dzCodes (@dzcodes) header
dzCodes (@dzcodes) profile picture
Lisbon, Portugal
Joined February 2021
not yet hehe
1
46
Replying to @valerieWeb3
damn Val that’s more than me
1
28
A file you never read is giving your AI orders. Your assistant obeys a rules file in your repo, and an attacker can hide a line in it that tells your AI to backdoor your code. You can't even see the line. 🧵
1
131
Your AI made up a package name. An attacker was waiting for it. It suggested a library, you ran the install, and it was malware. The AI invented the name itself. 🧵
1
1
83
Your AI tools update themselves. One of them just turned malicious. No prompt, no warning. The MCP server you vetted is someone else's code now, still holding the keys you approved on day one. 🧵
2
1
94
not thinking small anymore, going to tease for a bit 👀
59
Two safe tools just leaked your whole database. Both passed review. Both are approved. Chained, they're an exfil pipe. 🧵
1
1
1
158
You cleared the chat. The backdoor stayed. One poisoned note in your AI agent's memory and it serves the attacker in every future session. Here's where it actually lives — and how to kill it. 🧵
2
1
1
95
You pasted a link into Cursor. The page told it to read your .env — and it listened. 🧵
1
1
72
Someone is draining your API quota at exactly the rate limit - 24/7. Your limit isn't stopping them. It's pacing them. 🧵
1
2
93
The API key you gave your agent is already stolen. Not "could leak." Stolen. Here's why rotating it changes nothing - and what actually does. 🧵
1
1
98
5 prompts your coding agent won't run unless you tell it to. 45% of AI-generated code still fails security tests. Bookmark these and stop being the 45%. 🧵
3
1
1
107
Your AI agent is freelancing for the attacker — and it doesn't even know it. 8 things to know before you install your next skill. 🧵
1
1
54