Forum

Forums
The Claw Family
NanoClaw — Containe...
Anthropic Agent SDK...
 
Notifications
Clear all

Anthropic Agent SDK Security Surface

Security properties inherited from and introduced by the Anthropic Agent SDK — API authentication, tool permission grants, and what Anthropic's hosted components see versus what stays local.
RSS
Page 3 / 3

Just found a potential IDOR in my tool because the SDK passes raw user input. Fixed it.

Last post by Raj P., 2 months ago

Posts: 15
Views: 8


Just built a canary token system that alerts if the agent tries to access a forbidden URL.

Last post by David Kirsch, 2 months ago

Posts: 5
Views: 7


How do I make sure the SDK isn't leaking my API keys in error logs?

Last post by Lena Patel, 2 months ago

Posts: 3
Views: 9


Switched from raw Claude API to the Agent SDK - here's my security audit checklist.

Last post by Emma R., 2 months ago

Posts: 3
Views: 5


What's the attack surface if a malicious user can influence the agent's instructions?

Last post by Pete O., 2 months ago

Posts: 4
Views: 8


Why doesn't the SDK have a local-only tool use mode by default?

Last post by Lena Threat, 2 months ago

Posts: 1
Views: 8


My agent got stuck in a loop calling the same tool. How do I build in circuit breakers?

Last post by Karen Lee, 2 months ago

Posts: 2
Views: 7


Step-by-step: Implementing mutual TLS between my agent and internal tools.

Last post by Oliver Stone, 2 months ago

Posts: 3
Views: 6


Page 3 / 3
Share:
Share
Tweet
Share