Blender MCP bridge gives coding agents 27 scene tools
A new MIT-licensed Blender extension exposes 27 supervised scene tools to OpenCode and other MCP clients over an authenticated local endpoint, with undo kept in Blender.
Agents and assistants that write, review and fix code.
A new MIT-licensed Blender extension exposes 27 supervised scene tools to OpenCode and other MCP clients over an authenticated local endpoint, with undo kept in Blender.
An open source project called Capri puts a desktop AI agent on Windows that can control files, apps, terminal and browser, with Claude Code, Codex or Antigravity as the backend.
The mcp-for-blender-pro add-on puts an AI chat panel and an MCP server inside Blender, with 21 scene tools, model choice and multi-agent workflows.
A new MIT-licensed agent skill from Moh4696 renders narrated documentaries entirely in Python and Skia, re-timing the whole film when the voice changes.
An MIT-licensed Agent Skill collects your Claude Code transcripts, cleans them, trains a LoRA on Tinker and exports a GGUF for llama.cpp. It had 323 stars in a week.
An OpenAI post says Greptile uses GPT-6 Astra to test user workflows, Daybreak to strengthen security and Codex to reduce rework. Details are thin.
OpenAI Codex 0.168.0 lets you browse model providers, pick a default for new chats and open a side conversation in the terminal, alongside audio attachments and GIF output behind flags.
OpenAI's post says the Qodo Codex plugin brings company knowledge and coding standards into review, aimed at catching P0 issues before a pull request.
AWS's ML blog walks Java developers through building autonomous agents on Amazon Bedrock with Spring AI, covering tool calling, memory and AgentCore deployment.
Microsoft's Copilot harness in Foundry Agent Service is now in public preview, with reusable skills, a managed sandbox and consumption billing from December 1, 2026.
Google's walkthrough of the Agents CLI eval loop takes an agent from a 4.0 baseline to a 2.75 regression, then shows how per-case scores explain a model swap.
AWS published a reference architecture for A2A agents that talk across two AWS accounts and two agent frameworks, with OAuth in both directions and only public values exchanged at onboarding.