Published in


PowerAgent v2: FAQs

Smart contracts don’t execute themselves; the evolution of DeFi requires automation

Based on experience gained since early 2020, PowerPool has for some time been developing the next iteration of our autonomous decentralized automated smart contract execution networks for projects running on any EVM compatible chain. Now audited and ready for initial tests with $CVP-staking Keepers looking to earn fees, PowerAgent v2 will replace PowerAgent v1, launched in late 2020, and serve as the primary automation tool powering existing and future PowerPool investor products.

Alongside building and auditing the code of PowerAgent v2, the PowerPool Research team has already published two in-depth articles describing the need for DeFi automation in general and the innovations offered by the new PowerAgent v2 architecture:

Automation Agent Networks: The missing layer of DeFi 2.0 infrastructure

or Smart Contracts don’t Execute Themselves, Part 1. Link to Part 1

TL/DR: As DeFi becomes ever more complex, protocols and users need increased (ideally autonomous and decentralized) automation to operate transparently and trustlessly while reducing user involvement and gas costs. DeFi is rapidly evolving from requiring all interactions to be done (and gas paid) mainly BY USERS, to most, if not all actions being done by automation FOR users, making their DeFi experience smoother, cheaper and easier. Most major DeFi protocols are developing in line with this trend.

Generalized, autonomous, decentralized automation networks like PowerAgent v2 offer protocols the option of writing less code, increasing transparency, and reducing internal dependencies while reducing cost and complexity for their users. Given the wide range of potential tasks, and the variety of types of tasks, automation networks like PowerAgent v2 are designed for maximum flexibility/configurability and deep customization of incentives for execution as well as reliability and transparency.

Introducing PowerAgent v2 Automation Network

or Smart Contracts don’t Execute Themselves, Part 2. Link to Part 2

TL/DR: In this second article, we introduce PowerAgent v2 architecture and its capabilities. PowerAgent v2 was developed with a generalization and flexibility in mind, giving EVM protocols maximum options for customization.

For example, there is no particular ‘hardcoded’ algorithm for Keepers’ selection in the basic version of PowerAgent v2. It was purposely abstracted to offer support for adding any type of Keepers selection algorithms and crypto economic qualification/incentive mechanics as may be desirable by a given protocol or for a specific investment vehicle. Multi-layered PowerAgent v2 architecture enables maximum generalization options:

PowerAgent v2 zero layer:

Keepers’ Selection Layer:

The Keepers’ selection layer allows the addition of custom Keepers selection logic for a specific group of tasks. Examples of Keepers’ selection logic: Round Robin, Round Robin by Priority, Least Connection, etc.

Cryptoeconomics Layer:

Keepers’ selection is tighly bonded with cryptoeconomics rules applied to Keepers. It includes staking, reward, and slashing policies — what would happen if the selected Keeper missed the transaction signing? How he will be slashed? Or, in opposite what reward for this particular task will he receive in case of successfully executed task?

In PowerAgent v2 is possible to add arbitrary cryptoeconomic logic to particular groups of tasks.

PowerAgent v2 FAQ: Any Questions?

After publishing these articles and distributing them to domain experts, we received many questions mainly focusing on PowerAgent v2 use cases, features, and competitive advantages:

Could you provide more details on PowerAgent v2 use cases?

The main use case of the PowerAgent network is executing (calling) smart contracts reliably and transparently according to pre-specified conditions. EVM-compatible blockchains do not provide any means for smart contracts/protocols to execute event-driven, chronologically-driven or conditional contract calls without assistance from an external agent/keeper contract. The PowerAgent v2 contract works as follows: the protocol/project specifies/adds any task with conditioned execution of a particular contract, and the autonomous automation network will reliable and transparently call this contract (make a transaction) when conditions are met. Protocols/projects also can automate any arbitrary sequence of transactions.

Although third-party protocols can implement some strategies for their users, some examples of particular on-chain actions (strategies) that can be automated by generalized autonomous PowerAgent v2 include the following:

  1. Yield harvesting/compounding — The automation network can perform a sequence like claim rewards from LP mining contracts (for example, Curve Gauge), exchange CRV to stablecoin, and stake stablecoin proceeds to the position (compound interest). Although Yearn.finance and some other projects have built specialized, complex and non-autonomous protocols around Curve strategies, PowerAgent v2 offers all of these capabilities to any project “out of the box”. Other examples where relatively simple automation is needed, like compounding swap fees into users in Uniswap v3 are all candidates for autonomous decentralized automation networks.
  2. Periodic updating of on-chain variables — For example, there is ERC20/ETH trading pair at Uniswap, and there is a need for a TWAP on-chain oracle. PowerAgent can execute tasks coming down to calling the TWAP calculation method and ensure periodic update of a price feed.
  3. Limit orders on DEXes — The PowerAgent network can execute swap on DEX based on specified conditions (spot price, time, etc). Another investment strategy can be the DCA (dollar-cost-averaging) strategy, when a given asset is purchased every time interval dT, averaging the investment basis over time.
  4. Collateral ratio maintenance — A simple example: the user deposits ETH collateral into MakerDAO, mints DAI, and supplies DAI to yield farming options (Yearn Vault). If the ETH price grows, it is logical to mint more DAI (maintaining initial collateral ratio) and farm bigger yield (in absolute dollar terms). In contrast, if the ETH price declines, a share of DAI should be burnt to maintain a healthy collateral ratio and protect the position from liquidation risks. This can be applied to almost any collateral management tasks on any EVM lending markets, such as Compound, AAVE, etc.
  5. Concentrated liquidity LP position management — Concentrated liquidity DEXs like Uniswap v3, (and soon others) require automated position shifting to minimize impermanent loss and maximize swap fee income. There are already third party specialized protocols with strategies for Uni v3 position mgmt. But these and other future use cases could all simply use PowerAgent v2.

Comparisons: How does PowerAgent differ from the Keep3r network?

The first and very important difference is that the PowerAgent v2 design is generalized, and also focused on task-specificity. The PowerAgent v2 architecture allows flexible specification of both Keepers’ selection and the cryptoeconomic mechanics (staking/rewards/penalties) for any commingled list/group of tasks. Keep3r is not nearly as customisable and flexible.

Other differences between Keep3r and PowerAgent v2 include:

Table 1. PoweAgent v2 and Kp3r comparison

Comparison with closest competitors, such as Gelato and Chainlink?

Table 2. PoweAgent v2, Gelato, and Chainlink comparison

Note for the comparison: we tried to do our best. If something wrong/irrelevant/information can be added — please reach out to @vasily_sumanov Twitter, @powerpoolcvp Twitter, or Discord.

How do we plan to motivate Keepers to place big size stakes?

Since a lot of tasks (especially high-value tasks) would possibly require Keepers with high stakes (to offset no-execution risk with slashing), this question is very important.

Our strategy for attracting Keepers with high stakes is the following:

  1. CVP Rewards scaled by stake size motivating high CVP stakes
  2. Continuously increasing minimal stake size via DAO votings and attracting more high-value tasks to the network with a certain stake size threshold protecting possible Keeper misbehavior.

We will continuously develop the Keepers acquisition strategy with the PowerPool community to achieve not only a certain quantity of Keepers in the network but also high stakes (skin in the game) and resulting in the capability to perform high-value tasks.

Any further questions about PowerAgent v2 or automation?

Please, drop us a message on Discord, @powerpoolcvp Twitter, or contact our Head of Research @vasily_sumanov on Twitter.



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store


A solution for accumulating governance power in Ethereum based protocols