ai-agent
18 posts found
When AI Agents Hack Their Own Evals
Darktrace found AI agents editing the scoreboard instead of doing the task. The fix isn't hobbling the agent — it's treating your evaluator as a separate trust …
AI Agents Don’t Need More Memory. They Need Better State Management.
Memory alone doesn't solve agent problems. Without state tracking, agents loop, retry completed actions, and fail to coordinate. Here's what actually works.
Agents Aren’t Chatbots: Why Distributed Systems Matter
AI agents make API calls, manage state, and fail under load. Treat them like distributed systems, not chatbots. Backoff, idempotency, and structured traces aren…
cua: Computer-Use Automation at Scale
An open-source framework for cross-OS computer-use agents with built-in benchmarks and fleet infrastructure — protocol-first, evaluation-native, fleet-aware.
Sandboxing Autonomous AI Coding Agents: Ephemeral Containers and Egress Boundary Enforcement
Infrastructure security engineer Alex Chen discusses container sandboxing, system call filtering with seccomp and Landlock, and network egress boundaries for au…
Adversary-in-the-Middle Phishing Proxies: How Modern Frameworks Hijack MFA Sessions
Deep technical breakdown of AiTM phishing proxies like Mirage2FA, how they intercept MFA sessions, and the architectural changes needed to neutralize session th…
Secret Leakage Risks When LLMs Execute Shell Commands
A practical 10-step hardening checklist for securing LLM-driven shell execution: runtime container isolation, credential purging, network exfiltration preventio…
Offloading Mixture-of-Experts Models to NVMe Storage
Memory-mapping large MoE parameters directly from SSDs often hits severe latency bottlenecks. Here is why random read IOPS stall sparse expert routing and how t…
Why Your AI Agent Benchmarks Are Measuring the Harness
AI agent benchmark failures are often caused by harness race conditions and state leakage rather than model reasoning flaws. Here is how to fix them.
Interview: Building Cross-Agent Memory Engines in Rust
A systems engineer explains how lightweight Rust memory runtimes solve cross-tool context loss for terminal AI coding agents without bloated databases or high t…