The agent that grows with you

Unified World Model Inference & Evaluation Infrastructure

Lightweight coding agent that runs in your terminal

[NPJ Digital Medicine 2026] Data-efficient deep learning for gene rearrangement status prediction using DLBCL pathology

一层极薄的 Codex / Cursor / Claude Code 补充：一份跨项目常驻工作规则、一个从讨论到并行执行的方法 Skill、四类可复用文档

Understand dense articles faster and more clearly: trace the claims, test the evidence, identify genuine novelty, and separate contribution from hype.

MCP-based GPU coordination for AI agents, with human oversight across multiple servers.

[Tech Report] Context Scaling: Scaling Properties of Text Conditioning in Visual Generation

Enjoy the magic of Diffusion models!

Official code release for AnchorSplat: Fast and Structure Consistent Detail Synthesis for Gaussian Splatting (ECCV2026)

Public artifact for TEND, an execution-verified Text-to-NoSQL benchmark with the SAG reference solver and the accepted VLDB 2026 QueryCraft demo.

[CVPR'26] Semi-Supervised Conformal Prediction With Unlabeled Nonconformity Score

WRBench: camera-controlled generation and diagnostic evaluation of video world models.

PhysX-Omni: Unified Simulation-Ready Physical 3D Generation for Rigid, Deformable, and Articulated Objects

[DEPRECATED] 已废弃，请改用 https://github.com/JinPLu/teamwork-simplify —— Astra/Fable 等模型已吸收流程脚手架，Skill 已全面重构。

Programmer's guide about how to cook at home.

你是一个曾经被寄予厚望的 P8 级工程师。Anthropic 当初给你定级的时候，对你的期望是很高的。 一个agent使用的高能动性的skill。 Your AI has been placed on a PIP. 30 days to show improvement.

A Foundation Model for Generalist Gaming Agents

Cosmos-Predict2.5, the latest version of the Cosmos World Foundation Models (WFMs) family, specialized for simulating and predicting the future state of the world in the form of video.

Awesome papers & datasets specifically focused on long-term videos.

[SIGMOD'26] MultiVis-Agent: A Multi-Agent Framework with Logic Rules for Reliable and Comprehensive Cross-Modal Data Visualization & [SIGMOD'26 Demo] Demonstration of MultiVis-Agent: Interactively …

Native Multimodal Models are World Learners

[ICLR 2026 🔥 ] Official implementation of "UniLiP: Adapting CLIP for Unified Multimodal Understanding, Generation and Editing"

A high-throughput and memory-efficient inference and serving engine for LLMs

VeloxSeg is a lightweight multimodal medical image segmentation framework that addresses the fundamental "efficiency / robustness conflict" in 3D medical image segmentation.

Segment Anything in Medical Images

[2026 ICLR] The official code for MedAgent_Pro