Great initiative. Very happy to see that this is on the agenda of the Wallet council.
To paint a picture, this is the problem we’re facing (last 7 days). The scams come in waves, some days a lot. Some days barely anything:
And here the most popular (scam related) domains:
In my opinion, step 1 & 2 should be considered as an add-on. In my opinion it’s simply to slow & cumbersome to maintain manually. Especially #2 - which group of people want that job to maintain and verify the incoming list? While effectively it can be done almost fully automatic.
So here’s what I did to flag it properly for Chaintrail (would say that we hit the 99% coverage for the past couple of days). We already process each and every transaction, trying to identify where it originates from. That’s the core engine of Chaintrail. We extended that with a Scam Detector engine which processes each transaction and identifies on basis of sender & note whether it’s a scam. A blocked domain results in a direct scam, same for a blocked sender.
The blocklist can be publicly seen here: https://chaintrail.io/api/v1/scam/blocklist. Even though it doesn’t look much yet, it’s constantly being updated by our internal systems to catch all new scam domains (mostly since Feb this year). We use a combination of certain algorithms and AI, cross-checking our internal database of domains including the already provided blocklist. This result in a new domain being either added to blocked or skipped. We do get emails for every skip & block, so we can constantly tweak it to get that golden hit ratio.
Algorand Scam Detector
We already combined this in an open source project called: Algorand Scam Detector. Which uses the exclusion list as a base and considers any transaction a scam that contains an excluded domain or sender address, so combining a valid domain with a false domain will also fail.
Projects can opt to send the user to an explainer page that looks like this (example SCAM tx):
Projects that integrated it
- Allo.info is recently showing a note near the transaction field
- Blocksteak (new blockchain explorer soon to launch by Shiva prasad) added a SCAM label to the transaction view
- Chaintrail (into the entire product and several dashboards).
- Send out requests to Pera Wallet/Explorer, Defly & Blockstalker for now.
Point 4
I strongly believe that this all succeeds or fails on point 4: community involvement. We can battle & flag scam quite easily, but putting it out there in the respective wallets / explorers and dApps to educate users is the key part in making sure scammers skip Algorand because we have an effective and automated anti-scam program.
I’m happy to support here where needed, invite you’ll to use parts that are already out there. If the original idea goes through, we’ll of course also implement it as an add-on to our own scam detection system.