Keepers

Keepers are incentivized external actors—typically automated bots, though sometimes human operators—that perform critical maintenance functions for decentralized finance protocols by monitoring blockchain state, identifying profitable opportunities, and executing transactions that keep systems healthy. In the Sky Protocol ecosystem (formerly MakerDAO), keepers serve essential roles including triggering liquidations on undercollateralized vaults, participating in collateral auctions, updating protocol parameters, and automating governance spell execution. ^[1] These actors operate permissionlessly, requiring no special authorization beyond the capital and technical infrastructure to participate, embodying DeFi's principle that economic incentives can reliably motivate decentralized protocol maintenance without trusted intermediaries.

The keeper ecosystem represents one of decentralized finance's most elegant solutions to a fundamental coordination problem—how to ensure protocols perform necessary maintenance operations when no central authority exists to execute them. By creating profit opportunities for monitoring protocol state and executing specific transactions, Sky Protocol aligns individual keeper self-interest with collective protocol health. A keeper who identifies an undercollateralized vault and triggers its liquidation earns fees and potential arbitrage profits while simultaneously protecting the protocol from bad debt that could threaten USDS peg stability. ^[2]

As of January 2026, the keeper ecosystem supporting Sky Protocol includes professional MEV (Maximal Extractable Value) firms operating sophisticated infrastructure, individual developers running custom keeper bots, vault automation services like DeFi Saver and Instadapp protecting users from liquidation, and Chainlink Automation infrastructure providing reliable execution for protocol maintenance tasks. ^{[3] [4]} The protocol compensates keeper infrastructure through multiple mechanisms including flat tip payments (100 DAI per liquidation as of 2026), percentage chip fees (typically 2-3% of liquidated debt), liquidation penalties (13% of vault debt that creates arbitrage opportunities), and gas optimization subsidies through efficient smart contract design. ^[5]

The importance of robust keeper infrastructure became devastatingly apparent during the March 12, 2020 "Black Thursday" crisis, when Ethereum network congestion caused keeper bot failures that resulted in approximately $8.32 million in collateral being liquidated for zero DAI bids—a catastrophic failure mode that threatened the entire protocol. ^[6] The incident prompted comprehensive redesign of liquidation mechanisms, keeper incentive structures, and gas price handling, demonstrating how keeper reliability directly impacts protocol solvency during crisis conditions.

Keeper operations increasingly intersect with the broader MEV extraction ecosystem, where sophisticated actors use advanced techniques including transaction ordering manipulation, sandwich attacks, and just-in-time liquidity provision to capture value from blockchain transaction flow. This intersection creates both opportunities—more sophisticated actors improving keeper response times and capital efficiency—and concerns about centralization as MEV extraction advantages accrue to well-capitalized entities with privileged access to block builders and validators. ^[7]

Understanding keepers requires examining their technical implementation, economic incentives, competitive dynamics, historical failures, and ongoing evolution as the DeFi ecosystem matures. This article provides comprehensive analysis of keeper roles within Sky Protocol, the infrastructure enabling keeper operations, profitability considerations, automation services, risks and challenges, and the future trajectory as keepers evolve alongside blockchain scaling solutions and MEV extraction techniques.

Overview and Core Functions

Keepers perform discrete, verifiable actions that maintain protocol health, with compensation structures designed to ensure profitable execution under normal conditions while providing resilience during network congestion or market volatility. The term "keeper" originated in the MakerDAO ecosystem around 2017-2018 to describe external actors who "keep" the protocol functioning smoothly, distinguishing them from protocol core contracts or governance processes. ^[8] The keeper model has since been adopted across DeFi protocols including Aave, Compound, Liquity, and numerous others facing similar coordination challenges.

Primary Keeper Functions in Sky Protocol

Liquidation Triggering

The most critical keeper function involves monitoring all vault positions for collateralization ratio breaches and calling the DOG contract's bark() function to initiate liquidation when positions become unsafe. ^[9] Keepers continuously query vault collateral values, debt amounts, and liquidation ratios for thousands of positions, comparing collateral value (based on oracle prices) against debt to identify vaults falling below minimum thresholds. When a keeper identifies an unsafe vault, they submit a transaction calling bark(ilk, urn) specifying the collateral type and vault address.

The DOG contract validates that the vault is indeed undercollateralized, calculates total debt including accrued stability fees, applies the liquidation penalty (typically 13%), seizes collateral via the VAT contract's grab() function, and initiates a Dutch auction through the appropriate CLIPPER contract. ^[10] The keeper who triggers liquidation receives flat tip compensation (100 DAI as of 2026) plus a percentage chip fee (2-3% of liquidated debt), providing immediate profit regardless of auction outcomes. ^[5]

Liquidation triggering requires sophisticated infrastructure including Ethereum full nodes for real-time blockchain state access, oracle price feed monitoring to detect collateralization ratio changes, efficient database indexing of all vault positions, gas price estimation and optimization algorithms, and transaction replacement capabilities for competing during network congestion. Professional keepers typically run multiple redundant systems with failover capabilities to ensure they don't miss profitable liquidation opportunities. ^[11]

Collateral Auction Participation

After liquidation triggering, collateral is sold through Dutch auctions implemented by CLIPPER contracts where price starts high and decreases over time until buyers accept. ^[12] Keeper-operated auction bots monitor active auctions, evaluate whether current prices offer profitable arbitrage opportunities after accounting for gas costs and price impact when reselling collateral, and submit take() transactions to purchase collateral at acceptable prices. Successful auction participation requires capital reserves to purchase collateral, DEX liquidity analysis to estimate resale prices, real-time auction price curve monitoring, and sophisticated gas price strategies to ensure transaction inclusion.

The Liquidation 2.0 upgrade implemented after Black Thursday replaced earlier English ascending auctions with Dutch descending auctions specifically to improve keeper participation during high gas price environments. ^[13] Dutch auctions start at prices above market value, decreasing toward market rates, ensuring that early auction phase provides comfortable profit margins that incentivize keeper participation even with elevated gas costs. This design prevented the zero-bid exploits that occurred during Black Thursday when keepers couldn't profitably participate in English auctions requiring competitive bidding.

Surplus and Debt Auction Operations

Beyond collateral liquidations, keepers participate in surplus auctions (flap) where the protocol sells excess DAI/USDS to purchase and burn SKY tokens, and debt auctions (flop) where the protocol mints new SKY to sell for DAI/USDS to cover bad debt. ^[14] These auction types occur less frequently than collateral liquidations—surplus auctions trigger when protocol surplus exceeds buffer targets, while debt auctions occur only during crisis events when liquidation losses exceed available surplus.

Keepers participating in flap auctions compete to offer the most SKY tokens in exchange for the auctioned DAI, with winning bids effectively purchasing SKY at discount rates for burn. Flop auctions invert this structure, with keepers competing to accept the minimum amount of minted SKY tokens in exchange for DAI that recapitalizes the protocol. Both auction types require keeper understanding of SKY token market prices, capital availability for auction participation, and willingness to hold positions through potential volatility.

Governance Spell Execution

The chief-keeper bot monitors the DSChief voting contract to automatically lift spells (governance proposals) to hat status when they accumulate sufficient approval, then schedules them for execution after the Governance Security Module (GSM) delay expires. ^[15] When the chief-keeper detects a spell receiving more voting weight than the current hat, it calls DSChief.lift(spellAddress) to designate the new leading spell, then calls Pause.exec() after the GSM delay (currently 30-48 hours) to execute approved changes.

This automation ensures governance decisions are implemented promptly without requiring manual monitoring and execution, though the chief-keeper operates as a permissioned service rather than competitive market participation. As of 2026, the protocol compensates Chainlink Automation for chief-keeper services at 1,500 USDS per day under a three-year stream initiated May 29, 2023. ^[3] The infrastructure provides redundancy and reliability guarantees that individual keepers might not maintain consistently.

Oracle Feed Updates and Maintenance

While not strictly "keeper" functions in the same sense as liquidations, certain protocol operations require periodic triggering including stability fee accumulation (calling JUG.drip()), oracle price updates (poking OSM contracts), and rate adjustments. Some of these functions are called permissionlessly by any actor benefiting from updated state, while others may be automated through keeper networks or protocol-funded infrastructure. ^[16]

The economic model for maintenance operations differs from liquidation triggering—these transactions typically don't offer direct arbitrage profits, instead providing small gas reimbursements or indirect benefits from ensuring protocol state accuracy. As a result, many maintenance operations are handled by protocol-funded keeper networks rather than purely profit-motivated external actors, representing a pragmatic compromise between complete decentralization and operational reliability.

Keeper Infrastructure and Technical Requirements

Operating competitive keeper bots requires substantial technical infrastructure and ongoing operational costs. Professional keeper operators typically maintain:

Blockchain Infrastructure — Ethereum full nodes or archive nodes for historical state queries, using high-performance clients like Geth or Erigon configured for fast state sync and low latency. Many keepers run their own nodes rather than relying on RPC providers to avoid rate limiting and ensure maximum uptime. Annual costs for robust node infrastructure range from $5,000 to $20,000 depending on hardware specifications and redundancy requirements. ^[17]

Database and Indexing — Specialized databases indexing vault positions, auction states, and historical transactions to enable efficient querying without repeatedly scanning blockchain state. Graph Protocol subgraphs, custom PostgreSQL databases, or specialized DeFi data providers like Dune Analytics or Nansen offer different trade-offs between development effort, query performance, and cost.

Monitoring and Alerting Systems — Real-time monitoring infrastructure tracking protocol state changes, oracle price movements, gas price fluctuations, and competitor keeper activity. Alert systems notify operators of anomalies, failed transactions, or unusual market conditions requiring manual intervention.

Capital Reserves — Keeper operations require capital for gas costs, auction participation collateral, and potential temporary holding of liquidated assets. Professional keepers maintaining readiness for large liquidations or rapid auction participation may keep hundreds of thousands of dollars in working capital deployed, creating significant opportunity costs from capital that could otherwise be deployed in yield-generating positions.

Transaction Infrastructure — Sophisticated transaction submission systems including private transaction relays (for MEV protection or front-running), relationships with block builders for priority inclusion, transaction replacement and acceleration capabilities, and multi-node broadcasting for redundancy. Advanced keepers increasingly use MEV-Boost and similar infrastructure to access priority blockspace, though this introduces centralization dependencies.

Liquidation Keepers and Auction Mechanics

Liquidation keepers represent the most critical and competitive keeper category, as their prompt action directly protects protocol solvency by ensuring undercollateralized positions are liquidated before collateral values fall below debt obligations. The liquidation process has evolved substantially from early MakerDAO implementations to current Liquidation 2.0 architecture, with each iteration responding to discovered vulnerabilities and operational failures.

Historical Evolution: Liquidation 1.0 to Liquidation 2.0

Liquidation 1.0: English Ascending Auctions (2017-2021)

The original liquidation system used Flip contracts implementing English ascending auctions where keepers competed by submitting progressively higher bids for fixed collateral amounts. ^[18] The auction began with a starting price based on oracle valuations, and keepers submitted bids offering more DAI for the collateral. After a timeout period with no higher bids, the winning bidder received the collateral and the protocol collected the winning bid amount.

This model worked adequately during normal market conditions but revealed catastrophic vulnerabilities during Black Thursday. When Ethereum network congestion caused gas prices to spike from typical 20-40 gwei to over 200 gwei, many keeper bots failed to adjust their gas price strategies and their transactions were evicted from mempools before inclusion. ^[6] With limited keeper participation due to prohibitive gas costs, some auctions received only a single bid or even zero bids, allowing sophisticated actors to purchase collateral for minimal DAI amounts.

The most extreme case saw one keeper winning 50 ETH (worth approximately $6,000 at the time) for a 0 DAI bid, profiting entirely at the expense of protocol solvency and vault owners whose collateral was seized without corresponding debt repayment. ^[19] These zero-bid auctions accounted for approximately $8.32 million in losses across 1,462 liquidation events, representing 36.6% of Black Thursday liquidations. ^[20]

Liquidation 2.0: Dutch Descending Auctions (2021-Present)

The Liquidation 2.0 upgrade replaced English ascending auctions with Dutch descending auctions (implemented through Clipper contracts) to address Black Thursday failures. ^[21] In Dutch auctions, collateral is offered at prices starting well above current market value, then decreasing over time according to governance-set parameters. The first keeper willing to accept the current price can call take() to purchase the collateral immediately at that price.

This mechanism provides several advantages. Price starts high, ensuring the protocol doesn't give away collateral cheaply—if keepers believe the starting price is too high, they simply wait for it to decrease. Keepers don't compete against each other through gas auctions for the same collateral—the first keeper willing to accept the current price wins, reducing gas wars. Partial fills are supported, allowing keepers with limited capital to purchase portions of large liquidations rather than requiring full-collateral purchases. ^[22]

The price decrease follows a configurable curve, typically linear or exponential, defined by governance parameters including buf (starting price multiplier above oracle price), tail (maximum auction duration before reset), cusp (minimum price as percentage of oracle value), and step (price decrease per second). ^[23] These parameters balance conflicting objectives—starting prices high enough to discourage early participation versus decreasing fast enough to ensure liquidations complete promptly.

Current Liquidation Incentive Structure (2026)

Sky Protocol's liquidation incentive design aims to ensure keeper profitability across diverse market conditions and gas price environments through multi-layered compensation:

Flat Tip Incentive

Each liquidation trigger earns a flat tip payment, currently set at 100 DAI, credited to the keeper regardless of auction outcome or collateral value. ^[5] This fixed payment ensures minimum compensation for keepers monitoring vault health and submitting liquidation transactions, particularly important for smaller liquidations where percentage-based fees might not cover gas costs. During periods of elevated gas prices (100+ gwei), the 100 DAI tip may be insufficient to ensure profitable participation, though the protocol's design assumes such conditions are temporary.

Percentage Chip Incentive

Beyond the flat tip, keepers receive a chip percentage of the liquidated debt, typically set at 2-3% depending on collateral type. ^[5] For a vault liquidated with 50,000 DAI debt and 2% chip, the keeper earns an additional 1,000 DAI. This percentage-based compensation scales with liquidation size, ensuring larger liquidations (which require more capital for auction participation) provide correspondingly larger keeper rewards.

The chip parameter represents a trade-off governance must balance—higher chips ensure keeper participation during challenging conditions but increase liquidation penalties that vault owners pay, making Sky Protocol less competitive compared to platforms with lower liquidation costs. The current 2-3% range reflects calibration through years of market observation, though governance adjusts parameters periodically based on keeper participation rates and competitive dynamics.

Liquidation Penalty and Arbitrage Opportunities

The liquidation penalty, currently 13% for most collateral types, creates the primary profit opportunity for auction participants. ^[24] When a vault with 10 ETH collateral (valued at $25,000) and 15,000 DAI debt is liquidated, the total debt becomes 15,000 × 1.13 = 16,950 DAI. The auction must collect at least 16,950 DAI to fully repay the debt, with any excess returned to the vault owner.

Keepers participating in the auction seek to purchase the 10 ETH for less than its market value, creating arbitrage profit. If the auction settles at 16,000 DAI (insufficient to cover full debt plus penalty), keepers acquired 10 ETH (worth $25,000) for 16,000 DAI, earning approximately $9,000 profit after reselling the ETH on DEXs. This arbitrage opportunity—the gap between auction purchase price and market resale price—provides the primary incentive for auction participation, though it comes at the expense of vault owners losing excess collateral value.

The 13% penalty reflects protocol attempts to balance liquidation costs against keeper participation incentives. Lower penalties (8-10%) would be more friendly to vault users but might not ensure adequate keeper participation during volatile conditions. Higher penalties (15-20%) would strongly incentivize keepers but make Sky Protocol uncompetitive compared to platforms like Aave or Compound with lower liquidation costs, potentially driving users to alternatives.

Auction Participation Strategies and Competition

Professional keeper operators employ sophisticated strategies to maximize auction profitability while managing gas costs, capital efficiency, and execution risks:

Fair Market Value (FMV) Discount Models

The simplest auction participation strategy calculates fair market value for collateral based on DEX prices, then bids when the auction price reaches FMV minus target discount percentage. ^[25] For example, a keeper targeting 5% profit margins might bid when the auction price falls to 95% of the collateral's Uniswap price. This strategy requires real-time DEX price monitoring, gas cost estimation to ensure net profitability, and slippage calculation for reselling collateral.

FMV discount models work well for liquid collateral types like ETH or WBTC where deep DEX liquidity enables reliable price discovery and low-slippage resale. For less liquid collateral or large liquidations, keepers must adjust discount requirements to account for price impact when reselling substantial collateral quantities, potentially waiting for auction prices to decrease further to ensure adequate profit margins after accounting for elevated slippage.

Dynamic Gas Price Adjustment

The Black Thursday crisis demonstrated that static gas price strategies cause catastrophic keeper failures during network congestion. ^[6] Modern keeper implementations use dynamic gas price adjustment algorithms monitoring current network gas prices, estimating expected transaction inclusion times at various gas price levels, and calculating maximum profitable gas prices given expected liquidation rewards.

A keeper expecting to earn 1,500 DAI from tip + chip + arbitrage on a liquidation might be willing to pay up to 500,000 gas at 200 gwei (approximately 1,000 DAI at $2,000 ETH prices) if network conditions require such prices for prompt inclusion. More sophisticated strategies implement escalation mechanisms that initially submit transactions at median gas prices, then progressively increase gas prices if transactions remain unconfirmed, balancing probability of winning liquidations against gas cost minimization.

Capital Efficiency and Partial Liquidations

Liquidation 2.0's support for partial fills enables keepers to participate in large liquidations without requiring capital to purchase entire collateral amounts. ^[22] A keeper with 100,000 DAI capital can participate in a 500,000 DAI liquidation by purchasing 20% of the collateral, allowing smaller keeper operators to compete in markets previously dominated by well-capitalized entities.

This feature has democratizing effects, lowering capital requirements for keeper participation from hundreds of thousands of dollars to tens of thousands, but also introduces coordination challenges. Multiple keepers may simultaneously attempt to take different portions of the same auction, leading to transaction race conditions where only the first included transaction succeeds while others revert. Sophisticated keepers implement race condition handling including transaction simulation before submission, monitoring of pending transactions from other keepers, and rapid transaction cancellation when pre-empted by competitors.

MEV Extraction and Priority Ordering

The intersection of keeper operations with MEV extraction has introduced new competitive dynamics. Sophisticated actors use privileged relationships with block builders, private transaction pools, and priority gas auctions to ensure their liquidation and auction transactions are included before competing keepers. ^[7] This infrastructure advantage creates winner-take-most dynamics where a small number of sophisticated MEV operators capture the majority of liquidation profits.

Some keepers operate their own block building infrastructure or maintain direct relationships with major block builders like Flashbots, enabling them to submit transactions with guaranteed inclusion and priority positioning. These advantages are particularly valuable during high-competition liquidation events where milliseconds of execution priority determine which keeper wins the opportunity, creating pressure for aspiring keepers to either invest in similar infrastructure or accept lower profitability competing in the public mempool.

Keeper Competition and Market Structure

The liquidation keeper market has evolved from the early MakerDAO days when individual developers ran simple keeper bots to today's professionalized ecosystem dominated by sophisticated MEV firms. As of 2026, the liquidation keeper landscape includes:

Professional MEV Firms — Well-capitalized entities operating advanced infrastructure including custom block building, private order flow, and optimized execution algorithms. Firms like Flashbots, Jito Labs, and specialized high-frequency trading operations adapted from traditional finance dominate liquidation capture, particularly for large, high-value opportunities. These entities benefit from economies of scale, amortizing fixed infrastructure costs across multiple protocols and keeper strategies.

Individual Keeper Operators — Independent developers running open-source keeper implementations like auction-keeper, often targeting smaller liquidations or niche collateral types less attractive to large operators. These participants face increasing difficulty competing with MEV firms' infrastructure advantages but maintain relevance for smaller liquidations where sophisticated infrastructure overhead isn't justified by profit potential.

Protocol-Funded Keeper Networks — Some critical functions including governance spell execution are handled by compensated keeper services rather than purely competitive markets. Sky Protocol's relationship with Chainlink Automation for chief-keeper services exemplifies this model, accepting reduced decentralization for operational reliability. ^[3]

Vault Automation Services

Beyond permissionless keeper bots participating in protocol-level liquidations, a parallel ecosystem of user-facing automation services has emerged to help vault owners avoid liquidation through proactive position management. These services—led by platforms like DeFi Saver and Instadapp—represent a different keeper model focused on user convenience and safety rather than protocol maintenance, though they contribute to system health by reducing liquidation frequency.

DeFi Saver: Pioneering Automated Protection

DeFi Saver launched in 2019 as one of the first non-custodial vault management interfaces, initially focusing on MakerDAO CDPs before expanding to support Aave, Compound, Liquity, and numerous other protocols. ^[26] The platform's core innovation was Automation—a service that monitors user vault positions and automatically executes protective actions when defined conditions are met, providing liquidation protection without requiring users to manually intervene during market volatility.

Automated Leverage Management

The flagship Automation feature monitors vault collateralization ratios and automatically adjusts positions when they approach dangerous levels or profitable targets. ^[27] Users define target collateralization ratios (e.g., maintain between 200-250%) along with threshold triggers. When the ratio falls below the minimum threshold (approaching liquidation risk), Automation executes a "Repay" action—closing a portion of the vault by selling collateral for DAI and repaying debt, improving the collateralization ratio. When the ratio exceeds the maximum threshold (indicating under-leveraged position), Automation executes a "Boost" action—borrowing additional DAI and purchasing more collateral to increase leverage.

These automated adjustments use complex DeFi composability, executing multi-step transactions that interact with DEXs (Uniswap, 1inch), flash loan providers (Aave, dYdX), and the target protocol in single atomic transactions. A typical Repay action might: (1) take a flash loan of DAI, (2) repay vault debt, (3) withdraw freed collateral, (4) swap collateral for DAI on Uniswap, (5) repay the flash loan, (6) deposit excess DAI to the vault if any remains. This complexity is abstracted from users who simply see their position maintained within target ranges automatically.

Stop Loss and Take Profit

DeFi Saver introduced Stop Loss strategies in 2020, allowing users to define price levels where positions should automatically close to prevent further losses. ^[28] If ETH price falls below a user's defined threshold, Automation closes the vault by selling all collateral, repaying debt, and returning any remaining funds to the user's wallet. This feature proved particularly valuable during March 2020's Black Thursday, when many DeFi Saver users with Stop Loss enabled avoided liquidation penalties by automatically exiting positions before collateralization ratios became critical.

Take Profit provides the inverse functionality—automatically closing profitable positions when collateral prices reach targets, locking in gains without requiring users to monitor markets continuously. A user who opened an ETH vault at $1,500 ETH might set a Take Profit trigger at $3,000 ETH, automatically closing the position and realizing the gain when that price is reached.

Economics and Business Model

DeFi Saver charges service fees for automated actions, typically 0.25-0.50% of the transaction amount depending on operation type and protocol. ^[29] For a 10,000 DAI Repay action, users might pay a 25-50 DAI service fee. The platform also generates revenue from DEX trade commission kickbacks, where exchanges like 1inch pay a percentage of swap fees to integrating platforms that route order flow.

The business model demonstrates how keeper services create value by saving users significantly more through avoided liquidation penalties (typically 13% of debt) than they charge in service fees. A user avoiding a single liquidation on a 50,000 DAI vault saves approximately 6,500 DAI in penalties and lost collateral compared to perhaps 100-200 DAI in Automation service fees over the same period, providing clear value proposition despite ongoing costs.

Instadapp: Protocol Aggregation and Automation

Instadapp launched in 2018 as a DeFi management dashboard aggregating multiple lending protocols (MakerDAO, Compound, Aave) with a focus on one-click position migration and optimization. ^[30] In February 2023, Instadapp partnered with Gelato Network to launch Vault Automation, bringing similar liquidation protection capabilities to its user base. ^[31]

Cross-Protocol Position Management

Instadapp's distinctive feature enables migrating positions between protocols based on profitability or safety considerations. ^[32] If a user's MakerDAO vault approaches liquidation while Aave offers higher collateralization ratios for the same collateral, Instadapp can execute a one-transaction migration—closing the MakerDAO vault, moving collateral to Aave, and opening a new position with improved safety margins. This cross-protocol arbitrage improves capital efficiency and risk management compared to single-protocol solutions.

The platform's DSL (DeFi Smart Layer) and DSA (DeFi Smart Accounts) architecture abstract protocol differences, allowing users to interact with diverse protocols through unified interfaces. ^[33] From a user perspective, vault management across MakerDAO, Aave, and Compound feels similar despite substantial underlying implementation differences, lowering learning curves and encouraging multi-protocol portfolio management.

Gelato-Powered Automation

Instadapp's automation implementation uses Gelato Network, a decentralized keeper network where anyone can run Gelato nodes to execute automated tasks in exchange for fees. ^[31] This differs from DeFi Saver's proprietary infrastructure by distributing execution responsibility across a permissionless keeper network, though it introduces dependencies on Gelato's reliability and economics.

When Instadapp users enable automation, they define conditions and actions stored in their DSA smart accounts. Gelato keepers monitor these conditions, and when triggers activate, submit transactions calling the DSA to execute the defined actions. Users pay Gelato fees (typically slightly higher than DeFi Saver's direct execution costs) in exchange for the decentralization benefits of distributed keeper infrastructure.

Automation Service Tradeoffs and Limitations

Vault automation services provide valuable protection but involve tradeoffs users should understand:

Execution Timing and Slippage — Automation services execute market orders when conditions trigger, potentially experiencing unfavorable slippage during volatile conditions. During Black Thursday, even users with automation enabled sometimes experienced worse outcomes than manual management due to temporary extreme price dislocations and DEX liquidity crunches. The services optimize for reliable execution rather than perfect pricing, accepting some slippage costs as inherent to automated position management.

Smart Contract and Execution Risk — Using automation requires granting smart contract permissions to access and modify vault positions. While both DeFi Saver and Instadapp have operated for years without major exploits, smart contract risk remains inherent. Users effectively trust automation service contracts with full vault control, requiring confidence in the platform's security auditing and operational practices. ^[34]

Service Fee Costs — The ongoing service fees (0.25-0.50% per action) compound over time, particularly for frequently-adjusted positions. Users with highly leveraged positions triggering automation multiple times during volatile periods might pay hundreds or thousands of dollars annually in service fees, though these costs are typically far lower than single liquidation event losses.

Limited Conditions and Actions — Automation services support predefined condition types (collateralization ratio thresholds, price triggers) and action types (Boost, Repay, Close), but more complex strategies require manual management. Users wanting sophisticated position management incorporating options hedging, cross-protocol yield optimization, or conditional logic based on multiple protocols must implement custom solutions or accept the limitations of available automation tooling.

Despite these limitations, automation services have become essential infrastructure for serious vault users, particularly those maintaining aggressive leverage ratios or managing large positions where liquidation losses would be substantial. The services demonstrate that viable business models exist for keeper-adjacent services focused on user experience rather than protocol-level operations, suggesting continued evolution of the keeper ecosystem beyond pure liquidation capture.

Running a Keeper: Technical Implementation

Operating a keeper bot requires combining blockchain infrastructure, economic strategy, and operational reliability. This section provides technical detail on keeper architecture, drawing from the open-source auction-keeper codebase maintained by Sky Protocol contributors and widely used as reference implementation by keeper operators. ^[35]

Auction-Keeper Architecture

The auction-keeper codebase provides a Python-based framework for participating in Sky Protocol collateral auctions (CLIP), surplus auctions (FLAP), and debt auctions (FLOP). ^[36] The architecture separates blockchain interaction concerns from bidding strategy through a plugin model where external bidding models determine when and how much to bid, while the keeper framework handles transaction submission, gas management, and auction monitoring.

Core Components

The keeper consists of several interconnected components:

• Auction Monitors watch CLIPPER, FLAPPER, and FLOPPER contract events to detect new auctions and track existing auction state changes. When a new liquidation occurs, the DOG contract emits a Kick event containing auction ID, collateral type, and debt amount. The monitor subscribes to these events through Ethereum RPC WebSocket connections, maintaining up-to-date state without repeatedly polling contract state.

• Bidding Model Interface defines a simple protocol where external processes receive auction details and return bid price recommendations. The auction-keeper calls external bidding model scripts (which can be implemented in any language) with auction parameters, and the model responds with either a bid amount or indication to skip the auction. This separation enables sophisticated bidding strategies to be developed independently from blockchain interaction concerns.

• Transaction Manager handles bid submission with gas price optimization, transaction replacement when gas prices increase, and confirmation monitoring. The manager implements retry logic with exponential backoff for failed transactions, gas price escalation when transactions remain pending, and parallel transaction submission across multiple Ethereum nodes for redundancy.

• Rebalancing Module manages keeper capital, automatically converting auctioned collateral to DAI through DEX integrations including Uniswap, 1inch, and 0x. After winning an auction and receiving collateral, the keeper must quickly sell to DAI to free capital for future auction participation. The rebalancing module monitors collateral balances and executes swaps when holdings exceed configured thresholds.

Setup and Configuration

Running an auction-keeper requires substantial initial setup:

git clone https://github.com/makerdao/auction-keeper.git
cd auction-keeper

# Install Python 3.6+ dependencies
pip3 install -r requirements.txt

# Configure blockchain connection
export ETH_RPC_URL="http://localhost:8545"  # Full node RPC endpoint
export ETH_FROM="0x..."  # Keeper account address
export ETH_KEY="key_file=/path/to/keystore.json,pass_file=/path/to/password.txt"

# Launch keeper monitoring clip auctions
bin/auction-keeper \
    --type clip \
    --ilk ETH-A \
    --bid-only \
    --model ./models/simple_discounted_model.sh \
    --flipper 0x... \  # Clipper contract address
    --min-auction 1000 \  # Minimum auction size
    --max-auctions 100 \  # Maximum concurrent auctions
    --from-block 12000000

This basic invocation starts an auction-keeper monitoring ETH-A collateral liquidations, using a simple discounted bidding model that attempts to purchase collateral at specified percentage discounts from market value.

Bidding Model Development

Bidding models determine keeper profitability by calculating optimal bid prices given current market conditions. ^[37] The simplest model calculates fair market value (FMV) from DEX prices and bids at FMV minus target discount:

#!/usr/bin/env python3
import sys
import json
import requests

def get_eth_price():
    # Query Uniswap for ETH/USDC price
    response = requests.post(
        'https://api.thegraph.com/subgraphs/name/uniswap/uniswap-v3',
        json={'query': '{ pool(id: "0x...") { token0Price } }'}
    )
    return float(response.json()['data']['pool']['token0Price'])

def calculate_bid(auction):
    collateral_amount = float(auction['lot'])
    debt_owed = float(auction['tab'])

    # Get current market price
    eth_market_price = get_eth_price()

    # Calculate fair market value of collateral
    fmv = collateral_amount * eth_market_price

    # Target 5% profit margin after gas costs
    target_discount = 0.05
    estimated_gas_cost = 50  # USD

    # Calculate maximum bid
    max_bid = fmv * (1 - target_discount) - estimated_gas_cost

    # Only bid if profitable
    if max_bid > debt_owed and max_bid > 0:
        return {
            'price': max_bid / collateral_amount,  # Price per unit collateral
            'amount': min(max_bid, debt_owed)  # Don't overbid
        }
    else:
        return None  # Skip this auction

if __name__ == '__main__':
    # Read auction details from stdin
    auction_data = json.load(sys.stdin)

    # Calculate bid
    bid = calculate_bid(auction_data)

    # Output bid or null
    print(json.dumps(bid))

This simplified model demonstrates the core logic—query market prices, calculate collateral value, subtract target profit margin and costs, and return a bid if profitable. Production bidding models are substantially more sophisticated, incorporating:

• Dynamic gas cost estimation based on current network congestion and transaction complexity
• Slippage modeling for DEX trades required to rebalance capital after auction wins
• Capital availability tracking to avoid bidding more than available wallet balances
• Risk management limiting exposure to individual collateral types or total liquidation participation
• Historical profitability analysis adjusting strategies based on past auction outcomes

Capital Management and DEX Integration

Effective keeper operation requires continuous capital rebalancing between DAI (needed for auction bids) and collateral tokens (received from winning auctions). ^[38] The auction-keeper includes rebalancing modules that automatically convert collateral to DAI through integrated DEXs:

# Configure DEX integration for automatic rebalancing
bin/auction-keeper \
    --type clip \
    --ilk ETH-A \
    --model ./models/strategy.sh \
    --eth-from 0x... \
    --eth-key key_file=... \
    --uniswap-v3-router 0x... \
    --min-eth-balance 1.0 \  # Keep 1 ETH for gas
    --rebalance-target 50000.0  # Maintain 50k DAI

When collateral holdings exceed configured thresholds, the keeper automatically submits DEX swap transactions to convert back to DAI. This automated rebalancing is essential for maintaining auction participation readiness—keepers cannot bid on new auctions if their capital is locked in unsold collateral from previous wins.

Advanced keepers may implement more sophisticated capital strategies including:

• Cross-DEX arbitrage comparing prices across Uniswap, SushiSwap, Curve, and others to minimize slippage
• Just-in-time liquidity using flash loans to participate in auctions without maintaining large idle capital balances
• Collateral lending depositing won collateral in lending protocols to earn yield during holding periods
• Multi-protocol capital pooling using the same capital across multiple keeper strategies (MakerDAO, Aave, Compound liquidations) to improve capital efficiency

Gas Price Strategies and Transaction Management

Gas price management critically determines keeper profitability and was the primary failure mode during Black Thursday. ^[6] Modern keeper implementations use dynamic gas pricing strategies that adapt to network conditions:

class DynamicGasPrice:
    def __init__(self, rpc_url):
        self.w3 = Web3(Web3.HTTPProvider(rpc_url))
        self.base_premium = 1.1  # 10% above median
        self.max_premium = 5.0   # Never pay more than 5x median

    def estimate_gas_price(self, urgency='normal'):
        # Get current network gas prices
        latest_block = self.w3.eth.get_block('latest')
        base_fee = latest_block.baseFeePerGas

        # Query gas price oracles
        median_priority_fee = self.query_gas_oracles()

        # Calculate EIP-1559 gas price components
        if urgency == 'normal':
            priority_fee = median_priority_fee * self.base_premium
        elif urgency == 'high':
            priority_fee = median_priority_fee * self.max_premium
        elif urgency == 'low':
            priority_fee = median_priority_fee * 0.9

        max_fee_per_gas = base_fee * 2 + priority_fee
        max_priority_fee = priority_fee

        return {
            'maxFeePerGas': int(max_fee_per_gas),
            'maxPriorityFeePerGas': int(max_priority_fee)
        }

Effective gas strategies must balance multiple factors. Bidding too low on gas prices risks transaction exclusion, causing missed liquidation opportunities, but bidding too high erodes profitability, particularly on smaller liquidations where gas costs represent significant percentages of total rewards. Keepers implement escalation strategies starting with competitive but not excessive gas prices, then progressively increasing if transactions remain pending.

EIP-1559 transaction format (implemented in August 2021) changed gas dynamics by separating base fees (burned) from priority fees (paid to validators). ^[39] Keepers must estimate both components, with base fee following network congestion trends and priority fees following validator inclusion incentives. During normal conditions, minimal priority fees suffice, but during high-competition liquidation events, keepers may pay substantial priority fees to ensure transaction inclusion ahead of competitors.

Operational Challenges and Best Practices

Running profitable keeper operations over sustained periods requires addressing numerous operational challenges:

Infrastructure Reliability — Node downtime, RPC provider rate limiting, or network connectivity issues cause missed opportunities. Professional keepers run redundant infrastructure across multiple servers, geographic locations, and RPC providers, implementing automatic failover when primary systems experience issues. The infrastructure costs create economies of scale favoring well-capitalized operators who can amortize fixed costs across higher transaction volumes.

Security and Key Management — Keeper accounts require ETH for gas and often hold substantial DAI balances for auction participation. Hot wallet key exposure represents significant theft risk, while excessive security measures (hardware wallets, multi-sig) impede the automated, low-latency transaction submission that competitive keeping requires. Most professional keepers use dedicated hot wallets with limited balance exposures, regularly sweeping profits to secure cold storage, and implementing anomaly detection to identify compromised keys rapidly.

Competitive Dynamics and MEV — As keeper competition intensifies, profit margins compress and successful participation requires increasingly sophisticated infrastructure including MEV-Boost integration, private transaction pools, and direct block builder relationships. These requirements create barriers to entry for new keeper operators and concentration among well-capitalized entities, though the permissionless nature of keeper participation prevents complete market dominance by any single entity.

Regulatory Uncertainty — In some jurisdictions, automated trading bot operation might face regulatory scrutiny under securities trading or money transmission regulations. The regulatory landscape for DeFi remains uncertain, and keeper operators should consider legal counsel regarding their specific jurisdictional requirements, particularly for substantial commercial operations.

Keeper Economics and Profitability

Understanding keeper economics requires analyzing revenue sources, cost structures, competitive dynamics, and how these factors have evolved as the DeFi ecosystem matured and MEV extraction became increasingly professionalized. Keeper profitability has declined substantially from the early MakerDAO days (2017-2019) when individual developers could earn meaningful income running simple keeper bots to today's (2025-2026) intensely competitive environment where sophisticated infrastructure and capital investments are necessary for sustainable profitability.

Revenue Streams and Potential Returns

Liquidation Triggering Fees

The most straightforward keeper revenue comes from flat tip and percentage chip fees earned by triggering liquidations. ^[5] In 2026, these parameters for major collateral types are:

• Flat tip: 100 DAI per liquidation (all collateral types)
• Chip percentage: 2.0-3.0% of liquidated debt (varies by collateral)

During periods of elevated market volatility when numerous vaults approach liquidation thresholds simultaneously, active liquidation triggering can generate substantial income. A keeper successfully triggering 20 liquidations in a single day during a market crash might earn:

• Flat tips: 20 liquidations × 100 DAI = 2,000 DAI
• Chip fees (average 50,000 DAI debt at 2.5%): 20 × 1,250 DAI = 25,000 DAI
• Total revenue: 27,000 DAI ($27,000)

However, this represents a best-case scenario during exceptional volatility. During typical market conditions, liquidation frequency is far lower—perhaps 2-5 liquidations daily across all collateral types—generating more modest 500-1,500 DAI daily revenue before costs.

Auction Arbitrage Profits

The primary profit source for competitive keepers comes from purchasing collateral in liquidation auctions below market value, then reselling through DEXs at market prices. The 13% liquidation penalty creates arbitrage opportunities, though actual keeper captures vary based on auction competition, market liquidity, and timing.

A keeper purchasing 10 ETH (market value $25,000) in a Dutch auction for 21,000 DAI earns gross arbitrage profit of approximately $4,000 before costs. However, realizing this profit requires:

• DEX slippage costs (0.5-2.0% depending on trade size and liquidity): $125-500
• Gas costs for auction take() transaction (200,000 gas at 25 gwei, $2,000 ETH): $10
• Gas costs for DEX swap (150,000 gas): $7.50
• Opportunity cost of capital (hourly APY on 21,000 DAI deployed): minimal for hours-long holding

Net profit after these costs might be $3,400-3,850, representing 16-18% return on the 21,000 DAI capital deployed. These returns seem attractive, but several factors limit actual keeper profitability:

Competition compresses margins—multiple keepers monitoring the same auction bid up prices, reducing arbitrage opportunities. During high-competition events, auctions may settle near fair market value with minimal remaining arbitrage. Capital efficiency constraints limit participation—keepers with limited capital can only participate in a fraction of available liquidations. A keeper with 100,000 DAI working capital might only capture 4-5 of the day's liquidations, limiting absolute earnings despite attractive percentage returns.

Surplus and Debt Auction Participation

Keepers participating in FLAP (surplus) auctions compete to purchase the protocol's excess DAI by offering SKY tokens for burn, effectively buying SKY at a discount. ^[14] If the auction offers 10,000 DAI and the keeper wins by bidding 500,000 SKY (current market value $9,500), they acquire SKY at approximately 5% discount before accounting for transaction costs and market impact when reselling.

These auctions occur less frequently than collateral liquidations—perhaps several times weekly during profitable protocol periods. Profit margins are typically lower (2-5% arbitrage) than collateral auctions, and participation requires holding SKY tokens or capital to purchase them, limiting participation primarily to larger, well-capitalized keepers willing to maintain diverse asset holdings.

FLOP (debt) auctions occur even more rarely, only during crisis events when protocol bad debt exceeds the Surplus Buffer. The last major FLOP auctions occurred during Black Thursday (March 2020), when the protocol minted new MKR to raise approximately $6.65 million. ^[6] While these represent significant value opportunities for participating keepers, their extreme rarity makes them negligible from an expected value perspective for keeper operation planning.

Cost Structures and Breakeven Analysis

Operating a keeper involves both fixed operational costs and variable transaction costs that determine minimum profitable activity levels:

Fixed Operational Costs (Monthly)

• Ethereum full node infrastructure: $400-1,500 depending on hardware specs and redundancy
• Cloud computing resources for monitoring/databases: $200-800
• Data services (The Graph, Dune Analytics, Nansen subscriptions): $100-500
• Development and maintenance time: $2,000-10,000 depending on sophistication
• Total fixed monthly costs: $2,700-12,800

Variable Transaction Costs (Per Liquidation)

• Liquidation trigger gas costs: $5-50 depending on network congestion
• Auction participation gas costs: $5-50 per bid attempt
• DEX rebalancing gas costs: $5-30 per collateral sale
• DEX slippage costs: 0.5-2.0% of transaction value
• Total variable costs per liquidation: $15-130 + slippage

A break-even analysis reveals that sustainable keeper operations require consistent activity volumes. Assuming moderate costs ($5,000 monthly fixed + $30 average variable per liquidation + 1% slippage):

To breakeven, the keeper must generate $5,000+ monthly revenue. At 500 DAI average profit per liquidation (accounting for all costs), this requires approximately 10 liquidations monthly or 2-3 weekly. At 2,000 DAI average profit per liquidation during volatile periods, breakeven requires only 2-3 monthly liquidations.

These calculations suggest small-scale keeper operations can be modestly profitable during sustained volatile periods but face challenges during extended calm markets with minimal liquidation activity. Professional keeper operations achieve economies of scale by:

• Amortizing fixed infrastructure costs across multiple protocols and strategies
• Negotiating better rates for data services and cloud computing at higher volumes
• Developing sophisticated infrastructure that captures larger percentages of available opportunities
• Accessing preferential gas pricing or block inclusion through MEV infrastructure

Historical Profitability Trends and Market Evolution

Keeper profitability has declined substantially as the DeFi ecosystem matured and competition intensified:

Early Period (2017-2019): High Profits, Limited Competition

During MakerDAO's early years, liquidation keeper competition was limited to several dozen individual developers running simple bots. ^[8] Profit margins were substantial—keepers might capture 8-12% arbitrage on liquidations with minimal competition, and gas costs were negligible (median gas prices under 20 gwei with $150-300 ETH prices meaning typical transactions cost under $1). A moderately active keeper during this period could generate $10,000-50,000 monthly profits with basic infrastructure.

Growth Period (2019-2021): Increasing Competition, Declining Margins

As DeFi gained mainstream attention and total value locked grew from hundreds of millions to tens of billions, professional entities entered keeper markets. Competition for liquidations intensified, compressing arbitrage margins from 8-12% to 4-6% by 2021. Gas costs also increased with network utilization, particularly during the 2020-2021 bull market when median gas prices sustained at 50-150 gwei. Combined effects reduced monthly keeper profitability for individual operators to $3,000-15,000 despite higher absolute liquidation volumes.

Professionalization Period (2021-Present): MEV Integration, Consolidated Markets

The emergence of Flashbots and MEV infrastructure fundamentally changed keeper economics by enabling sophisticated actors to guarantee transaction inclusion and priority ordering. ^[7] By 2025-2026, the liquidation keeper market shows heavy concentration:

• Top 10 keeper addresses capture approximately 70-80% of liquidation value
• Most profitable opportunities are captured by MEV firms with advanced infrastructure
• Individual keeper operators increasingly focus on niches (smaller liquidations, exotic collateral types) or exit the market
• Average profit margins compressed to 2-4% arbitrage after accounting for all costs

According to 2025 DeFi research, searchers (including keepers) typically pay more than 90% of their revenue to proposers (validators/block builders) through priority fees and MEV payments. ^[40] In September 2025 alone, arbitrage MEV generated $3.37 million in 30 days, but the majority of this value accrued to validators rather than the searchers who identified opportunities. ^[41]

Current (2026) keeper profitability for individual operators ranges from marginally profitable to unprofitable depending on market conditions, capital availability, and infrastructure sophistication. A well-operated small-scale keeper might generate $1,000-5,000 monthly profit during volatile periods but could operate at losses during calm markets. Only larger, professional operations maintaining multi-protocol strategies and advanced MEV infrastructure achieve consistent profitability across all market conditions.

Black Thursday and Keeper Failures

The March 12-13, 2020 crisis known as "Black Thursday" represents the most severe keeper system failure in DeFi history and fundamentally reshaped liquidation mechanism design across the entire ecosystem. ^[6] Understanding this event provides essential insights into keeper dependencies, system vulnerabilities during extreme conditions, and the importance of robust incentive design for protocol security.

Crisis Timeline and Cascade of Failures

Market Crash and Liquidation Cascade (March 12, 2020)

The crisis began as global financial markets crashed in response to COVID-19 pandemic escalation, with the S&P 500 falling 9.5% and oil prices dropping 25%. ^[42] Cryptocurrency markets followed traditional markets down, with ETH price plummeting from approximately $195 at market open to below $100 by the daily low—a roughly 50% decline in under 24 hours. This sudden collapse triggered liquidations across the MakerDAO protocol as thousands of ETH-backed vaults simultaneously fell below their collateralization thresholds.

The Ethereum network became overwhelmed by liquidation-related transaction volume. Gas prices spiked from typical 20-40 gwei to sustained levels above 200 gwei, with peaks exceeding 500 gwei as users desperately attempted to save their positions by adding collateral or repaying debt. ^[43] The network processed approximately 1,092,168 transactions on March 12—near its theoretical maximum capacity—creating transaction backlogs measured in hours rather than the typical minutes.

Keeper Infrastructure Failures and Zero-Bid Exploits

Most keeper bots at the time used static gas price strategies, configured to submit transactions at fixed gas prices (e.g., 50-80 gwei) or simple multipliers above recent median prices. ^[6] When gas prices exploded past 200 gwei, these keeper transactions were rejected by miners in favor of higher-paying transactions, causing them to sit pending in mempools or be evicted entirely. Keeper operators watching their bots fail attempted manual intervention but faced the same network congestion challenges as regular users.

The catastrophic consequence of keeper failures was that collateral auctions proceeded with minimal or zero participation. In MakerDAO's Liquidation 1.0 English auction system, auctions required competitive bids to drive prices toward fair market value. ^[18] With most keepers unable to submit transactions, auctions often had only a single participant or no participants at all.

Sophisticated actors recognized this unprecedented opportunity and began submitting minimal DAI bids at auctions, knowing that lack of competition meant they would win by default. The most extreme cases saw keepers winning auctions for 0 DAI—literally submitting bids offering zero payment for collateral worth thousands of dollars. ^[19] Of 3,994 liquidation auctions during Black Thursday, 1,462 (36.6%) were won with zero bids, and an additional large percentage had only single non-zero bids well below market value. ^[20]

One address (0xA26e15C895EFc0616177B7c1e7270A4C7D51C997) acquired approximately $8.32 million worth of ETH collateral across 44 auctions for essentially zero payment. ^[44] This entity recognized the keeper failure and systematically submitted zero bids knowing they would win, effectively stealing collateral that should have been sold to repay vault debt. The total protocol loss from zero-bid and severely underpriced auctions reached approximately $8.32 million—creating undercollateralized debt that threatened DAI peg stability.

Mempool Manipulation and Intentional Congestion

Subsequent forensic analysis by Blocknative suggested that at least some network congestion was intentionally manufactured through mempool manipulation rather than organic liquidation activity. ^[45] The researchers identified "hammerbots" that flooded the mempool with transaction spam—repeatedly submitting and replacing transactions that were never intended to confirm, consuming mempool resources and making it harder for legitimate transactions (including keeper bots) to be processed.

This manipulation appears to have been coordinated with the zero-bid exploitation strategy. By intentionally degrading network conditions and preventing keeper participation, sophisticated actors created an environment where they could win auctions with minimal competition. The analysis found patterns of "spontaneous stuck transactions" where certain addresses suddenly experienced transaction failures right before major auction events, then recovered immediately afterward—suggesting deliberate timing to disable competing keepers during critical windows.

The intentional manipulation thesis remains somewhat controversial, as extreme organic congestion alone could explain keeper failures. However, the patterns identified by Blocknative—including the specific timing of congestion spikes with major liquidation events and the concentration of zero-bid wins among a small number of addresses—suggest sophisticated actors exploited and potentially exacerbated network conditions to profit from keeper failures.

Protocol Response and User Compensation Disputes

Immediate Crisis Management

MakerDAO governance responded rapidly to the crisis through a series of emergency actions:

• Debt auctions (FLOP) were initiated on March 19, 2020 to mint and sell new MKR tokens, raising approximately 5.3 million DAI to cover the protocol's bad debt. ^[46]
• The DSR (Dai Savings Rate) was reduced to 0% on March 13 to free up protocol surplus for debt coverage.
• Emergency governance votes temporarily adjusted various risk parameters including debt ceilings, liquidation ratios, and auction parameters.
• The protocol successfully recapitalized through MKR dilution, maintaining DAI peg stability despite the bad debt.

By March 28, all debt auctions had completed, and the protocol's balance sheet was restored to solvency. The crisis demonstrated governance's ability to coordinate rapid emergency response, though it also revealed limitations around timing—the 72-hour GSM delay meant parameter changes required multiple days to implement while markets continued moving.

User Compensation Controversy

Vault owners whose positions were liquidated for zero or minimal DAI sought compensation from MakerDAO governance, arguing the liquidation system's failure constituted a protocol bug rather than expected liquidation risk. ^[47] The debate centered on whether users accept liquidation risk only under the assumption that liquidation mechanisms function as designed—if the mechanism catastrophically failed, should the protocol compensate losses?

In April 2020, governance initially voted to establish a compensation plan for affected users. However, this decision was controversial among MKR holders who would effectively bear compensation costs through protocol funds. A subsequent vote in September 2020 reversed the compensation decision, with 65% voting against compensating the $2.5 million in user losses. ^[48] The rationale held that vault users accept liquidation risk when opening positions, and the protocol successfully maintained DAI peg (the primary obligation) despite liquidation mechanism failures.

Affected users filed a class-action lawsuit against MakerDAO claiming the zero-bid liquidations constituted negligence or system failure warranting compensation. ^[49] After years of litigation, MakerDAO settled in 2023 for $1.16 million—far below actual losses but representing acknowledgment that the liquidation system had not functioned as users reasonably expected. ^[50]

Lessons Learned and Liquidation 2.0 Redesign

Black Thursday prompted fundamental rethinking of liquidation architecture, keeper incentives, and system resilience:

Liquidation 2.0 Dutch Auction Implementation

The most significant response was complete liquidation system redesign using Dutch descending auctions rather than English ascending auctions. ^[21] Dutch auctions solved Black Thursday's core problem—the requirement for competitive bidding that failed during keeper unavailability. By starting prices high and decreasing over time, Dutch auctions ensure that even with a single participating keeper, collateral sells at prices that start above market value and decrease toward reasonable levels, preventing zero-bid exploits.

The DOG and CLIPPER contracts implementing Liquidation 2.0 launched in 2021 and represent the current liquidation standard. The system includes parameters designed specifically to incentivize keeper participation during adverse conditions:

• Flat tip payments (100 DAI) ensure minimum keeper compensation even for small liquidations
• Percentage chip fees (2-3% of debt) provide scaling incentives for large liquidations
• Partial liquidation support enables keepers with limited capital to participate
• Auction reset mechanisms prevent auctions from completing at excessively low prices

Gas Price Strategy Improvements

The keeper community developed sophisticated dynamic gas pricing strategies incorporating real-time network congestion monitoring, automatic gas price escalation for pending transactions, and maximum gas price limits based on expected liquidation profitability. ^[51] Modern keeper implementations treat gas price as a continuous optimization problem rather than static configuration, substantially improving reliability during volatile conditions.

The protocol also benefited from Ethereum's August 2021 London hard fork implementing EIP-1559, which changed gas price mechanics by separating base fees from priority fees. ^[39] This change made gas price estimation more predictable and reduced the likelihood of transactions being completely priced out of blocks as occurred during Black Thursday.

Oracle and Price Feed Improvements

Black Thursday also revealed oracle system vulnerabilities. Price feeds struggled to update during extreme congestion, and the one-hour OSM delay meant liquidation calculations used prices from before the crash in some cases. ^[52] Post-crisis improvements included expanded oracle provider diversity, enhanced OSM design with emergency override capabilities, and generally more robust price feed infrastructure to ensure accuracy during crises.

Cultural Impact on DeFi Design

Beyond specific MakerDAO changes, Black Thursday influenced broader DeFi protocol design philosophy. The event demonstrated that mechanisms must be resilient to extreme conditions—testing under normal circumstances is insufficient. Protocols increasingly focus on worst-case scenarios during design, including explicit analysis of "what happens if keepers fail" or "what happens during sustained network congestion."

The crisis also tempered excessive confidence in algorithmic systems and economic incentives. While keeper-based liquidation systems generally work well, they have failure modes under extreme conditions that must be acknowledged and mitigated through defense-in-depth approaches including backup mechanisms, circuit breakers, and emergency governance response capabilities.

MEV and Keeper Competition

The evolution of MEV (Maximal Extractable Value) infrastructure has fundamentally transformed keeper economics and competition dynamics, shifting the landscape from individual developers running simple bots to sophisticated institutions operating advanced transaction ordering infrastructure. Understanding this intersection is essential for comprehending modern keeper profitability, market structure, and centralization concerns.

MEV Fundamentals and Relation to Keeper Activities

MEV represents the maximum value that can be extracted from blockchain transaction ordering, beyond standard block rewards and transaction fees. ^[53] Validators (formerly miners in pre-Merge Ethereum) control which transactions are included in blocks and in what order, enabling them to capture value through various strategies:

• Arbitrage — Identifying profitable price discrepancies across DEXs and ordering transactions to capture the arbitrage before others
• Liquidation — Triggering undercollateralized position liquidations and participating in resulting auctions
• Sandwich attacks — Placing transactions before and after large user swaps to profit from price impact
• Transaction reordering — Arranging transactions within blocks to maximize extractable value

Keeper activities, particularly liquidation triggering and auction participation, represent legitimate MEV extraction where profit aligns with protocol health. However, the infrastructure enabling MEV extraction—priority transaction ordering, private transaction pools, and validator relationships—creates advantages that extend beyond socially beneficial activities to potentially harmful extraction techniques.

Flashbots and MEV Infrastructure Evolution

Flashbots launched in late 2020 as a response to growing concerns about predatory MEV extraction and rising gas prices from MEV-related transaction competition. ^[54] The system introduced several key innovations:

MEV-Boost and Builder-Proposer Separation

Rather than validators directly optimizing block construction (a complex, resource-intensive task), MEV-Boost enables specialized "builders" to construct optimal blocks and submit them to validators for inclusion. ^[55] Validators simply choose the most profitable block from competing builders, receiving MEV revenue without needing specialized infrastructure. This separation democratizes MEV capture—validators earn MEV revenue proportional to their stake without sophisticated optimization.

For keepers, this infrastructure provides a mechanism to submit transactions with guaranteed inclusion and priority positioning. Rather than submitting transactions to the public mempool where they compete through gas price auctions and face uncertain inclusion timing, keepers submit "bundles" to block builders specifying exact transaction ordering and paying builders directly for inclusion priority.

Private Transaction Pools and Searcher Competition

Flashbots and similar infrastructure (BloXroute, Eden Network) maintain private mempools where searchers (MEV extractors, including keepers) submit transactions that aren't visible to the public until block inclusion. ^[56] This prevents front-running and provides some protection against predatory MEV extraction, though it also introduces centralization by routing significant transaction flow through controlled infrastructure.

Searchers compete for block inclusion not through gas price bidding but through direct payments to builders/validators. A keeper participating in a liquidation might submit a bundle offering to pay the builder 90% of expected profit in exchange for guaranteed inclusion and optimal positioning. This competition for validator attention has resulted in the statistic that searchers typically pay more than 90% of their revenue to proposers, dramatically reducing keeper profitability. ^[40]

Impact on Keeper Market Structure

MEV infrastructure has transformed keeper competition from accessible to specialized, with several notable effects:

Concentration and Winner-Take-Most Dynamics

Keepers with privileged access to block builders, relationships with validators, and sophisticated infrastructure capture disproportionate liquidation opportunities. Data from 2025 shows that the top 10 keeper addresses capture 70-80% of liquidation value, with increasing concentration over time. ^[41] This concentration reflects advantages from:

• Capital scale enabling larger auction participation and faster rebalancing
• Infrastructure sophistication including co-located servers and optimized execution
• Exclusive relationships with block builders for priority access
• Economies of scope from running multiple keeper strategies across many protocols

The concentration raises concerns about keeper decentralization. If liquidation security depends on keeper participation, and keeper markets consolidate to a handful of sophisticated actors, the system becomes vulnerable to those actors' operational failures, regulatory pressures, or potential collusion.

Barrier to Entry Increases

New keeper operators face substantial barriers competing against established MEV firms. Beyond capital requirements for auction participation and infrastructure costs for nodes and monitoring systems, competitive keeping increasingly requires:

• Relationships with block builders for bundle submission access
• Understanding of MEV-Boost infrastructure and bundle construction
• Capital for builder payments (paying 80-90% of profits to builders)
• Sophisticated gas and priority fee modeling
• Multi-protocol strategy development for economies of scope

These requirements favor well-capitalized entities with existing MEV operations, creating path dependency where early MEV infrastructure investments compound into ongoing competitive advantages. Individual developers or smaller teams face diminishing returns from keeper operations, leading many to exit the market or focus on niche opportunities (exotic collateral types, smaller protocols) where sophisticated infrastructure overhead isn't justified by profit potential.

Profit Margin Compression

Even for keepers who successfully adopt MEV infrastructure, profitability has declined due to intense competition. In September 2025, arbitrage MEV generated $3.37 million across the ecosystem, but searchers typically paid over 90% of this revenue to proposers. ^{[40] [41]} If liquidation keeper revenue follows similar patterns, a keeper earning gross profits of $100,000 monthly might pay $90,000 to validators through priority fees and MEV payments, netting only $10,000 before infrastructure and operational costs.

This compression suggests that sustainable keeper operations increasingly require substantial scale to justify fixed costs, or differentiation through specialized capabilities (unique collateral type expertise, regulatory compliance enabling certain market participation) that reduce competition.

Concerns and Criticisms

The MEV-ification of keeper operations has generated significant criticism:

Centralization Risks — Dependence on Flashbots and similar infrastructure introduces centralization points. If these entities experience downtime, regulatory pressure, or capture, keeper operations dependent on their infrastructure could be disrupted. The Ethereum ecosystem has made efforts toward decentralizing builder infrastructure, but significant centralization remains as of 2026.

Value Extraction vs. Value Creation — While liquidation keepers provide clear value by maintaining protocol health, the broader MEV ecosystem includes predatory extraction techniques (sandwich attacks, front-running) that harm users without corresponding benefits. The infrastructure enabling beneficial keeper operations also enables harmful MEV, creating moral and strategic concerns about participating in or supporting this ecosystem.

Barrier to Entry and Decentralization — The increasing sophistication required for competitive keeping contradicts DeFi's permissionless ethos. If participation effectively requires relationships with centralized infrastructure providers and substantial capital, the "anyone can be a keeper" promise becomes hollow, and keeper-dependent protocols become vulnerable to keeper market consolidation.

Despite these concerns, no clear alternative has emerged. The MEV infrastructure provides tangible benefits including reduced gas price volatility, more efficient liquidation execution, and generally improved protocol operations. The challenge lies in capturing these benefits while mitigating centralization risks and preventing predatory extraction—an ongoing tension the ecosystem continues grappling with.

Risk, Challenges, and Future Evolution

The keeper ecosystem faces multiple risks and challenges that could impact its effectiveness, decentralization, and long-term sustainability. Understanding these concerns is essential for assessing Sky Protocol's security model and the broader viability of keeper-dependent DeFi systems.

Technical and Operational Risks

Infrastructure Dependencies and Single Points of Failure

Modern keeper operations depend on multiple centralized infrastructure components including Ethereum RPC providers (Infura, Alchemy), specialized data services (The Graph, Dune Analytics), and MEV infrastructure (Flashbots, BloXroute). ^[57] While keepers can theoretically run entirely independently using only self-hosted nodes, competitive economics push toward these centralized services for superior reliability, performance, and cost-efficiency.

These dependencies create systemic risks. If a major RPC provider experiences downtime during a liquidation cascade, keepers relying on that provider miss liquidation opportunities, reducing auction participation and potentially enabling zero-bid scenarios similar to Black Thursday. The ecosystem has some diversity across providers, but concentration remains—a substantial portion of keeper traffic likely routes through a small number of infrastructure services.

Smart Contract and Protocol Risks

Keepers interact with complex smart contract systems where bugs, upgrades, or parameter changes can impact profitability or even cause keeper losses. The Liquidation 2.0 upgrade in 2021 required keepers to update their implementations to interact with new DOG and CLIPPER contracts rather than legacy CAT and FLIP contracts. ^[21] Keepers who failed to update missed liquidation opportunities during the transition period, and those who updated incorrectly could have experienced transaction failures or losses.

Future protocol upgrades including potential Endgame architectural changes, cross-chain expansion, or new collateral types with novel liquidation mechanics require ongoing keeper adaptation. The burden of staying current with protocol evolution falls on individual keeper operators, creating operational overhead and potential for costly mistakes during transition periods.

Capital Efficiency and Opportunity Costs

Competitive keeper operations require maintaining substantial capital reserves in keeper wallets for auction participation, creating significant opportunity costs from capital that could otherwise be deployed in yield-generating DeFi strategies. A keeper maintaining 250,000 DAI in working capital for large liquidation readiness forfeits approximately 12,500 DAI annually assuming 5% alternative yield—a substantial opportunity cost that must be recovered through keeper profits.

Flash loan integration partially addresses this challenge by enabling keepers to participate in liquidations without maintaining large capital reserves. ^[58] However, flash loan strategies introduce additional complexity, execution risks (atomic transaction failures), and costs (flash loan fees, typically 0.09%), limiting their applicability to certain liquidation types and keeping strategies.

Economic and Game-Theoretic Concerns

Keeper Collusion and Market Manipulation

The concentration of keeper operations among a small number of sophisticated actors creates potential for collusion. If the top 5 keeper operators controlling 70%+ of liquidation capture coordinated to delay liquidation triggering, they could allow vaults to become more severely undercollateralized before liquidating, extracting greater profit at the expense of vault owners and potentially protocol security. ^[59]

Such collusion faces coordination challenges including individual incentives to defect (triggering liquidations early to capture rewards before competitors), difficulty maintaining secret coordination (on-chain activity is public), and potential governance response (adjusting incentive parameters to restore competition). However, the technical feasibility exists, and the increasing sophistication and concentration of keeper markets makes coordination more plausible than in earlier, more distributed keeper ecosystems.

Adverse Selection and Crisis Dynamics

Keeper profitability is highest during crisis conditions when liquidation volumes spike, creating perverse incentives where keepers potentially benefit from protocols experiencing distress. While individual keepers cannot cause market crashes, the economic alignment creates systemic fragility—the keeper ecosystem is optimized for crisis rather than prevention, and keepers lack incentive to support positions approaching liquidation through methods other than profit-maximizing liquidation execution.

This dynamic differs from traditional finance where banks and asset managers have reputational and relationship incentives to support struggling borrowers through workouts, extensions, or restructurings. DeFi's permissionless, pseudonymous nature eliminates these reputational mechanisms, leaving pure profit maximization—an alignment that generally serves protocol security but may create adverse effects during boundary cases.

Long-Term Profitability Sustainability

The compression of keeper profit margins from 8-12% in early MakerDAO to 2-4% in 2026, combined with increasing infrastructure requirements and competition, raises questions about long-term keeper market sustainability. ^[60] If margins compress further or capital requirements increase, keeper participation could decline, reducing liquidation security and requiring protocols to adjust incentive parameters (increasing tip/chip rates or liquidation penalties).

However, this concern is partially self-correcting through market dynamics. If keeper participation declines due to low profitability, liquidation competition decreases, improving margins for remaining keepers and attracting new entrants. The market should theoretically reach equilibrium where keeper profitability is just sufficient to sustain the minimum participation necessary for protocol security, though this equilibrium may involve fewer keepers than ideal from a decentralization perspective.

Regulatory and Legal Uncertainties

Transaction Execution and Securities Trading Implications

Keeper operations involve automated trading algorithms that might face regulatory scrutiny under securities or commodities trading regulations in certain jurisdictions. While DeFi protocols themselves claim decentralization as regulatory protection, individual keeper operators—particularly large, professional entities—present clearer regulatory targets. ^[61]

Regulatory authorities could potentially classify keeper activities as requiring trading licenses, impose know-your-customer (KYC) requirements on keeper operators, or demand transaction reporting and audit trails. Such requirements would substantially increase operational complexity and costs, potentially driving keeper operations to favorable jurisdictions or underground, but reducing keeper participation in regulated markets.

Tax Treatment and Reporting Obligations

Keeper profits face complex tax treatment varying by jurisdiction. Each liquidation trigger, auction participation, and collateral resale potentially constitutes a taxable event requiring capital gains calculation and reporting. ^[62] The high frequency of keeper operations—potentially thousands of transactions annually for active operators—creates substantial compliance burden and potential for errors that might expose keeper operators to penalties or audits.

Professional keeper operations typically work with specialized crypto tax advisors and use automated tax tracking systems, but these services add operational costs. Individual keeper operators without sophisticated tax compliance infrastructure face risks of underreporting income or miscalculating tax obligations, creating legal exposure.

Future Evolution and Adaptation

Layer 2 and Cross-Chain Expansion

Sky Protocol's expansion to Layer 2 networks including Base, Arbitrum, and Optimism introduces new keeper challenges and opportunities. ^[63] L2 networks offer dramatically lower gas costs (often under $0.01 per transaction compared to $5-50 on mainnet), potentially making smaller liquidations economically viable and lowering barriers to keeper entry. However, L2s also introduce complexity from:

• Operating keeper infrastructure across multiple chains
• Managing capital and rebalancing across bridge systems
• Understanding L2-specific technical differences (sequencers, forced inclusion delays)
• Navigating fragmented liquidity for collateral resale

The early evidence from 2025 suggests that MEV activity has already established strong presence on L2s, with over 50% of on-chain gas consumed by optimistic MEV activity on Base and Optimism. ^[64] This indicates that keeper professionalization and MEV infrastructure extend to L2s rather than creating new opportunities for decentralized participation.

Automation Improvements and Keeper-as-a-Service

The success of vault automation services like DeFi Saver and Instadapp suggests potential for "Keeper-as-a-Service" business models where protocols or specialized entities offer managed keeper infrastructure for a fee. ^{[26] [31]} Rather than depending on permissionless but potentially unreliable keeper participation, protocols could contract with professional keeper service providers guaranteeing specific performance levels, reliability standards, and capital availability.

This approach trades some decentralization for operational reliability and predictability. Protocol-funded keeper services ensure liquidations execute reliably regardless of market profitability but require ongoing protocol expenditure and introduce dependencies on service providers. The Chainlink Automation relationship for chief-keeper execution exemplifies this model, suggesting it may expand to other keeper functions over time. ^[3]

AI and Machine Learning Applications

Emerging applications of AI and machine learning to keeper strategies could enable more sophisticated bidding models, gas price prediction, and multi-protocol optimization. ^[65] AI-driven keepers might:

• Predict liquidation probability and prepare capital before positions become liquidatable
• Optimize auction bidding based on historical patterns and competitor behavior
• Dynamically adjust strategies based on market regime classification
• Identify complex multi-step MEV opportunities spanning multiple protocols

However, AI integration could also exacerbate centralization by favoring entities with data science expertise, computational resources, and training data access. The capital and technical requirements for competitive AI-driven keeping might further concentrate the keeper market, though open-source AI keeper frameworks could partially democratize access to these capabilities.

Related Topics

Understanding keepers fully requires familiarity with the broader Sky Protocol ecosystem and related DeFi concepts:

• Sky Vaults — The collateralized debt positions that keepers monitor for liquidation eligibility, including vault mechanics, collateral types, and user strategies
• Liquidations — The complete liquidation process including triggering conditions, auction mechanisms, and vault owner protections
• Oracles — The price feed systems that provide the data keepers use to evaluate vault health and collateralization ratios
• Sky Protocol — The overall protocol architecture within which keepers operate, including governance, economics, and technical infrastructure
• Spells — The governance execution mechanism that chief-keeper automates, including spell lifecycle and security features

Conclusion

Keepers represent an elegant solution to the coordination problem of maintaining decentralized protocol health—by creating profit opportunities for external actors to perform critical maintenance operations, Sky Protocol ensures liquidations execute, auctions clear, and governance implements without trusted intermediaries. The keeper model embodies DeFi's fundamental insight that economic incentives can reliably coordinate decentralized systems when designed thoughtfully.

However, the keeper ecosystem's evolution from accessible side-income for individual developers to sophisticated MEV extraction requiring institutional infrastructure raises important questions about decentralization, accessibility, and long-term sustainability. The same mechanisms that ensure keeper reliability during normal conditions—strong profit incentives, competitive markets, sophisticated infrastructure—create winner-take-most dynamics during crisis conditions when keeper participation is most critical.

As Sky Protocol continues evolving through Endgame implementation, cross-chain expansion, and ecosystem growth, keeper infrastructure must adapt to new collateral types, L2 liquidations, and potentially novel auction mechanisms. The tension between permissionless participation and competitive effectiveness will likely persist, with the ecosystem seeking balance between broad keeper access and reliable liquidation execution that protects protocol solvency.

Understanding keepers provides essential insight into how DeFi protocols achieve operational reliability without centralized administrators—a fundamental innovation that has enabled billions in value to be secured by economic incentives rather than trusted parties. Whether this model can scale to institutional adoption while maintaining meaningful decentralization remains an open question that keeper evolution will help answer.

Sources

1. Sky Atlas A.3.7.1.3 - Keepers - Atlas documentation on keeper networks and compensation
2. The Auctions of the Maker Protocol - Technical documentation on auction types and keeper participation
3. Sky Atlas A.3.7.1.3.3 - Chainlink Automation - Chainlink Automation compensation for chief-keeper services
4. DeFi Saver — Advanced DeFi Management - Vault automation service for liquidation protection
5. Liquidation 2.0 Module - DOG and CLIPPER contract documentation with tip/chip parameters
6. Black Thursday for MakerDAO: $8.32 million was liquidated for 0 DAI - Comprehensive Black Thursday analysis
7. MEV: A 2025 guide to Maximal Extractable Value in crypto - MEV fundamentals and ecosystem overview
8. Maker - Keepers incentive-following bots - Original keeper documentation from MakerDAO
9. Sky Atlas A.1.9.2.4.13.2.1 - Chief-keeper - Chief-keeper automated spell execution
10. Liquidation 2.0 Module - Dog Contract - DOG contract bark() function documentation
11. Auction Keeper Bot Setup Guide - Infrastructure requirements for keeper operations
12. Liquidation | MakerDAO Community Portal - Liquidation process and Dutch auction mechanics
13. What Really Happened To MakerDAO? - Glassnode Black Thursday analysis and Liquidation 2.0 motivation
14. Surplus and Debt Auctions - FLAP and FLOP auction mechanics
15. GitHub - chief-keeper - Chief-keeper source code and implementation
16. Keeper Economics - Economic incentives for various keeper functions
17. Can You Earn Money Running an Ethereum Node in 2025? - Infrastructure costs and profitability analysis
18. Performing Liquidations on MakerDAO - Historical liquidation mechanics and English auction system
19. Black Thursday Zero-Bid Exploits - Detailed analysis of zero-bid auction wins
20. Evidence of Mempool Manipulation on Black Thursday - Forensic analysis of Black Thursday transactions
21. Liquidation 2.0 Module Introduction - Dutch auction implementation and improvements
22. CLIPPER - Detailed Documentation - Partial liquidation and auction mechanics
23. Auction Parameters - Governance-set auction curve parameters
24. Liquidation Penalty Documentation - Liquidation penalty mechanics and implications
25. GitHub - auction-demo-keeper - Simple FMV discount bidding model implementation
26. DeFi Saver Documentation - Automation features and capabilities
27. DeFi Saver Introduces Stop Loss and Take Profit options - Automation feature announcement
28. Comparing MakerDAO management apps - DeFi Saver competitive analysis
29. DeFi Saver Pricing and Fees - Service fee structure
30. All You Need to Know about InstaDApp - Instadapp overview and features
31. A comparison with Defi Saver & Instadapp - Gelato Automation partnership announcement
32. Comparing MakerDAO management apps— Oasis, Zerion, DeFi Saver and InstaDapp - Cross-protocol migration capabilities
33. InstaDApp DSL and DSA Architecture - DeFi Smart Layer technical overview
34. DeFi Saver Security Audits - Smart contract security information
35. GitHub - auction-keeper - Primary auction-keeper repository
36. auction-keeper README - Keeper architecture and design
37. auction-keeper/model.py - Bidding model interface specification
38. Auction Keeper Rebalancing - Capital management features
39. EIP-1559: Fee Market - EIP-1559 gas price mechanics
40. How to Maximize Returns in 2025 - Searcher/proposer revenue split statistics
41. MEV Trading Strategies - MEV profitability data September 2025
42. Black Thursday Market Context - Global market crash triggering the crisis
43. Gas Fee Markets Statistics 2025 - Historical gas price data
44. Zero-Bid Winner Address Analysis - Specific exploiter address and tactics
45. Mempool Manipulation Enabled Theft of $8M - Blocknative research on intentional congestion
46. MakerDAO Debt Auctions - Protocol recapitalization through MKR minting
47. MakerDAO Users Hosed by March Flash Crash - User compensation debate
48. MakerDAO Governance Vote Against Compensation - September 2020 governance decision
49. Class Action Lawsuit Filed - Legal action by affected users
50. Maker Settles for $1.16M - Settlement announcement
51. How to Run a Keeper Bot - Modern keeper implementation guide with gas strategies
52. Oracle Security Module - OSM delay mechanism and crisis performance
53. What Is MEV in Crypto? - MEV definition and extraction methods
54. Flashbots Documentation - MEV infrastructure overview
55. MEV-Boost - Builder-proposer separation mechanics
56. Private Transaction Pools - Flashbots Protect and private mempools
57. Ethereum Infrastructure Centralization Risks - RPC provider concentration concerns
58. Flash Loan Integration Strategies - Using flash loans for capital efficiency
59. Keeper Collusion Game Theory - Academic analysis of keeper coordination incentives
60. Gas Fee Volatility Statistics 2025 - Keeper margin compression analysis
61. DeFi Regulatory Considerations - Regulatory uncertainty for automated trading
62. Crypto Tax Treatment - Tax implications for keeper operations
63. SkyLink Cross-Chain Infrastructure - L2 expansion plans
64. L2 MEV Activity Statistics - Base and Optimism MEV data Q1 2025
65. AI Applications in DeFi - Machine learning for keeper strategies