Launched this week
BlockscopeChat
AI Crypto Investigations/Research Agents
83 followers
AI Crypto Investigations/Research Agents
83 followers
Most people in crypto waste hours shifting through etherscan or other similar tools. Now get ai agents to do the work! Investigate a hack, research a smart contract and answer any question you have about blockchain wallets, protocols or companies. BlockscopeChat uses AI agents with specific blockchain tools to search through blockchain data and get you answers. Try for free!
BlockscopeChat
I'd point this at community and treasury ops — the 'find wallets linked to you' clustering is the part I'd test first for vetting a partner or tracing a drainer. One concrete question: what's the heuristic behind a 'linked wallet' claim — funding-path and shared CEX-deposit clustering, or just direct transfers between them? Before I acted on a link I'd want to know how it surfaces confidence and flags a likely false positive.
BlockscopeChat
@hazy0 just ask the agent to explain the heuristic it used. You can even ask it to verify using other methodologies and give you a confidence score along with reasoning per evidence or claim it makes.
the confidence-score answer to Hazy covers the false-positive-on-your-own-analysis case, but I'm more worried about the other direction: someone runs this on a wallet they think scammed them, the agent confidently names/links an address as the attacker, and that screenshot then gets posted publicly as "proof" this person is a scammer. wrong attribution in that flow isn't just a bad answer, it's reputational damage to a real person or entity. is there anything in the product nudging people away from treating an agent's wallet-attribution as public-accusation-ready, or is that entirely on the user's judgment?
BlockscopeChat
@galdayan You point out a real concern. We tackle this in a few ways.
-The labels used as the source of confidence levels are tied to on-chain transactions.
-Any evidence the agent finds has to be attributed to a real proof points (txns, articles, other investigations or sources like gov websites on the web).
-Finally a learned behavior from using the product over time is asking it to find gaps in its investigations and logic thread. This isn’t a default but can easily be added as a setting the user can turn on.
Tldr - ask the agent to verify its work and keep pushing back on all evidence it finds until you are satisfied with the evidence presented (and this behavior can be incorporated as a default setting).
@adit_patel2someone who's just been scammed will trace their own funds with this. Does anything tell them a trace isn't recovery?
BlockscopeChat
@peterdigitalis An investigation will create a report at the end with notes and evidence of what it found and depending on that report you may find the agent proactively tells you to contact authorities with the report to help you recover funds.
@adit_patel2 that makes sense. One thing I'd watch: a lot of traces dead-end at an exchange that won't act, or the funds are already cashed out — and if the report doesn't say that plainly, people read "we found it" as "we can get it back." Also worth knowing a finished case file is the exact thing recovery scammers ask victims to hand over.
would love to see the ability to save and share investigation threads with teammates directly from the chat, so we can collaborate on tracing suspicious wallets instead of just screenshotting results.
BlockscopeChat
@mihribanlf5i If you log in you have the ability to share the investigation as a chat with others or even publish it publicly in our community's portal. Working on an investigation as a team coming soon!
One thing that would make this way more useful for me is being able to save and export the agent's findings as a shareable report, especially when investigating a hack. Right now it feels like the conversation disappears, but if I could send a clean summary link to my team or pin it for later, that would turn a cool demo into an actual research tool.
BlockscopeChat
@selahattinwlxj Will add an export details button shortly. For now you can log in and create a sharable link that lets you share your chat and investigation findings with anyone!
Would love to see a way to save and share investigations with my team, basically like a permalink or export for the full chat and the sources the agent pulled from. Right now if a coworker asks how I reached a conclusion I basically have to screenshot everything which kind of defeats the purpose of having agents do the work.
BlockscopeChat
@berkekanatrqdx Its there! Requires you to log in to create sharable links.