[Proposal] Gitcoin Anti-Fraud Stewards Draft Grant Proposal

Hi folks!  The gitcoin Anti-Fraud stewards are also building a proposal to be submitted before the close of GR10.  The proposal is still in rough form but here is a link for anyone interested to take a look.  Feedback is welcomed.  The proposal is addressing rewards for:



Grant Verification & Oversight

Human Participation in Machine Learning Algorithms

Policy Documentation for Round 10 anti-fraud organisation.  


The group is having weekly meetings on Tuesdays at 2200h CET.  Drop by the meetings if interested.

Hello @blazingthirdeye! Thank you for the proposal.

I have some questions about the doc.

Shaping the future of anti sybil: Work in gamifying the approvals process, working with other DAOs to provide heads up about users found to colluding

How do you identify a user in this case? That’s the entire problem of sybil. It’s just an ethereum address. They just abandon it and make another one. So what user data would you forward to other DAOs?

Operational functions of fraud detection

That section needs better formatting and re-wording a bit. I found it quite hard to read and understand. At least split the subsections into their own paragraphs.

All in all if this is for a grant in the gitcoin builds gitcoin category I think this is a really good idea. Would be interested to learn more about the processes you plan to follow but that is offtopic to the proposal here.

1 Like

For example, during a recent hackathon, I think one of the sponsors committed fraud by not awarding the bounty, relying on friendly language and vague requirements. 20 submissions for the prize were all rejected with no clear explanation.


Yes these questions are great. Thanks for the feedback,

There are many criteria including names, websites, wallets and social media accounts. Usually the entire application is scrutinized and links get followed. We try to get a sense of the applicant. When a team is made of OG, OG1, OG2, and there is no additional information the grant is typically flagged or denied.

1 Like