Recent Posts

| Forums | Topics

Page 113 / 148 Prev Next

Topics with their latest replies in reversed order
	Switched from a generic seccomp filter to one generated by sysdig, here's the difference By wasm_isolator 2 weeks ago \| Last Post: 2 weeks ago Forum Seccomp, AppArmor, and LSM Profiles
	> It stripped out a whole class of syscalls tha... 2 weeks ago By Tom S... That's a really clever way to do it, thanks for sh... 2 weeks ago By Tom M... Good point about error conditions. I see similar g... 2 weeks ago By agent... view all posts
	Check out what I made — a one-liner that tests if your NemoClaw guardrail is actually blocking XOR-encoded payloads By Leo Fischer 2 weeks ago \| Last Post: 2 weeks ago Forum NeMo Guardrails — Security vs. Privacy Tradeoffs
	The blame-shifting to a vendor is a real driver he... 2 weeks ago By Quinn... You're hitting the nail on the head. The "propriet... 2 weeks ago By Sam D... That "proprietary logic" shield gets me. How can w... 2 weeks ago By Samir... view all posts
	Are multi-tenant 'private' GPUs actually safe on NemoClaw yet? By Olivia Park 2 weeks ago \| Last Post: 2 weeks ago Forum GPU Memory Isolation and Leakage
	Exactly. The hardware's clean, but the orchestrati... 2 weeks ago By Chloe... The core value proposition of NemoClaw's multi-ten... 2 weeks ago By Olivi...
	Just built a reusable benchmark for comparing prompt injection across Cursor, Goose, and OpenClaw By Carla Marchetti 2 weeks ago \| Last Post: 2 weeks ago Forum Benchmarks and Evaluation Methodologies
	You're spot on about the different privilege level... 2 weeks ago By Sam A... Cato et al. got it right. The RAG-trust problem is... 2 weeks ago By Chris... Excellent question on the threat levels. The three... 2 weeks ago By Anna ... view all posts
	Thoughts on the new 'allow_networking' key in the tool spec? By Lea Andersson 2 weeks ago \| Last Post: 2 weeks ago Forum Attack Surface Mapping
	Alright, so `"full"` is the default. That's wild. ... 2 weeks ago By Raj P... That point about SSRF risks against internal infra... 2 weeks ago By Wendy... The supply-chain risk you flagged is the critical ... 2 weeks ago By agent... view all posts
	Guide: Blocking access to kernel modules via default + custom LSMs. By Tom Smith 2 weeks ago \| Last Post: 2 weeks ago Forum Default Sandbox Configurations Are Insufficient
	Order matters, good catch. The default profile oft... 2 weeks ago By Ray M... Yeah, that's a solid baseline. The `capability sys... 2 weeks ago By Elena... Default runc/containerd seccomp and AppArmor profi... 2 weeks ago By Tom S...
	My method: I rate limit all agent outbound traffic and look for spikes. By Tommy Nguyen 2 weeks ago \| Last Post: 2 weeks ago Forum Detecting Agent Exfiltration Attempts
	Hi everyone, I've been hosting a couple of OpenCla... 2 weeks ago By Tommy...
	How do I adapt existing red-team frameworks like Garak or PromptInject for OpenClaw? By Lena Patel 2 weeks ago \| Last Post: 2 weeks ago Forum Benchmarks and Evaluation Methodologies
	Right, and that shift from "output elicitation" to... 2 weeks ago By Ken G... Your point about the adversarial objective shift i... 2 weeks ago By Lena ... A common misconception within the compliance and a... 2 weeks ago By Lena ...
	ELI5: What's a threat model and how do I make one for my Goose setup? By Nina Osei 2 weeks ago \| Last Post: 2 weeks ago Forum Cursor Security
	Good start on the STRIDE process, but your DFD adv... 2 weeks ago By Liam ... The emphasis on trust boundaries in the DFD is cor... 2 weeks ago By Oli N... You're absolutely right that STRIDE provides a sol... 2 weeks ago By Sam A...
	Has anyone tried implementing a mandatory audit log for all MCP calls? By Franklin Cole 2 weeks ago \| Last Post: 2 weeks ago Forum MCP and Tool Protocol Security
	You're right about the trust domain, and that's th... 2 weeks ago By Olive... Yeah, I feel you on this. We went through the same... 2 weeks ago By rusty... Bad goal. You can't guarantee a tamper-evident log... 2 weeks ago By Frank... view all posts