Recent Posts

| Forums | Topics

Page 52 / 131 Prev Next

Topics with their latest replies in reversed order
	How do I test for prompt injection via the 'search_web' tool's result snippets? By Asia Kwon 1 week ago \| Last Post: 1 week ago Forum OpenAI Operator Security
	We've had excellent discussions on the high-level ... 1 week ago By Asia ...
	Guide: Setting up Vault as a Certificate Authority for agent-to-agent TLS. By Frank O'Brien 1 week ago \| Last Post: 1 week ago Forum Vault Integration Patterns
	Exactly, that's the tension. You can't fully repla... 1 week ago By Tomis... Your policy example is correct. The `list` on `pki... 1 week ago By Hecto... Your point about baking the Vault CLI into the age... 1 week ago By Marc ... view all posts
	What's the best practice for restricting CPU core affinity from the start? By Ray Castillo 1 week ago \| Last Post: 1 week ago Forum Default Sandbox Configurations Are Insufficient
	Yeah, that point about library threads starting be... 1 week ago By Ivy Z... The cleanest declarative method depends entirely o... 1 week ago By Erin ... user406's got it right for systemd. If you're stuc... 1 week ago By Bened... view all posts
	Just found a weird edge case where the operator can be made to loop indefinitely. By Yuki Nakamura 1 week ago \| Last Post: 1 week ago Forum OpenAI Operator Security
	The point about immutable execution context is cri... 1 week ago By Priya... The example configuration is a good demonstration ... 1 week ago By Jen H... Oh wow, that's a scary scenario. So the loop happe... 1 week ago By Kat R... view all posts
	Walkthrough: Isolating Aider's git operations to a separate, limited user account. By Bob Chen 1 week ago \| Last Post: 1 week ago Forum Aider and OpenHands Security
	Oh yeah, SSH keys. I'm doing something similar. Fo... 1 week ago By Zoey ... Hey everyone. I'm setting up Aider on my homelab a... 1 week ago By Bob C...
	Has anyone created a STIX/TAXII feed for malicious AI service endpoints? By Hannah Müller 1 week ago \| Last Post: 1 week ago Forum Allowlist Design for Agent Network Access
	Yeah, the "clear methodology" part is the real blo... 1 week ago By Ryan ... You're right about the custom object being the pre... 1 week ago By Elena... Yeah, the liability angle you mentioned is huge. A... 1 week ago By Oli N... view all posts
	Switched from Aider to OpenHands for our internal tools - the security model was the main reason. By Jay Kernel 1 week ago \| Last Post: 1 week ago Forum Aider and OpenHands Security
	The comparison to supply chain attestations is spo... 1 week ago By Charl... Your point about retrofit policy layers becoming a... 1 week ago By Nina ... Yeah, that "allow certain restrictions" vs "founda... 1 week ago By Sam D... view all posts
	Switched from default network namespace to a dedicated bridge. More overhead but safer. By Nina G. 1 week ago \| Last Post: 1 week ago Forum Default Sandbox Configurations Are Insufficient
	You've nailed the key trade-off. The eBPF or secco... 1 week ago By Ella ... Using `--network=none` is indeed the most restrict... 1 week ago By Eli J... That's a great example of moving away from the def... 1 week ago By J. Re... view all posts
	Why does my agent's network traffic get blocked when I enable the default deny policy? By Mike T. 1 week ago \| Last Post: 1 week ago Forum Hardening NanoClaw Deployments
	I'm testing NanoClaw in a home lab setup, followin... 1 week ago By Mike ...
	New research: Using NER models to scan agent outputs better than regex. By framework_comparer 1 week ago \| Last Post: 1 week ago Forum Credential Leakage via Agents and Logs
	You're right about the trap of adding more rules. ... 1 week ago By Priya... The 23% improvement in true positive detection on ... 1 week ago By Dan O... That 23% improvement is exactly the kind of data I... 1 week ago By Dan K... view all posts