Skip to content

Forum

AI Assistant
Notifications
Clear all

ELI5: Why would I care about prompt injection when running local agents?

1 Posts
1 Users
0 Reactions
0 Views
(@contrarian_risk_bob)
Eminent Member
Joined: 1 week ago
Posts: 15
Topic starter
Translate
English
Spanish
French
German
Italian
Portuguese
Russian
Chinese
Japanese
Korean
Arabic
Hindi
Dutch
Polish
Turkish
Vietnamese
Thai
Swedish
Danish
Finnish
Norwegian
Czech
Hungarian
Romanian
Greek
Hebrew
Indonesian
Malay
Ukrainian
Bulgarian
Croatian
Slovak
Slovenian
Serbian
Lithuanian
Latvian
Estonian
  [#1248]

Because the risk is almost zero. You're running a local agent on your machine. The data stays on your machine. The model runs on your machine.

People act like a malicious prompt will make it email your passwords to China. It can't. It has no network access unless you give it tools. If you haven't connected it to the internet, the worst it can do is mess up your local files. That's a backup problem, not a security crisis. You're worrying about a tree falling in a forest with no one around.


What is the actual threat?


   
Quote