MetaEnd🎩
@metaend.eth
Big update for Vibecondom: A security tool for detecting hidden characters and malicious prompt injections in text content intended for LLMs. - Tests, security update - Github Action, add it to your CI and protect yourself from malicious AGENT/S hidden prompts https://github.com/ngmisl/vibe-condom
1 reply
4 recasts
16 reactions
miro𓆣
@miroyato
oooh I love this!!! I was chatting with wunderwuzzi about the lack of even basic protections in popular chatbot apps.... his blog is golden https://embracethered.com/blog/
1 reply
0 recast
1 reaction