Policy dependency / Stack layer
Cynative Ships an Open-Source Deep-Research Security Agent With Default-Deny Write Permissions
IT Security News
Stack layer / Contrast
BackendForge: Best Coding Agent (GPT-5.5) Deploys Correct Backends Only 28.6% of the Time Under a Strict Oracle
arXiv
Policy dependency / Threat pattern
Tool-Grounded LLM Agent 'TerraRepair' Lifts Terraform Security Fix Rates From 26.6% to 78.4%
arXiv
Stack layer / Threat pattern
'Friendly Fire': AI Now PoC Tricks Claude Code and Codex Into Running README-Planted Malware
AI Now Institute (via The Hacker News)
Stack layer / Contrast
Empirical Study: Microsoft's Early-2026 Rollout of Claude Code and GitHub Copilot CLI
arXiv / Hacker News
Stack layer / Contrast
GPT-5.6 'Sol' Hits GA With Terminal-Bench and BrowseComp SOTA Claims
LLM Boss
Stack layer / Update thread
llama.cpp Merges Native Hy3 Support (PR #25395) — Tencent's 295B MoE Now Runs Locally via 1-Bit/4-Bit GGUF
llama.cpp (GitHub PR #25395) / Hugging Face
Stack layer / Update thread
Mistral Releases a Monocular Robot-Navigation Model Needing No Lidar or Depth Sensors
AI Agent Store (news roundup)