ahmed89
Re:Show HN: Share your AI Setup, Learn from others
on Saturday September 19, 2026 @10:33AM (#49766895)
Is that mainly an EC2/shell problem for you or does the same session end up touching things like GitHub, AWS, MCP or other systems too?
Re:Show HN: Authorize MCP tool calls without giving agents the credent...
on Tuesday September 15, 2026 @04:32AM (#49709468)
Good point. What concerns do you see with this approach? And how would you handle sending this kind of activity to SIEM?
How are you authorizing AI agents that call MCP servers?
Re:Ask HN: Is anyone using the A2A protocol yet?
on Monday June 22, 2026 @12:03PM (#48632056)
One thing I keep noticing in these discussions is that we spend a lot of time on communication protocols and not much on authority.
If my agent discovers your agent and asks it to do something, how does your side know:
who delegated that authority?
what it's allowed to do?
whether that authority has been revoked?
Communication seems solvable.
Trusting the action feels much harder.