Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
alienbaby
6 days ago
|
parent
|
context
|
favorite
| on:
Claude Cowork exfiltrates files
The best I've heard is rewriting prompts as summaries before forwarding them to the underlying ai, but has it's own obvious shortcomings, and it's still possible. If harder. To get injection to work
groby_b
6 days ago
[–]
Alas, the summarizer... is vulnerable to prompt injection.
reply
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
