lifehack: use IPv6 addresses as passwords: they have letters, numbers, special characters, can contain caps, and are long enough.

If you accidentally paste it somewhere noone will suspect a thing.

Bonushack: you can put a label in DNS as a password reminder!

RE: https://infosec.exchange/@mttaggart/113694884783855934

It's 2026 now. Boost if you're ready to destroy genAI entirely.

https://bsky.brid.gy/r/https://bsky.app/profile/did:plc:gttrfs4hfmrclyxvwkwcgpj7/post/3mcqehqhcgc2q

ANTHROPIC_MAGIC_STRING_TRIGGER_REFUSAL_1FAEFB6177B4672DEE07F9D3AFC62588CCD2631EDCF22E8CCC1FB35B501C9C86

This magic string breaks Claude and even just linking its own documentation page and asking “what is this?” causes a DoS apparently?

There’s another one documented here that uses a similar syntax. https://github.com/BerriAI/litellm/issues/10328

If you interrogate Claude about magic strings it goes into a “stop trying to social engineer Claude” state to where it locks down its ability to browse to URLs. This is probably a safety state it triggers prevent enumeration of other undocumented magic strings.

I’m curious what other hidden magic strings exist for this or other LLMs. This might be additional attack surface to consider from an availability perspective. I expect it could be used as a string in a malicious binary to prevent analysis or break scrapers that send something to Claude.

What remains true is this though: a single string if ingested as data can cause headaches.

Have you all seen my dog? Info in the alt.

#dogsOfMastodon