The roadmap of long-horizon agents

Framework for evaluating and improving agents

EdgeBench: Unveiling scaling laws of learning from real-world environments

A benchmark for evaluating AI agents on realistic business workflows

Harness for running and evaluating AI agents against RL environments

Benchmark self-evolving Agent upon realistic large-scale file workspaces

Implementation for: Towards On-Policy Data Evolution for Visual-Native Multimodal Deep Search Agents

Context engineering for AI Agents. Manage your LLM's context window with caching, compaction, summarization, and graceful degradation.

Trae Agent is an LLM-based agent for general purpose software engineering tasks.

🔥 A collection of the Claude Code open source

The agent benchmark that scores the full stack — harness, config, and model — not just the LLM. Trace-based scoring, reliability metrics, configuration diagnostics.

The agent that grows with you

Official PyTorch implementation of "Visually-grounded Humanoid Agents"

"百战百胜，非善之善者也；不战而屈人之兵，善之善者也。"

毛选.skill — 让毛泽东的思维框架帮你分析问题、制定策略、看透本质。7个核心心智模型 · 10条决策启发式 · 完整表达DNA。不是复读语录，是用他的认知框架帮你看问题。

An agent-managed museum exhibit, built in Rust with Gajae-Code / LazyCodex — developed and maintained with no human intervention.

feishu-cli 是一个功能完整的飞书开放平台命令行工具。它将飞书文档、知识库、电子表格、消息、日历、任务等操作封装为简洁的命令行接口，核心能力是 Markdown ↔ 飞书文档双向无损转换。

Create beautiful slides on the web using a coding agent's frontend skills

Toolathlon-Gym for testing AI agents real-world tool-use capabilities across diverse MCP servers.

Claw-Eval is an evaluation harness for evaluating LLM as agents. All tasks verified by humans.

Bash is all you need - A nano claude code–like 「agent harness」, built from 0 to 1

Create, Evaluate, and Connect AI Skills

Agent World Model: Infinity Synthetic Environments for Agentic Reinforcement Learning

Yunjue Agent: A Fully Reproducible, Zero-Start In-Situ Self-Evolving Agent System for Open-Ended Tasks