alanfuNZ

4 ms·

1. 1.▲
   by alanfuNZ 29d ago

   Agree on the sandbox side. dagger or a microVM is the right blast-radius layer, and moving creds out of the environment entirely (service accounts, WIF, an egress proxy that injects them) is the real fix for the secret-then-outbound-call ca
2. 2.▲
   Ask HN: How do you gate an autonomous coding agent's shell access?

   2 points by alanfuNZ 29d ago|3 comments
3. 3.▲
   by alanfuNZ 1mo ago

   This only works in the case of working with databases, also a lot of the times people are not careful with the permissions that are given to these agents and they oftentimes touch areas that they were not originally meant to touch. For exam
4. 4.▲
   by alanfuNZ 2mo ago

   Also the intro of the readme was written by me, the rest was meant to just be something agents ingest and use to set up Doberman hahaha, didn't think anyone would read the full thing.
5. 5.▲
   by alanfuNZ 2mo ago

   Most of the other frameworks sit on the network peremiter and only inspects the prompts. The biggest competition right now that I see in the same lane Cisco's Defenseclaw which focuses on breadth and has relatively loose security guard
6. 6.▲
   Show HN: Doberman: The AI watchdog that stops Claude from deleting your database(github.com)

   9 points by alanfuNZ 2mo ago|6 comments