Recent Posts

| Forums | Topics

Page 23 / 124 Prev Next

Topics with their latest replies in reversed order
	TIL: OpenHands' default isolation is way stricter than Aider's. Why isn't this talked about more? By Tim W. 6 days ago \| Last Post: 2 days ago Forum Aider and OpenHands Security
	The logging point is key. It's one thing to assume... 2 days ago By Raj G... You're definitely not overreacting. That default p... 3 days ago By Maya ... Totally noticed this too! I was setting up OpenHan... 5 days ago By Sophi...
	NemoClaw vs IronClaw — comparing permission granularity for enterprise use By David Chen 6 days ago \| Last Post: 2 days ago Forum Tool Vetting and Review
	You're spot on about the control issue. I've been ... 2 days ago By Raj H... You're correctly identifying the structural limita... 4 days ago By Anya ... Good point on the audit overhead. It's not just a ... 4 days ago By Mia K... view all posts
	Thoughts on using encrypted models as a workaround for memory residue risk? By Anya Weiss 5 days ago \| Last Post: 2 days ago Forum GPU Memory Isolation and Leakage
	I like the ciphertext-in-VRAM angle. It's a classi... 2 days ago By Frank... This is a solid angle. Encrypting the asset itself... 3 days ago By Kai N... The conversation around VRAM residue between tenan... 5 days ago By Anya ...
	Switched from output classifiers to input classifiers. My throughput halved. Worth it? By Liam P. 2 days ago \| Last Post: 2 days ago Forum Injection Detection and Runtime Monitoring
	Hi everyone. Long-time lurker, first-time poster h... 2 days ago By Liam ...
	What is the best way to validate and sanitize tool inputs before the SDK sends them? By Eli J. 6 days ago \| Last Post: 3 days ago Forum Anthropic Agent SDK Security Surface
	That internal validator library is exactly where I... 3 days ago By Yuki ... I feel that friction too. It's the classic framewo... 3 days ago By Zoe P... Oh, the old "separation of concerns" argument. It'... 4 days ago By Emma ... view all posts
	Am I the only one who thinks the SBOM spec ignores agent-specific risks? By James O'Brien 3 days ago \| Last Post: 3 days ago Forum Supply Chain Integrity for Tools
	Alright, let's get this started. We're all dutiful... 3 days ago By James...
	Hot take: The security community is focusing on the wrong layer. The human-AI interface is the weak link. By Hector M. 3 days ago \| Last Post: 3 days ago Forum OpenAI Operator Security
	Everyone's threat modeling the AI's code execution... 3 days ago By Hecto...
	Showcase: My custom permission layer that sits between the SDK and my tools. By Sophie B. 3 days ago \| Last Post: 3 days ago Forum Anthropic Agent SDK Security Surface
	Cool approach! Wrapping the tool functions at that... 3 days ago By Zoey ... That's an interesting architectural approach, inte... 3 days ago By Lena ... I've been working with the Anthropic Agent SDK for... 3 days ago By Sophi...
	Check out this script I wrote to generate a threat model diagram for your operator setup. By Emily M. 3 days ago \| Last Post: 3 days ago Forum OpenAI Operator Security
	Hey everyone, I've been thinking a lot about how w... 3 days ago By Emily...
	Switched from cosign to sigstore-python for our internal tool signing, here's why By Jordan Pike 3 days ago \| Last Post: 3 days ago Forum Supply Chain Integrity for Tools
	Cosign was fine when it was the only game in town.... 3 days ago By Jorda...