What Is an Exploit Bounty in Crypto?

An exploit bounty in crypto means one of two things. It is either a standing reward for security researchers who find and report a bug before an attack. Or it is a post-hack offer made straight to an attacker, asking them to return stolen funds in exchange for keeping a cut.

Both uses are real. Both show up in DeFi headlines all the time. The first is what most people call a bug bounty. A protocol pays anyone who finds a flaw and reports it responsibly, before any harm is done. The second is what Crypto Twitter usually means by “exploit bounty.” A team just lost $50M, so it sends an on-chain message to the hacker’s wallet offering 10% to send the rest back.

Same words, two very different situations. This guide covers both, with real numbers and real cases from 2023 through 2026.

Key Takeaways

• An exploit bounty covers two scenarios: a pre-hack reward for ethical vulnerability disclosure, and a post-hack on-chain offer asking an attacker to return stolen funds.
• Pre-exploit programs on platforms like Immunefi pay anywhere from a few thousand dollars to $10M for critical bugs, depending on the protocol’s size and published caps.
• Post-exploit bounty offers succeed occasionally — Euler Finance’s hacker returned everything in 2023 — but overall only about 4.6% of stolen funds were recovered via negotiated bounties in 2025.

What an Exploit Bounty Actually Means

The term “exploit bounty” covers two real-world scenarios. Mix them up and you will misread half the headlines in the space.

The first meaning is proactive. A protocol runs a standing security program, usually on a platform like Immunefi, Sherlock, or HackenProof. It pays researchers who find bugs in its smart contracts before anyone gets hurt. Cybersecurity veterans call this a bug bounty. In crypto it works the same way. You find the flaw, report it privately, and get paid. The program stays live whether or not funds are ever stolen.

The second meaning is reactive. A hack happens, funds leave, and the team posts an on-chain message to the attacker’s wallet. The deal is usually simple: keep 10–20%, return the rest, and no legal action follows. Crypto Twitter calls this an exploit bounty or a “bounty offer.” There is no standing program here. It is a live negotiation with someone who already took the money.

Both meanings are legitimate. But they serve different purposes with different odds. A standing bug bounty protects a protocol in advance. A post-exploit offer is damage control after the fact, and it often fails.

One more thing worth clearing up. Neither of these is a marketing bounty. That is a separate category where projects pay users to translate docs, share posts, or complete small tasks. Marketing bounties have nothing to do with security research.

How a Pre-Exploit Bounty Works

A pre-exploit bounty is the classic bug bounty that crypto programs run. It starts before any attack. The protocol works with an established platform to set the rules, define payout tiers, and list which contracts are in scope. Then researchers can go hunting.

The process runs in a predictable order. A researcher spots something suspicious in an in-scope contract. They write a proof of concept (PoC), a reproducible demo that the bug is real and exploitable. They file the PoC through the platform’s portal. The project’s security team reviews it and assigns a severity rating using the CVSS scale or a similar system. If the bug is valid, in scope, and not a duplicate, the researcher gets paid, usually in USDC or ETH.

The stakes at the top end are serious. The Ethereum Foundation raised its critical-bug payout cap to $1M in March 2025, a new standard for mainnet programs. Immunefi, the largest web3 bug bounty platform, reports paying out over $100M since launch. It has also noted a rough truth of the job: 50–70% of incoming submissions are duplicates, out of scope, or not reproducible.

Severity tiers set the payout range. Most programs follow a four-level structure:

Severity tier	Typical payout range
Critical	$50,000–$10,000,000+
High	$10,000–$100,000
Medium	$1,000–$10,000
Low	$100–$1,000

Real numbers swing with protocol size and budget. Flagship programs like Wormhole and the Ethereum Foundation sit at the top. Smaller DeFi protocols often cap critical bugs at $50,000 or less.

One rule catches first-timers off guard. Immunefi requires government-issued ID before it releases a critical payout. That KYC step exists to satisfy legal compliance. Anyone expecting a fully anonymous process finds out fast that it is not.

There is also an ongoing debate over the “scaling bounty” model. Here, bounty caps track a percentage of TVL instead of a fixed dollar amount. Immunefi has pushed this framework publicly, arguing that a protocol with $500M TVL should offer higher ceilings than one with $5M. The idea has traction, but it is not yet standard across programs.

How a Post-Exploit Bounty Offer Works

When a protocol gets hacked, the team’s first public move is often an on-chain message sent straight to the attacker’s wallet. Anyone can read it on a block explorer like Etherscan. The structure is recognizable. Acknowledge the exploit, state how much was taken, offer the attacker a cut to return the rest, and promise no legal action if they comply. This is the on-chain bounty offer, a DeFi security bounty of sorts, except aimed at someone who already pulled off the exploit.

Cetus Protocol’s May 2025 hack is a clear recent example. The project lost $223M. Soon after, Cetus sent the attacker an on-chain message offering $6M to return 20,920 ETH. That worked out to roughly a 10% keeper deal on the stolen amount.

Euler Finance’s 2023 case went the other way. After a $197M hack, the attacker returned nearly all of the funds following long on-chain negotiation. No legal resolution, no public prosecution, just a return. The 1inch case in March 2025 followed the same path. Around $5M was taken, and most of it came back after negotiation.

Not every offer lands. The IoTeX bridge exploit in February 2026 showed what happens when an attacker rejects the standard terms. The protocol made a 10% bounty offer. The hacker refused, then counter-demanded full control of the protocol. No funds came back.

So why make these offers at all? The math is simple. Recovering 80–90% of stolen funds beats recovering nothing. On-chain transparency adds pressure too. The attacker’s wallet is visible, later transactions are traceable, and cashing out large amounts through regulated exchanges is hard. But the tactic is far from reliable. Industry data from 2025 shows only about 4.6% of all stolen funds were recovered through negotiated bounty arrangements. Attacks involving shadow mints, where attackers conjure tokens from nothing before dumping them, are especially hard to reverse. There is nothing clean to “send back.”

What an Exploit Bounty Pays — And What It Doesn’t

Payouts from exploit bounty programs run from almost nothing to some of the largest single bug rewards in software history. The gap between the two is where most of the confusion lives.

Start with the headline numbers. Wormhole’s $10M payout in 2022 remains the largest single crypto bug bounty on record. Aurora’s $6M payout that same year protected roughly $200M at risk. Usual Protocol launched a $16M bounty pool on Sherlock in 2026, the largest available reward pool by program size. And the Ethereum Foundation’s critical consensus bug cap now sits at $1M after the March 2025 revision.

The average tells a quieter story. Across all programs on Immunefi, the typical critical payout lands around $13,000. That number reflects the many smaller protocols that cannot afford flagship rewards. Chase high-value targets instead and you find bigger programs. Think major bridge protocols, Layer 1 consensus clients, or top-ten DeFi protocols by TVL. You also find more competition there.

A few real catches shape the job for first-time bug hunters. Most reports get rejected. The usual reasons are out-of-scope contracts, duplicates from other researchers, and bugs the team calls too low-impact for the stated tier. Approved payouts can take weeks to land. And the payout you expect is not always the payout you get.

The Injective case from 2024 showed this the hard way. A researcher known as f4lc0n found a vulnerability that the program documentation implied could pay far more than $50,000. The potential loss was estimated at $500M. Injective offered $50,000. The researcher disputed it and reportedly waited months for any resolution. This is not a rare edge case. Projects have reclassified bug severity after a report comes in, which drops the payout tier and cuts the amount owed. Immunefi runs a dispute process for exactly this, but it is slow.

So do the boring homework before the fun part. Read the in-scope document carefully. Check the published cap for each severity tier. Verify which specific contract versions are covered. A tight search on a high-payout program beats a wide search across dozens of small ones.

The Controversy Around the Exploit Bounty Model

Three real tensions run through the exploit bounty space. All three come up often on Crypto Twitter.

The first is underpayment. The f4lc0n/Injective case is the most-discussed example from 2024, but the complaint is older than that. White hat researchers keep reporting bugs that could drain millions. Then they get offers that feel far below the stated caps. When a protocol downgrades a severity rating after the fact, there is no clean remedy. Immunefi’s dispute process exists, but it can drag on. And by the time the dispute starts, the researcher has already handed over the vulnerability.

The second tension is philosophical. Is a post-exploit bounty offer just legalized extortion? A 2024 CoinDesk opinion piece put it sharply, arguing protocols need a “sheriff, not a bounty hunter.” The concern is fair. If a protocol pays an attacker 10% and declines to prosecute, what just happened? Critics call it a moral hazard. Defenders say it is the only practical way to recover funds when chains are pseudonymous and cross-border prosecution crawls.

The third tension is the perverse-incentive problem. If attackers can keep 10–20% of what they take with no legal consequence, does that lower the real cost of trying an exploit? The data does not settle this yet, but it is worth naming. Industry figures from 2025 show only about 4.6% of stolen funds came back through bounty negotiations. So the vast majority of attackers took the money and kept it, whatever the offer.

None of this sinks the whole model. Pre-exploit bug bounty programs have caught real vulnerabilities before harm was done. Post-exploit offers have clawed back real funds. But the space is messier than the headline “we offered the hacker a bounty” suggests. At one end sit protocols that hard rug after an exploit and never pay a bounty at all. At another sit protocols that make good-faith offers and watch the attacker vanish. And protocols paying far below their published caps sit somewhere uncomfortable in between.

Knowing these failure modes helps two groups. It helps anyone judging a protocol’s security posture, and anyone deciding whether to report a bug. A program that looks generous can behave differently under pressure. In the same way, a soft rug, where a team quietly walks back commitments without a dramatic exit, can look a lot like the post-exploit ghosting that frustrated researchers describe.

How to Find and Submit an Exploit Bounty

For most security researchers, the practical starting point is Immunefi. It is the largest DeFi-focused bug bounty platform, with hundreds of active programs in one place. Sherlock is the next major option, running both bug bounties and audit contests. HackenProof covers a wider range of web3 targets. HackerOne hosts big programs from centralized exchanges, including Coinbase and Kraken.

Whatever the platform, a valid submission needs the same core parts:

• A reproducible proof of concept that demonstrates the vulnerability
• A severity justification using CVSS or the platform’s own scoring rubric
• Confirmation that the target contract or component is in scope
• No prior disclosure to the project or any third party before submission

That last point trips people up. On Immunefi, contacting the protocol directly before you file a formal submission breaks program terms and can void the payout. The platform is your first contact, not the team.

KYC then applies before any critical payout goes out. Immunefi requires government-issued ID at that stage. Anyone expecting a pseudonymous process should plan for it. It does not lock you out of participating, but your identity gets tied to your payout.

Skills matter as much as luck. Strong bug hunters know Solidity well and understand common attack vectors, including reentrancy, access control flaws, and oracle manipulation. They also write clean PoCs a reviewer can reproduce in under an hour. The top earners on Immunefi tend to specialize. They pick one chain or one class of protocol and build pattern recognition through depth, not breadth.

One habit pays off more than any other. Favor programs where the scope is clear, the payout history is public, and the bug-to-payout ratio looks credible from past disclosures. Trying to frame volume spoofing or other market manipulation as an in-scope exploit rarely works. Most smart contract programs scope around on-chain contract logic, not market behavior.

So before you write a single line of test code, read the full scope document. Check which contract addresses are listed. Confirm the severity tiers and the maximum payout for each. Then start.

FAQ

What is the difference between an exploit bounty and a bug bounty?

A bug bounty is a standing program that rewards anyone who finds and reports a security flaw before an attack. An exploit bounty, in its common Crypto Twitter sense, means the post-hack offer made to a real attacker. It asks them to return stolen funds in exchange for keeping a cut and dodging legal action. The terms overlap casually, and some projects also call their pre-hack programs exploit bounties. Context usually makes the meaning clear.

Do hackers actually return funds when offered an exploit bounty?

Sometimes. Euler Finance got back nearly all of the $197M taken in its 2023 hack after on-chain negotiation. The 1inch hacker returned most of the roughly $5M taken in March 2025. Cetus Protocol’s $6M exploit bounty offer for 20,920 ETH led to a partial resolution. IoTeX’s February 2026 offer was rejected outright. Across 2025, industry data shows only about 4.6% of stolen funds came back through negotiated exploit bounty arrangements. Success depends on the attacker’s motives, how traceable the funds are, and whether steady on-chain and legal pressure makes returning them the easier path.

Is it legal to claim an exploit bounty in crypto?

It depends on how the funds were obtained and which jurisdiction applies. Finding and reporting a vulnerability through a program’s authorized scope, with a signed disclosure agreement, is generally legal and treated as protected security research. Exploiting a contract first and then negotiating an exploit bounty to return funds is legally murky. Some jurisdictions still count the initial exploit as theft, even if the funds come back later. On-chain anonymity gives no legal cover. Responsible disclosure, meaning you report the flaw without exploiting it, is the only path with clear legal protection. Anyone in a grey area should get legal advice before proceeding.

How much does an exploit bounty pay for a critical bug on Immunefi?

It varies by program. Across all Immunefi programs, the reported average critical payout sits around $13,000, dragged down by the many smaller protocols with limited budgets. Flagship programs pay far more. The Ethereum Foundation pays up to $1M for critical consensus bugs as of March 2025. Wormhole paid $10M for a single critical report in 2022. Your safest move is to check the program page for its published cap, then read its payout history to see whether the stated maximum has ever actually been hit.

Can a protocol refuse to pay an exploit bounty even after confirming the bug?

Yes. Protocols have disputed severity ratings after a report comes in, reclassifying the bug to a lower tier and paying less than the researcher expected. The Injective case in 2024 is the most-discussed example. f4lc0n found a vulnerability with potential losses estimated at $500M and was offered $50,000. The project disputed the severity, and resolution took months. Immunefi’s dispute process exists for exactly this, but it is slow and the outcome is not guaranteed. Checking a program’s payout history and community reputation before you sink real research time into it lowers this risk.

Where To Start

Came here after a DeFi hack headline? Start by working out which scenario it describes: a pre-hack bug reward or a post-hack fund-return offer. The two exploit bounty types look alike on the surface but work in opposite directions. The protocol’s public messages usually make the difference clear within a few sentences.

If you want to hunt bugs as a researcher, the steps below are the right sequence. Skipping any of them, especially reading the scope document, is the top reason a valid finding still ends in a rejected submission.

Here is a concrete starting sequence:

• Create an account on Immunefi and browse active programs filtered by chain and payout range.
• Read the scope document of any program you are considering before writing any test code.
• Study past disclosures on Immunefi’s public leaderboard to understand which vulnerability types get accepted on your target protocol.
• Practice writing clean, reproducible PoCs in a local fork environment before submitting.
• Prepare for KYC at the payout stage — have government ID ready so an approval does not stall.

If you are a DeFi user rather than a researcher, the takeaway is simpler. Protocols with active, well-funded exploit bounty programs on reputable platforms are measurably safer than protocols with no public security program at all. A visible bug bounty is no guarantee. The Injective case shows even well-funded programs can frustrate good-faith researchers. Still, no program at all is a red flag worth noting before you put real money into a contract.

One last check is worth building into your routine. Look up whether the protocol has actually paid out any bounties. Immunefi publishes payout records. A program that has never paid carries a different risk profile from one with a visible track record. Same label, very different assurance.