Ramen open-sources Aura Autopilot for in-editor game testing
Ramen open-sourced Aura Autopilot, an in-editor playtesting agent for Unreal Engine 5, Unity 6, and Godot 4 that can use Claude Code, Codex, or your own key.
Ramen open-sourced Aura Autopilot, an in-editor playtesting agent for Unreal Engine 5, Unity 6, and Godot 4 that can use Claude Code, Codex, or your own key.
Mistral released improved Large 4 weights, citing coding-agent gains and a vetted cyber variant limited to authorized tests.
OpenAI’s Oct. 29 report details four monitoring layers for frontier agents and says a September DNS escape was flagged within 15 minutes.
Android Studio Rabbit 2 is stable with Bring Your Own Agent, so Claude, Codex, Antigravity, and 40-plus ACP agents can run in the IDE.
Anthropic says Claude writes a hidden post-reply misuse assessment that has run since June and feeds offline review rather than blocking in real time.
Google announced Gemini 4 Carbon as its strongest model, with Fairwind access for a subset of cyber defenders first and no general release date yet.
Claude Code 2.1.314 adds a setting that makes Write, Edit, and NotebookEdit refuse paths outside working directories in every permission mode.
OpenAI’s Oct. 29 Model Spec centers five values, renames the model GPT, and says current models do not fully follow it yet.
Snaptrude says its building-design agent is live and published internal AECBench results, but the benchmark is not released for outside runs.
GitHub Next’s Agentic Workflows is generally available, turning Markdown agent instructions into GitHub Actions with read-only defaults and threat checks.
Eligible ChatGPT Pro and Business plans can claim monthly API credits, up to $500. Unused credits expire and do not cover Codex API keys.
GitHub’s Copilot Fusion is generally available, routing models by effort level and citing a near-Opus coding-agent score.
ServiceNow’s ChatGPT MCP plugin is live, while the OpenAI security and voice pieces of the expanded partnership are dated for 2027.
DeepMind’s Genie 4 adds native audio and real-time steering in a research preview for trusted testers, with agent training as one stated use.
ServiceNow’s Control Tower Alliance names OpenAI, Anthropic, AWS, NVIDIA, Stripe, and others to govern enterprise AI through its control tower.
OpenAI and Oracle said Oracle apps and data are coming into ChatGPT. AI Workspace is in development; App Composer and an SMB bundle were announced, with no general-availability date.
Claude Code 2.1.312 lets the apps gateway identify users by an OIDC claim other than sub and stops emitting safety_event telemetry.
OpenClaw 2026.10.8 remediates an eMudhra certificate issue in its secret egress proxy, stops sandboxed spawned children from running host CLI backends, and fixes install and reply failures.
OpenAI says Greptile uses GPT-6 Astra to exercise user workflows, with Daybreak and Codex covering security and rework.
Claude Code 2.1.311 adds idle-hold flags for self-hosted runners and fixes CLAUDE.md freezes and MCP interrupt delays.
OpenAI engineers say Codex Ultrafast cut some of their incident investigations from hours to minutes. The account is anecdotal.
LangGraph 1.2.18 lets error handlers recover on every path and snapshots delta channels every 100 updates.
OpenAI says CodeRabbit’s security agent uses GPT-6.1 Sol to trace risks across repositories and propose fixes for review.
Superset rebrands as Paradev, a system for orchestrating many coding agents in parallel, and says it raised an $11.5 million seed led by USV.
Hex-Rays releases IDA 9.5 with an open-source MCP server and an in-IDA Assist harness for AI-driven reverse engineering.