News & Updates

2018-10-16: Peckshield has climbed higher on the leaderboard, with another 5K points for geth DoS vulnerabilities, and 1200 points for Constantinople flaws. Also, Feeker earned 500 points for a Geth DoS via RPC.
2018-08-09: We have updated the scope of the bounty program, to be more explicit about what we are interested in – what is included in the scope and what is not.
2018-08-02: PeckShield has been awarded 12000 points for three separate reports.
- Geth DoS via discovery protocol (2000 points)
- LES Server DoS via malicious query (5000 points)
- LES Server DoS via AnnouceMsg (5000 points)
2018-03-28: Several new entries on the leaderboard! Congratulations, and thanks!
- First off, we have awarded 10 000 points to the researchers from Boston University: Sharon Goldberg, Yuval Marcus and Ethan Heilman, for their research about eclipse attacks on geth nodes. The fixes were included in 1.8.0.
- Dominic Brütsch was awarded 7500 points for a vulnerability which could be used to trigger slow block processing in geth, which has been fixed in 1.8.3.
- Two DoS vulnerabilities in geth RPC methods were reported by Vasily Vasiliev, awarded 500 points each.
- We also awarded ‘jazzybedi’ with 500 points for alerting us about DNS rebind vulnerabiliites, which were also fixed in 1.8.0.
2018-02-12: Barry Whitehat has been added to the leaderboard, for a discrepancy in how Geth vs Parity treated ‘future’ blocks. The discrepancy could potentially cause a mining minority be at a disadvantage. This has been fixed by aligning how Geth and Parity treats such blocks.
2018-01-16: See the blog for a security announcement concerning the Mist Browser. The Mist browser is not considered production software, and we will not pay full rewards for upstream bugs.
2017-12-13: Yoonho Kim, of team Hithereum, has scored another 5000 points; again for an RCE in Mist/Electron. Also, Peter Stöckli submitted a Mist-vulnerability which granted him 5000 points. Congratulations both! A very important reminder: the Mist browser is not secure for browsing the internet. Marcin Noga of Cisco/Talos security also submitted some issues to cpp-ethereum, which gained him 500 points.
2017-11-28: Juno Im, has scored another 5000 points; again for an RCE in Mist/Electron.
2017-10-09: Yoonho Kim, of team Hithereum, has scored 15000 points for a 0-day vulnerability (remote command execution) in Mist/Electron, which was subsequently patched upstream and made into the last Mist-release. Juno Im has been awarded another 500 points for a Geth access control issue.
2017-09-19: In order to get some extra eyes on the Byzantium implementations, we’ve temporarily increased the rewards: Between now and the Byzantium mainnet hardfork, we will double the ratio of points-to-USD for any vulnerabilities affecting cross-client consensus or Geth denial-of-service. A ‘High’ can thus yield up to $30K USD, and ‘Critical’ up to $50K USD. All Byzantium functionality is considered in-scope, as if it was already enabled on the mainnet.
2017-09-14: Harry Roberts has been awarded 5000 points for discovering a bug in how Solidity implemented ecrecover. See release notes for v0.4.14 for further details.
2017-07-28: Juno Im has been awared 5000 points for a Mist-vulnerability regarding importing of maliciously crafted wallet-files.
2017-05-31: Whit Jackson has been awarded 2000 points for hex-encoding ambiguities in EthereumJS, Christoph Jentzsch has been awarded 2000 points for the solidity optimizer bug, and ‘Tintin’ was awarded another 2000 points for a bug in a third-party component for CPP-ethereum.
2017-05-02: Yaron Velner has been awarded 1000 points for an ENS-submission, where by ENS second price could be manipulated via replay, forcing winners to pay the full amount offered.
2017-04-07: EthHead and Steve Waldman have been added to the leaderboard for their ENS findings. Bug 1 and bug 2
2017-04-07: ENS is now officially included in the program.
2016-12-01: Solidity is now officially included within the bug bounty program.
2016-11-10 We’re please to have three new names on the leaderboard, Bertrand Masius (Solidity bug), tintinweb (Mist vulnerabilities) and Yaron Velner (EXP opcode mispriced).
2016-07-15: The Ethereum hard fork code is in scope of the Ethereum bounty program. Please see the latest hard fork specification.
2016-01-26: BTC RELAY is now in scope of the Ethereum bounty program. Please see BTC RELAY Bounty Program and BTC RELAY Spec for more info and exact scope.
2015-09-02: With Martin (@mhswende) finding another consensus protocol bug in the Python client, he’s now climbed ahead of nickler and we have a new leader on the leaderboard! We’ve also clarified reference to the Python client and it’s scope within the bounty program (see link below in the references).
2015-07-30: As we are launching Frontier, we will continue the bounty program throughout and at least until Homestead. One extension, and one change: From now on, core CPP libraries will be in scope as well. The genesis block inscription reward is altered to an entry in the namereg. Happy hunting!
2015-06-11: As the Ethereum clients are becoming more stable and secure, we’re happy to announce Proof-of-Work (Ethash) and the Go P2P implementation are now also in scope and eligibile for rewards. The develop branch is the target.
2015-03-19: The bounty program will remain running for at least the duration of the upcoming Ethereum frontier release. Please see the Ethereum blog for more information about Frontier!
2015-02-27: These scripts by Jonas Nick can be helpful to build the Ethereum Go client and test it. Please see the bash scripts for build commands and the python script for a simple example of calling the JSON-RPC API. Please note the currently known issues
2015-02-27: Another major vulnerability found by Jonas Nick. Awarded with 5 BTC, this exploit triggers a bug in the Ethereum VM to create ether out of thin air.
2015-01-30: Friendly reminder: Ethereum websites are out of scope for the bounty program and not eligible for rewards. With that said, we are thankful for submissions relating to webpage security and will work to fix these issues.

RULES & REWARDS

Please have a look at the bullets below before starting your hunt!

Issues that have already been submitted by another user or are already known to the Ethereum team are not eligible for bounty rewards.
Public disclosure of a vulnerability makes it ineligible for a bounty.
You can start or fork a private chain for bug hunting. Please respect the Ethereum main and test networks and refrain from attacking them.
Ethereum’s core development team, employees and all other people paid by the Ethereum project, directly or indirectly, are not eligible for rewards.
Anyone who works with the codebase as a professional Ethereum developer is not eligible for rewards.
Ethereum websites or Ethereum Foundation infrastructure in general, are NOT part of the bounty program.
Ethereum bounty program considers a number of variables in determining rewards. Determinations of eligibility, score and all terms related to an award are at the sole and final discretion of the Ethereum Foundation bug bounty panel.

The value of rewards paid out will vary depending on Severity. The severity is calculated according to the OWASP risk rating model based on Impact and Likelihood :

severity

Reward sizes are guided by the rules below, but are in the end, determined at the sole discretion of the Ethereum Foundation bug bounty panel.

Critical: up to 25 000 points
High: up to 15 000 points
Medium: up to 10 000 points
Low: up to 2 000 points
Note: up to 500 points

1 point currently corresponds to 1 USD (payable in ETH or BTC), something which may change without prior notice.

OBS! Between 2017-09-19 and Byzantium hard-fork on Mainnet, each point corresponds to 2 USD for issues related to cross-client consensus or geth DoS vulnerabilities.

Beyond monetary rewards, every bounty is also eligible for listing on our leaderboard with points accumulating over the course of the program.

In addition to Severity, other variables are also considered when the Ethereum Foundation bug bounty panel decides the score, including (but not limited to):

Quality of description. Higher rewards are paid for clear, well-written submissions.
Quality of reproducibility. Please include test code, scripts and detailed instructions. The easier it is for us to reproduce and verify the vulnerability, the higher the reward. Please see the wiki and repos to learn more about our test suite in the official documentation.
Quality of fix, if included. Higher rewards are paid for submissions with clear description of how to fix the issue.

Important Legal Information

The bug bounty program is an experimental and discretionary rewards program for our active Ethereum community to encourage and reward those who are helping to improve the platform. It is not a competition. You should know that we can cancel the program at any time, and awards are at the sole discretion of Ethereum Foundation bug bounty panel. In addition, we are not able to issue awards to individuals who are on sanctions lists or who are in countries on sanctions lists (e.g. North Korea, Iran, etc). You are responsible for all taxes. All awards are subject to applicable law. Finally, your testing must not violate any law or compromise any data that is not yours.

BOUNTY SCOPE

Our bug bounty program spans end-to-end: from soundness of protocols (such as the blockchain consensus model, the wire and p2p protocols, proof of work, etc) and protocol/implementation compliance to network security and consensus integrity. Classical client security as well as security of cryptographic primitives are also part of the program. When in doubt, send an email to bounty@ethereum.org and ask us.

Here is some guidance on what we are typically interested in hearing about:

Geth security

Geth is an Ethereum client written in Go. Areas that typically are in scope are:

Consensus protocol compliance: are there any flaws that would make Geth deviate from consensus?
- This are includes EVM operations, precompiles, block validation etc.
Account management flaws: flaws which would put end-user accounts at risk.
DoS issues: flaws that makes Geth crash or perform very slowly.
- Includes suboptimal EVM operation, failure to handle attacks on the protocol stack

Some areas of Geth are ‘experimental’, and not yet enabled by default. Yes, these are also included, but the ‘Impact’ of issues in the areas below will be counted as low.

LES

The LES (light clients) parts of Geth are twofold: server and client. For LES, we are interested in

LES Server or Client: RCE-type of vulnerabilities
LES Server: DoS vulnerabilities

Swarm

Swarm is not yet production ready, and has very limited bounty scope. We are always interested in RCE-types of vulnerabilities, but not (yet) DoS via swarm protocols.

Whisper

Whisper is also not yet production ready, and has very limited bounty scope.

RCE-types of vulnerabilities.
Cryptographic flaws which would break the Whisper protocol confidentiality.
- For example, if an multi-node adversary can recover plaintext, or identify a particular sender of encrypted messages
Not included: DoS via whisper protocol

EthereumJ/Harmony

EthereumJ EthereumJ is a pure-Java implementation of the Ethereum protocol, and the basis of Harmony – a full Ethereum client. EthereumJ/Harmony is not included in the bounty program, since there are still too many known issues, and not a full mainnet client yet.

Aleth

Aleth is an impementation of an Ethereum node in C++. This client is included, but any issues found will have rather low Impact rating since it’s not commonly used. Typically, we would be interested in consensus or p2p DoS issues, but not so much e.g. DoS via RPC attacks.

Py-evm/Trinity

Py-evm is a python implementation of the Ethereum Virtual Machine, and the basis for Trinity. The Trinity client is currently in an alpha release stage and is not suitable for mission critical production use cases. Both of these components are included in the bounty scope, but any issues reported will have a lowered Impact since there are already known issues and they are not considered production release.

Solidity language security

This category includes:

Incorrect behaviour of the Solidity code generator or optimizer, which could cause unintended functionality (bugs) in the generated contract code.

Here is an example of a submitted Solidity bug.

Solidity does not hold security guarantees regarding compilation of untrusted input – and we do not issue rewards for crashes of the solc compiler on maliciously generated data.

Mist Browser Beta / Ethereum Wallet app

Mist is a Dapp browser that connects users to the blockchain. The scope of bounty submissions includes, but are not limited to:

Privilege escalation issues.
Flaws breaking into the victim’s filesystem.
Flaws compromising any information outside each website scope (e.g. localStorage leaks, cross-website interaction).

Flaws affecting Mist that were already made publicly available by the vendors (e.g. Electron, Chromium, v8) may or may not be accepted by us.

Vyper language security

The Vyper language is a new, experimental programming language for the EVM. It is still beta software, and as such is not expected to be bug-free.

Vyper is included in the bug bounty, but due to it still being under development, the Impact of bugs found will be downgraded accordingly.

Typical bugs that could qualify are:

Flaws that produce bytecode which is clearly erroneous, as in ‘not what the programmer intended’
Semantic flaws, where a programmer might very easily make an error, e.g. due to misleading syntax or keywords

As with Solidity, crashes of the Vyper compiler in the face of malicious input is not included in the bounty program.

LLL language security

LLL is not included in the bug bounty.

Pyethereum/Pyethapp

Pyethereum is a legacy Ethereum implementation, and the basis for the Pyethapp python client implementation. Both of these are now deprecated, in favour of py-evm/Trinity, and not not in scope of the bounty program.

ENS security

This category includes:

Flaws making it possible to gain unauthorized access to, or prevent the authorized withdrawal of, funds locked in Deeds.
Flaws making it possible to interfere with, or make modifications to, an ENS-domain belonging to another user.
Flaws in the auction that affect the legitimacy of auction results.

Here is an example of a bug in the initial ENS registrar that would have allowed people to bid during the reveal period, thus affecting the legitimacy of auction results.

Other clients/things

Clients not developed by the Ethereum Foundation would typically not be covered by the bounty program. For Parity, please visit their bounty program.

ERC20 contract bugs are typically not included in the bounty scope. However, we can help reach out to affected parties, such as authors or exchanges in such cases.

Ethereum.org infrastructure

Our infrastructure; such as webpages, dns, email etc, are not part of the bounty-scope.

FAQ

So, what should a good vulnerability submission look like?

Here is an example of a real issue which was previously present in the Go client:

Description: Remote Denial-of-service using non-validated blocks

Attack scenario: An attacker can send blocks that may require a high amount of computation (the maximum gasLimit) but has no proof-of-work. If the attacker sends blocks continuously, the attacker may force the victim node to 100% CPU utilization.

Impact: An attacker can abuse CPU utilization on remote nodes, possibly causing full DoS.

Components: Go client version v0.6.8

Reproduction: Send a block to a Go node that contains many txs but no valid PoW.

Details: Blocks are validated in the method Process(Block, dontReact). This method performs expensive CPU-intensive tasks, such as executing transactions (sm.ApplyDiff) and afterward it verifies the proof-of-work (sm.ValidateBlock()). This allows an attacker to send blocks that may require a high amount of computation (the maximum gasLimit) but has no proof-of-work. If the attacker sends blocks continuously, the attacker may force the victim node to 100% CPU utilization.

Fix: Invert the order of the checks.

So, the bug bounty program is time limited?

No end date is currently set. See the “News & Updates” section above, and the Ethereum blog for the latest news.

How are bounties paid out?

Rewards are paid out in ETH or BTC after the submission has been validated, usually a few days later. Local laws require us to ask for proof of your identity. In addition, we will need your ETH/BTC address.

Can I donate my reward to charity?

Yes. We can donate your reward to an established charitable organization of your choice.

I reported an issue / vulnerability but have not received a response!

We aim to respond to submissions as fast as possible. Feel free to email us if you have not received a response within a day or two.

I want to be anonymous / I do not want my name or nick on the leader board.

Submitting anonymously or with a pseudonym is OK, but will make you ineligible for BTC rewards. To be eligible for BTC rewards, we require your real name and a proof of your identity. Donating your bounty to a charity doesn’t require your identity.

Please let us know if you do not want your name/nick displayed on the leader board.

What are the points in the leaderboard?

Every found vulnerability / issue is assigned a score. Bounty hunters are ranked on our leaderboard by total points.

I have further questions.

Email us at bounty@ethereum.org.

Do you have a PGP key?

Please use AE96 ED96 9E47 9B00 84F3 E17F E88D 3334 FA5F 6A0A

1.		Martin Holst Swende* holiman 33500 pts
2.		Yoonho Kim (team Hithereum) uknowy 20000 pts
3.		PeckShield peckshield 18200 pts
4.		Bertrand Masius catageek 15000 pts
5.		Tin tintinweb 12000 pts
6.		Juno Im junorouse 10500 pts
7.		Heilman/Marcus/Goldberg 10000 pts
8.		Jonas Nick jonasnick 10000 pts
9.		Dominic Brütsch 7500 pts
10.		Harry Roberts HarryR 5000 pts
11.		Peter Stöckli p- 5000 pts
12.		EthHead EthHead 5000 pts
13.		Yaron Velner yaronvel 2000 pts
14.		Whit Jackson whitj00 2000 pts
15.		Christoph Jentzsch CJentzsch 2000 pts
16.		Vasily Vasiliev 1000 pts
17.		talko talko 1000 pts
18.		Steve Waldman swaldman 1000 pts
19.		Panu Kekäläinen ptk 1000 pts
20.		henrit henrit 1000 pts
21.		Marc Bartlett BlameByte 1000 pts
22.		Barry Whitehat 1000 pts
23.		Lucas Ryan badmofo 1000 pts
24.		Marcin Noga (Cisco/Talos Security) 500 pts
25.		jazzybedi 500 pts
26.		Jonathan Brown ethernomad 500 pts

ETHEREUM Bounty Program

Leaderboard