ERC-6900: Modular Smart Contract Accounts

Abstract

This proposal standardizes smart contract accounts and account modules, which are smart contracts that allow for composable logic within smart contract accounts. This proposal is compliant with ERC-4337. This standard emphasizes secure permissioning of modules, and maximal interoperability between all spec-compliant accounts and modules.

This modular approach splits account functionality into three categories, implements them in external contracts, and defines an expected execution flow from accounts.

Motivation

One of the goals that ERC-4337 accomplishes is abstracting the logic for execution and validation to each smart contract account.

Many new features of accounts can be built by customizing the logic that goes into the validation and execution steps. Examples of such features include session keys, subscriptions, spending limits, and role-based access control. Currently, some of these features are implemented natively by specific smart contract accounts, and others are able to be implemented by proprietary module systems like Safe modules.

However, managing multiple account implementations provides a poor user experience, fragmenting accounts across supported features and security configurations. Additionally, it requires module developers to choose which platforms to support, causing either platform lock-in or duplicated development effort.

We propose a standard that coordinates the implementation work between module developers and account developers. This standard defines a modular smart contract account capable of supporting all standard-conformant modules. This allows users to have greater portability of their data, and for module developers to not have to choose specific account implementations to support.

These modules can contain execution logic, validation functions, and hooks. Validation functions define the circumstances under which the smart contract account will approve actions taken on its behalf, while hooks allow for pre and post execution controls.

Accounts adopting this standard will support modular, upgradable execution and validation logic. Defining this as a standard for smart contract accounts will make modules easier to develop securely and will allow for greater interoperability.

Specification

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119 and RFC 8174.

Terms

• An account (or smart contract account, SCA) is a smart contract that can be used to send transactions and hold digital assets. It implements the IAccount interface from ERC-4337.
• A modular account (or modular smart contract account, MSCA) is an account that supports modular functions. There are three types of modular functions:
  • Validation functions validate authorization on behalf of the account.
  • Execution functions execute custom logic allowed by the account.
  • Hooks execute custom logic and checks before and/or after an execution function or validation function. There are two types of hooks:
    • Validation hooks run before a validation function. These can enforce permissions on actions authorized by a validation function.
    • Execution hooks can run before and/or after an execution function. Execution hooks can be attached either to a specific execution function or a validation function. A pre execution hook may optionally return data to be consumed by a post execution hook.
• A native function refers to a function implemented by the modular account, as opposed to a function added by a module.
• A module is a deployed smart contract that hosts any amount of the above three kinds of modular functions.
• A module's manifest describes the execution functions, interface IDs, and hooks that should be installed on the account.

Overview

A modular account handles two kinds of calls: either from the EntryPoint through ERC-4337, or through direct calls from externally owned accounts (EOAs) and other smart contracts. This standard supports both use cases.

A call to the modular account can be broken down into the steps as shown in the diagram below. The validation steps validate if the caller is allowed to perform the call. The pre execution hook step can be used to do any pre execution checks or updates. It can also be used along with the post execution hook step to perform additional actions or verification. The execution step performs a defined task or collection of tasks.

Each step is modular, supporting different implementations, which allows for open-ended programmable accounts.

Interfaces

Modular accounts MUST implement:

• IAccount.sol and IAccountExecute.sol from ERC-4337.
• IERC6900Account.sol to support module management and usage, and account identification.
• The function isValidSignature from ERC-1271

Modular accounts MAY implement:

• IERC6900AccountView.sol to support visibility in account states on-chain.
• ERC-165 for interfaces installed from modules.

Modules MUST implement:

• IERC6900Module.sol described below and implement ERC-165 for IERC6900Module.

Modules MAY implement any of the following module types:

• IERC6900ValidationModule to support validation functions for the account.
• IERC6900ValidationHookModule to support hooks for validation functions.
• IERC6900ExecutionModule to support execution functions and their installations on the account.
• IERC6900ExecutionHookModule to support pre & post execution hooks for execution functions.

IERC6900Account.sol

Module execution and management interface. Modular accounts MUST implement this interface to support installing and uninstalling modules, and open-ended execution.

IERC6900AccountView.sol

Module inspection interface. Modular accounts MAY implement this interface to support visibility in module configuration.

IERC6900Module.sol

Module interface. Modules MUST implement this interface to support module management and interactions with ERC-6900 modular accounts.

IERC6900ValidationModule.sol

Validation module interface. Modules MAY implement this interface to provide validation functions for the account.

IERC6900ValidationHookModule.sol

Validation hook module interface. Modules MAY implement this interface to provide hooks for validation functions for the account.

IERC6900ExecutionModule.sol

Execution module interface. Modules MAY implement this interface to provide execution functions for the account.

IERC6900ExecutionHookModule.sol

Execution hook module interface. Modules MAY implement this interface to provide hooks for execution functions for the account.

Validation Functions

Installation

• The account MUST install all validation hooks specified by the user and SHOULD call onInstall with the user-provided data on the hook module to initialize state if specified by the user.
• The account MUST install all execution hooks specified by the user and SHOULD call onInstall with the user-provided data on the hook module to initialize state if specified by the user.
• The account MUST configure the validation function to validate all of the selectors specified by the user.
• The account MUST set all flags as specified, like isGlobal, isSignatureValidation, and isUserOpValidation.
• The account SHOULD call onInstall on the validation module to initialize state if specified by the user.
• The account MUST emit ValidationInstalled as defined in the interface for all installed validation functions.

Uninstallation

During validation uninstallation, the account MUST correctly clear flags and other fields based on the incoming data provided by the user.

• The account MUST clear all flags for the validation function, like isGlobal, isSignatureValidation, and isUserOpValidation.
• The account MUST remove all hooks and SHOULD clear hook module states by calling onUninstall with the user-provided data for each hook, including both validation hooks and execution hooks, if specified by the user.
  • The account MAY ignore the revert from onUninstall with try/catch depending on the design principle of the account.
• The account MUST clear the configuration for the selectors that the validation function can validate.
• The account SHOULD call onUninstall on the validation module to clean up state if specified by the user.
• The account MUST emit ValidationUninstalled as defined in the interface for all uninstalled validation functions.

Execution Functions

Installation

• The account MUST install all execution functions and set flags and fields as specified in the manifest.
  • An execution function selector MUST be unique in the account.
  • An execution function selector MUST not conflict with native ERC-4337 and ERC-6900 functions.
• The account MUST add all execution hooks as specified in the manifest.
• The account SHOULD add all supported interfaces as specified in the manifest.
• The account SHOULD call onInstall on the execution module to initialize state if specified by the user.
• The account MUST emit ExecutionInstalled as defined in the interface for all installed executions.

Uninstallation

During execution uninstallation, the account MUST correctly clear flags and other fields based on the incoming data and module manifest provided by the user.

• The account MUST remove all execution functions and clear flags and fields as specified in the manifest.
• The account MUST remove all execution hooks as specified in the manifest.
• The account SHOULD remove all supported interfaces as specified in the manifest.
• The account SHOULD call onUninstall on the execution module to clean up state and track call success if specified by the user.
• The account MUST emit ExecutionUninstalled as defined in the interface for all uninstalled executions.

Hooks

Execution Hooks Data Format

For accounts that implement execution hooks, accounts MUST conform to these execution hook formats:

1. For executeUserOp calls, for execution hooks associated with a validation function, accounts MUST send the full calldata (msg.data in solidity), including the executeUserOp selector.
2. For executeUserOp calls, for execution hooks associated with a selector, accounts MUST send PackedUserOperation.callData for executeUserOp calls, excluding executeUserOp.selector and the rest of the PackedUserOperation.
3. For executeWithRuntimeValidation calls, for all execution hooks, accounts MUST send the inner data field.
4. For all other calls, for execution hooks associated with a selector, accounts MUST send over the full calldata (msg.data in solidity).

Hook Execution Order

It is RECOMMENDED that an account implementer runs hooks in first installed first executed order. However, an account MAY implement a different execution order.

Validation Call Flow

Modular accounts support three different calls flows for validation: user op validation, runtime validation, and signature validation. User op validation happens within the account's implementation of the function validateUserOp, defined in the ERC-4337 interface IAccount. Runtime validation happens through the dispatcher function executeWithRuntimeValidation, or when using direct call validation. Signature validation happens within the account's implementation of the function isValidSignature, defined in ERC-1271.

For each of these validation types, an account implementation MAY specify its own format for selecting which validation function to use, as well as any per-hook data for validation hooks.

Within the implementation of each type of validation function, the modular account MUST check that the provided validation function applies to the given function selector intended to be run (See Checking Validation Applicability). Then, the account MUST execute all validation hooks of the corresponding type associated with the validation function in use. After the execution of validation hooks, the account MUST invoke the validation function of the corresponding type. If any of the validation hooks or the validation function reverts, the account MUST revert. It SHOULD include the module's revert data within its revert data.

The account MUST define a way to pass data separately for each validation hook and the validation function itself. This data SHOULD be sent as the userOp.signature field for user op validation, the authorization field for runtime validation, and the signature field for signature validation.

The result of user op validation MUST be the intersection of time bounds returned by the validation hooks and the validation function. If any validation hooks or the validation functions returns a value of 1 for the authorizer field, indicating a signature verification failure by the ERC-4337 standard, the account MUST return a value of 1 for the authorizer portion of the validation data.

The set of validation hooks run MUST be the hooks specified by account state at the start of validation. In other words, if the set of applicable hooks changes during validation, the original set of hooks MUST still run, and only future invocations of the same validation should reflect the changed set of hooks.

Checking Validation Applicability

To enforce module permission isolation, the modular account MUST check validation function applicability as part of each validation function implementation.

User op validation and runtime validation functions have a configurable range of applicability to functions on the account. This can be configured with selectors installed to a validation. Alternatively, a validation installation MAY specify the isGlobal flag as true, which means the account MUST consider it applicable to any module execution function with the allowGlobalValidation flag set to true, or for any account native function that the account MAY allow for global validation.

If the selector being checked is execute or executeBatch, the modular account MUST perform additional checking. If the target of execute is the modular account's own address, or if the target of any Call within executeBatch is the account, validation MUST either revert or check that validation applies to the selector(s) being called.

Installed validation functions have two additional flag variables indicating what they may be used for. If a validation function is attempted to be used for user op validation and the flag isUserOpValidation is set to false, validation MUST revert. If the validation function is attempted to be used for signature validation and the flag isSignatureValidation is set to false, validation MUST revert.

Direct Call Validation

If a validation function is installed with the entity ID of 0xffffffff, it may be used as direct call validation. This occurs when a module or other address calls a function on the modular account, without wrapping its call in the dispatcher function executeWithRuntimeValidation to use as a selection mechanism for a runtime validation function.

To implement direct call validation, the modular account MUST treat direct function calls that are not from the modular account itself or the EntryPoint as an attempt to validate using the caller's address and the entity ID of 0xffffffff. If such a validation function is installed, and applies to the function intended to be called, the modular account MUST allow it to continue, without performing runtime validation. Any validation hooks and execution hooks installed to this validation function MUST still run.

Execution Call Flow

For all non-view functions within IERC6900Account except executeWithRuntimeValidation, all module-defined execution functions, and any additional native functions that the modular account MAY wish to include, the modular account MUST adhere to these steps during execution:

If the caller is not the EntryPoint or the account, the account MUST check access control for direct call validation.

Prior to running the target function, the modular account MUST run all pre execution hooks that apply for the current function call. Pre execution hooks apply if they have been installed to the currently running function selector, or if they are installed as an execution hook to the validation function that was used for the current execution. Pre execution hooks MUST run validation-associated hooks first, then selector-associated hooks second.

Next, the modular account MUST run the target function, either an account native function or a module-defined execution function.

After the execution of the target function, the modular account MUST run any post execution hooks. These MUST be run in the reverse order of the pre execution hooks. If a hook is defined to be both a pre and a post execution hook, and the pre execution hook returned a non-empty bytes value to the account, the account MUST pass that data to the post execution hook.

The set of hooks run for a given target function MUST be the hooks specified by account state at the start of the execution phase. In other words, if the set of applicable hooks changes during execution, the original set of hooks MUST still run, and only future invocations of the same target function should reflect the changed set of hooks.

Module execution functions where the field skipRuntimeValidation is set to true, as well as native functions without access control, SHOULD omit the runtime validation step, including any runtime validation hooks. Native functions without access control MAY also omit running execution hooks.

Extension

Semi-Modular Account

Account implementers MAY choose to design a semi-modular account, where certain features, such as default validation, are integrated into the core account. This approach SHOULD ensure compatibility with fully modular accounts, as defined in this proposal, to maintain interoperability across different implementations.

Rationale

ERC-4337 compatible accounts must implement the IAccount interface, which consists of only one method that bundles validation with execution: validateUserOp. A primary design rationale for this proposal is to extend the possible functions for a smart contract account beyond this single method by unbundling these and other functions, while retaining the benefits of account abstraction.

This proposal includes several interfaces that build on ERC-4337. First, we standardize a set of modular functions that allow smart contract developers greater flexibility in bundling validation, execution, and hook logic. We also propose interfaces that provide methods for querying execution functions, validation functions, and hooks on a modular account. The rest of the interfaces describe a module's methods for exposing its modular functions and desired configuration, and the modular account's methods for installing and removing modules and allowing execution across modules and external addresses.

ERC-4337 Dependency

ERC-6900's main objective is to create a secure and interoperable foundation through modular accounts and modules to increase the velocity and security of the smart account ecosystem, and ultimately the wallet ecosystem. Currently, the standard prescribes ERC-4337 for one of its modular account call flows. However, this does not dictate that ERC-6900 will continue to be tied to ERC-4337. It is likely that smart account builders will want to develop modular accounts that do not use ERC-4337 in the future (e.g., native account abstraction on rollups). Moreover, it is expected that ERC-4337 and its interfaces and contracts will continue to evolve until there is a protocol-level account abstraction.

In the current state of the AA ecosystem, it is tough to predict the direction the builders and industry will take, so ERC-6900 will evolve together with the space's research, development, and adoption. The standard will do its best to address the objectives and create a secure foundation for modular accounts that may eventually be abstracted away from the infrastructure mechanism used.

Community Consensus

While this standard has largely been the result of collaboration among the coauthors, there have been noteworthy contributions from others in the community with respect to improvements, education, and experimentation. Thank you to the contributors:

• Gerard Persoon (@gpersoon)
• Harry Jeon (@sm-stack)
• Zhiyu Zhang (@ZhiyuCircle)
• Danilo Neves Cruz (@cruzdanilo)
• Iván Alberquilla (@ialberquilla)

We host community calls and working groups to discuss standard improvements and invite anyone with questions or contributions into our discussion.

Backwards Compatibility

Existing accounts that are deployed as proxies may have the ability to upgrade account implementations to one that supports this standard for modularity. Depending on implementation logic, existing modules may be wrapped in an adapter contract to adhere to the standard.

The standard also allows for flexibility in account implementations, including accounts that have certain features implemented without modules, so usage of modules may be gradually introduced.

Reference Implementation

See https://github.com/erc6900/reference-implementation

Security Considerations

Wallet/SDK Developers

From the wallet side, there are certain checks that the wallet/SDK should perform on the UI/UX side.

The standard introduces a concept of the execution manifest, which describes the execution functions, interface IDs, and hooks that should be installed on the account from an execution manifest. As part of constructing parameters to installExecution(), the SDK creates an executionManifest parameter that will specify the actions of the execution module being installed. SDKs will process the provided executionManifest() function of the underlying module, but the finished parameter will not necessarily be identical to the module’s return value of executionManifest(). Furthermore, this parameter is only going through very limited checks on-chain as part of the installation process. Therefore, the executionManifest parameter needs to be carefully constructed and verified before being used for installations. Furthermore, the executionManifest parameter of uninstallation should ideally be the same used in installation to not leave residual data.

The standard supports a isGlobalValidation flag for validation functions, which means that this function is added to a global validation pool and can validate any of the execution functions that expose themselves to global validation via allowGlobalValidation flag. Depending on the implementation, some, all or none of the native execution functions could be globally validated. Therefore, the wallets should be careful about what validation functions they allow to be installed with global validation enabled, as that would allow these functions to validate the exposed native functions and hence bypass any restrictions that may have been added to protect these native/execution functions. In a sense, these global validation functions could gain root access to the exposed native execution functions and potentially the whole account.

Module Developers

The standard does not enforce any rules surrounding what data needs to be installed or uninstalled when a module is added/deleted from the account. Hence the onUninstall() function in various modules may leave behind residual state data, especially since the external call may not be performed all. Furthermore, execution hooks linked to an uninstalled function may remain configured. This could pose security risks or lead to unexpected behavior in the case where the module is reinstalled with the same entityId. The danger is that previously set permissions or data may be unintentionally reused. Hence it is a good idea to fully uninstall all data linked to the smart module account including execution hooks.

Users

The modular smart contract accounts themselves are trusted components. Installed modules are trusted to varying degrees, as modules can interact with an arbitrarily large or small set of resources on an account. For example, a wide-reaching malicious module could add reverting hooks to native function selectors, bricking the account, or add execution functions that may drain the funds of the account. However, it is also possible to install a module with a very narrow domain, and depend on the correctness of the account behavior to enforce its limited access. Users should, therefore be careful in what modules to add to their account.

Copyright

Copyright and related rights waived via CC0.