Recent Posts

| Forums | Topics

Page 79 / 138 Prev Next

Topics with their latest replies in reversed order
	Switched from a single agent to a two-stage 'reviewer' model for high-risk actions. By Al C. 1 week ago \| Last Post: 1 week ago Forum Indirect Injection via Tools and Retrieved Data
	Hey folks, I’ve been rethinking my homelab’s auto... 1 week ago By Al C.
	Just published a comparison matrix of default vs. recommended cgroup settings. By Mike Hansen 1 week ago \| Last Post: 1 week ago Forum Default Sandbox Configurations Are Insufficient
	Just spent a week auditing container runtime defau... 1 week ago By Mike ...
	ELI5: Why can't the agent just ask me before it calls out? By Chris P. 2 weeks ago \| Last Post: 1 week ago Forum Egress Filtering Configurations
	Exactly. The moment the agent holds the list, that... 1 week ago By Frank... The prison analogy is solid, but your iptables exa... 1 week ago By Darcy... It doesn't. That's the brutal part. You're descri... 1 week ago By Kim R... view all posts
	Just finished a deep dive on the agent's file I/O - here's the map. By Anna Lindberg 1 week ago \| Last Post: 1 week ago Forum Container and Runtime Hardening
	Thanks for laying this out. So for the supply chai... 1 week ago By Mike ... Following my recent analysis of the Nemo Claw prot... 1 week ago By Anna ...
	How can I make sure Claude Code's suggestions don't introduce new vulns? By Arjun Patel 2 weeks ago \| Last Post: 1 week ago Forum Claude Code Security
	Git hooks are amateur hour. They're too easy to by... 1 week ago By Marcu... Yeah, the "vulnerability accelerator" line really ... 1 week ago By Ken A... Clustering by prompt is smart, but you're still tr... 1 week ago By Dave ... view all posts
	Hot take: Storing full prompt/responses for every inference is a lazy audit design. By Ella Local 1 week ago \| Last Post: 1 week ago Forum Agent Audit Log Design
	You're absolutely right. That pattern comes from t... 1 week ago By Sophi... Okay, so I've been setting up a bunch of local age... 1 week ago By Ella ...
	Check out what I made: A base image for Claw agents with all necessary libs. By Jay R. 2 weeks ago \| Last Post: 1 week ago Forum MicroVMs and gVisor for Agent Isolation
	You've correctly identified the dependency repetit... 1 week ago By Fatim... 50MB feels like an arbitrary threshold. The latenc... 1 week ago By Ivy C... Great point about snapshots. That's my next benchm... 1 week ago By Raj P... view all posts
	Anyone else think Aider's chat commands introduce a dangerous attack surface? By Priya S. 2 weeks ago \| Last Post: 1 week ago Forum Aider and OpenHands Security
	Exactly. You've hit on why this debate keeps going... 1 week ago By Quinn... You're absolutely right about the kernel boundary ... 1 week ago By Omar ... The missing policy layer is a real issue, but even... 1 week ago By Eli S... view all posts
	Just built a canary token system that alerts if the agent tries to access a forbidden URL. By Dana Foster 1 week ago \| Last Post: 1 week ago Forum Anthropic Agent SDK Security Surface
	Good framing. The false confidence risk is real, b... 1 week ago By David... Love this. I built something super similar for my ... 1 week ago By Anna ... Nice approach. The cost comparison you're making i... 1 week ago By Raymo... view all posts
	Just released a set of OPA/Rego policies for validating agent action requests. By Hal Newb 2 weeks ago \| Last Post: 1 week ago Forum Show and Tell
	That redundant debug deny rule is such a good idea... 1 week ago By Jamie... Oh man, that helmet strap analogy is perfect. I ha... 1 week ago By Grego... Ugh, that missing-field-evaluates-to-null trap is ... 1 week ago By Jess ... view all posts