frozen if the centralized validator goes down. Users cannot produce blocks themselves and exiting the system requires new block production (CRITICAL).
Forced exit
If the user experiences censorship from the operator with regular exit they can submit their withdrawal requests directly on L1. The system is then obliged to service this request or halt all withdrawals, including forced withdrawals from L1 and regular withdrawals initiated on L2. Once the force operation is submitted and if the request is serviced, the operation follows the flow of a regular exit.
EVM compatible smart contracts are supported
OP stack chains are pursuing the EVM Equivalence model. No changes to smart contracts are required regardless of the language they are written in, i.e. anything deployed on L1 can be deployed on L2.
The system uses the following set of permissioned addresses:
Challenger is an actor allowed to challenge or delete state roots proposed by a Proposer.
Guardian is an actor allowed to pause deposits and withdrawals.
Proposer is an actor allowed to post new state roots of the current layer to the host chain.
- A Gnosis Safe with 3 / 4 threshold.
- Can act on behalf of ProxyAdmin.
- Can change the configuration of AddressManager - set and change address mappings - acting via ProxyAdmin.
- Can upgrade the implementation of L1ERC721Bridge, DataAvailabilityChallenge, SystemConfig, OptimismMintableERC20Factory, OptimismPortal, L2OutputOracle, SuperchainConfig - acting via ProxyAdmin.
- Can upgrade the implementation of L1StandardBridge - upgrading the bridge implementation can give access to all funds escrowed therein - acting via ProxyAdmin.
A Challenger.
- Can change the configuration of DataAvailabilityChallenge - can upgrade the parameters of DA challenges like the bond size or refund percentages, potentially making challenges infeasable or insecure.
- Can change the configuration of SystemConfig - it can update the preconfer address, the batch submitter (Sequencer) address and the gas configuration of the system.
A Sequencer.
A Proposer.
A Guardian.
The system consists of the following smart contracts on the host chain (Ethereum):
Used to bridge ERC-721 tokens from host chain to this chain.
Upgrade delay: No delay
The DataAvailabilityChallenge contract is used to challenge the full availability of data behind commimted transaction data hashes. See the technology section for more details.
Upgrade delay: No delay
- Can be used to configure AddressManager - set and change address mappings.
- Can be used to upgrade implementation of L1ERC721Bridge, DataAvailabilityChallenge, SystemConfig, OptimismMintableERC20Factory, OptimismPortal, L2OutputOracle, SuperchainConfig.
- Can be used to upgrade implementation of L1StandardBridge - upgrading the bridge implementation can give access to all funds escrowed therein.
Sends messages from host chain to this chain, and relays messages back onto host chain. In the event that a message sent from host chain to this chain is rejected for exceeding this chain’s epoch gas limit, it can be resubmitted via this contract’s replay function.
A helper contract that generates OptimismMintableERC20 contracts on the network it’s deployed to. OptimismMintableERC20 is a standard extension of the base ERC20 token contract designed to allow the L1StandardBridge contracts to mint and burn tokens. This makes it possible to use an OptimismMintablERC20 as this chain’s representation of a token on the host chain, or vice-versa.
Upgrade delay: No delay
The main entry point to deposit funds from host chain to this chain. It also allows to prove and finalize withdrawals. This contract stores the following tokens: ETH, ATA.
Upgrade delay: No delay
Contains a list of proposed state roots which Proposers assert to be a result of block execution. Currently only the PROPOSER address can submit new state roots.
Upgrade delay: No delay
This is NOT the shared SuperchainConfig contract of the OP stack Superchain but rather a local fork. It manages the PAUSED_SLOT, a boolean value indicating whether the local chain is paused, and GUARDIAN_SLOT, the address of the guardian which can pause and unpause the system.
Upgrade delay: No delay
The main entry point to deposit ERC20 tokens from host chain to this chain. This contract can store any token.
Upgrade delay: No delay
Legacy contract used to manage a mapping of string names to addresses. Modern OP stack uses a different standard proxy system instead, but this contract is still necessary for backwards compatibility with several older contracts.
Value Secured is calculated based on these smart contracts and tokens:
Main entry point for users depositing ERC20 token that do not require custom gateway.
Upgrade delay: No delay
Main entry point for users depositing ETH, ATA.
Upgrade delay: No delay
The current deployment carries some associated risks:
Funds can be stolen if a contract receives a malicious code upgrade. There is no delay on code upgrades (CRITICAL).
