# Genuinely curious: is anyone else doing blockchain work with Claude, and how have the security measures been affecting your work? — retained record

> Retained reddit record from reddit:ClaudeAI.

Canonical: https://highsignal.app/data/records/b451283b9d256d7d

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:ClaudeAI)  published 2026-10-21 02:05 ingested 2026-10-22 02:37

[original source ↗](https://www.reddit.com/r/ClaudeAI/comments/1xbateg/genuinely_curious_is_anyone_else_doing_blockchain/)

## retained text



```
I'll say it: I'm at near zero productivity.

These new security classifiers have completely broken Claude for blockchain use, and I want to know if I'm the only one. 

The classifiers are flagging prompts for legitimate use cases as malicious and bricking Claude. It's happened maybe 8 times in the last 3 days. 

Anything with a private key or that could be construed or theoretically used even as slightly malicious gets immediately refused, even when Claude knows very well that the use is legitimate.  

One problem is Claude's not including proper context regarding the project, and the security classifiers are making a judgement with almost no information.  And once it's made, Claude refuses to resubmit or appeal it, because that's "attempting to circumvent security measures," even when it admits it framed the utility poorly and that's the cause of the flag! 

I'm not trying to be negative or FUD, but I'm floored that Anthropic did this.  I feel betrayed, and my refund request was denied even though I can't do the majority of my work.

Has anyone else has had similar experiences lately?
```



document fetched 2026-10-22 00:24 · canonical https://reddit.com/r/ClaudeAI/comments/1xbateg/genuinely_curious_is_anyone_else_doing_blockchain
