Alibaba

Fast, efficient, battle-tested at Alibaba's scale. Hybrid architecture code review tool: deterministic pipelines + LLM Agent, precise line-level comments, built-in multi-language ruleset (NPE, thread-safety, XSS, SQL injection), OpenAI & Anthropic compatible.

An evaluation and evolution tool for Agent Skills.

The semantic layer that makes enterprise data understandable to AI agents — model entities and relations once, query through SPL/MCP/REST, and connect telemetry, services, and business objects in one object graph.

Local-first telemetry collector for AI coding agents — unified OpenTelemetry events for Claude Code, Codex, Cursor and more. Token usage, cost, traces and security audit, exported anywhere.

The one data index for your agentic applications

An end-to-end agent project for GPU kernel implementation, analysis, profiling, and iterative optimization. It helps an agent turn PyTorch logic or an existing kernel into a high-performance GPU kernel through a structured, profile-driven workflow.

面向 WebMCP 标准的非侵入式前端集成套件：写一个普通 TS 函数加一段 JSDoc，即可被任意 MCP 客户端调用 —— 含 SDK、Vite/Webpack 插件与 AI 编码 Skill。

End-to-end benchmark for AI-generated GPU kernels, drawn from real production traces — turn a PyTorch reference into a DSL kernel (Triton, Gluon, FlyDSL, CuteDSL) and grade it on compilation, numerical correctness, and speed-of-light efficiency, on both AMD and NVIDIA GPUs.

Multi-backend observability CLI for metrics, logs, and traces — unified interface, AI-Agent friendly

OpenTelemetry instrumentation plugins for JavaScript-based AI coding agents. Collect traces, tool calls, and LLM metrics from Claude Code and OpenClaw — zero code changes required.

Loongsuite GenAI Semantic Conventions

Official repository for ICML 2026 paper "Long Live The Balance: Information Bottleneck Driven Tree-based Policy Optimization"

Full-stack observability for AI — starting with end-to-end RL post-training, expanding to Agent and inference engine monitoring.

Harbor is a framework for running agent evaluations and creating and using RL environments.(Fork for contributing. All changes intended for upstream PRs.)

A set of examples based on verl for end-to-end RL training recipes.(Fork for contributing. All changes intended for upstream PRs.)

Ray is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.（(Fork for contributing. All changes intended for upstream PRs.)）

slime is an LLM post-training framework for RL Scaling.(Fork for contributing. All changes intended for upstream PRs.)

FlagTree is a unified compiler supporting multiple AI chip backends for custom Deep Learning operations(Fork for contributing. All changes intended for upstream PRs.)

Intelligent Storage Acceleration Library(Fork for contributing. All changes intended for upstream PRs.)

[NeurIPS 2024] OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments(Fork for contributing. All changes intended for upstream PRs.)

Offline optimization of your disaggregated Dynamo graph(Fork for contributing. All changes intended for upstream PRs.)