From e8760bb7cc579aa75e0c098ac28a22d7428af22f Mon Sep 17 00:00:00 2001 From: genitrix Date: Tue, 28 Jul 2026 18:34:14 +0800 Subject: [PATCH] Add ClawBench evaluation benchmark --- AWESOME.md | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/AWESOME.md b/AWESOME.md index f17ae32..2bfd00d 100644 --- a/AWESOME.md +++ b/AWESOME.md @@ -123,6 +123,12 @@ A carefully curated list of AI coding assistant tools, including Skills, MCP ser - Description: Universal skills loader - Usage: `npm i -g openskills && openskills install anthropics/skills` +## 📊 Agent Evaluation & Benchmarks + +- **[ClawBench](https://github.com/TIGER-AI-Lab/ClawBench)** by TIGER-AI-Lab + - Description: Open-source live-web benchmark for browser and computer-use agents, covering 283 V1/V2 tasks across 163 live websites. + - Features: Isolated Docker runs, selectable agent harnesses, request interception, and replayable action, screenshot, recording, HTTP, and agent-message evidence; [paper](https://arxiv.org/abs/2604.08523) and [project](https://claw-bench.com). + ## 🔧 Recommended MCP Servers ### Official MCP Servers (Model Context Protocol)