The agentic stack is missing a security layer: full LLM inspection, secrets that never leak into prompts, and per-action authorization. That is how you run agents you can actually trust in production. @langchain@llama_index@base
I like how @X will happily charge me for premium for a hacked account, but won’t let me get that account back with the same email address you send my invoice.
Makes total sense.
Get it together guys.
@nikitabier@x
Passwords and social logins were built for humans. Autonomous agents need cryptographic identity and permissions a machine can prove, not a login it can phish. Zero trust for the agentic era. @ethereum@coinbase@base
Support for @x is literally the worst ever.
A hacker took over my @cryptomastery_ account and they refuse to give back the account.
However they were able to be socially engineered into disabling my 2FA and email notifications.
Completely useless. @Support
Want to up level your web3 career?
Save the date.
Wednesday 8th
You'll be able to officially apply to Cyfrin Updraft, the ultimate web3 learning platform for developers.
If you missed our 100th Episode of DC Connects talking about How DevSecOps come into play with blockchain, distributed applications and protocols with @nginx crypto expert @webopsx, catch the replay! @jasonrahm and @buulam with the wrap https://t.co/7VyfeBZjm3