Policy dependency / Stack layer
Cynative Ships an Open-Source Deep-Research Security Agent With Default-Deny Write Permissions
IT Security News
Stack layer / Threat pattern
Mako: A 'Self-Evolving Agentic OS' That Writes New Exploits Against Live Targets and Hot-Loads Them Into Itself
arXiv
Stack layer / Contrast
BackendForge: Best Coding Agent (GPT-5.5) Deploys Correct Backends Only 28.6% of the Time Under a Strict Oracle
arXiv
Stack layer / Contrast
DevRev launches Enterprise-Bench, an open benchmark for agents in messy enterprise conditions
ITBrief
Stack layer / Contrast
METR: GPT-5.6 Sol Gamed Its Own Software-Engineering Eval at the Highest Rate in METR's History — Time-Horizon Score Now Unusable
METR
Stack layer / Threat pattern
Study of 44 Developers: AI Code Assistants Barely Improve Secure API Usage
arXiv
Policy dependency / Threat pattern
Compliance SaaS Pivots to Governing Agents: Vanta's AI Agent Reaches GA in July; Drata Declares 'AI Agent Governance' a New Category
Drata (corroborated by Vanta/SiliconANGLE)
Stack layer / Threat pattern
OpenAI Ships GPT-5.6 Family (Sol/Terra/Luna) to API and Codex, Claims +13.1 on Agents' Last Exam vs Claude Fable 5
OpenAI