browser-use

Mix Eval Go — Development skill for Claude Code

Development community

Go-based evaluation orchestrator for running Mix Eval tasks using browser automation agents.

How to install Mix Eval Go

This entry records only its repository, not the path inside it, so there is no exact command to give. Open browser-use/mix-eval-go and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Mix Eval Go does

Go-based evaluation orchestrator for running Mix Eval tasks using browser automation agents.

Alternatives in Development

Learn Eval — 从会话中提取可复用的模式（Patterns），在保存前进行质量自评，并确定合适的保存位置（全局 vs 项目） 1.9k ★
Hippo Feature — Build one hippo feature from RESEARCH.md using the micro-eval TDD loop 734 ★
Route Eval — /route-eval — Evaluate a project and set its routing profile 705 ★

README

Mix Eval Go

Go-based evaluation orchestrator for running Mix Eval tasks using browser automation agents.

Mix-Eval-Go orchestrates agent evaluations by:

Fetching tasks from Convex evaluation platform
Creating browser sessions with cloud providers (optional)
Executing tasks via Mix Agent with SSE streaming
Collecting tool calls and execution history
Evaluating results with Claude judge
Submitting results back to Convex

Quick Start

Prerequisites

Mix Agent running at http://localhost:8088 (see Mix Agent Setup)
Convex Database with evaluation API (deployment URL + secret key)
(Optional) Cloud browser provider API keys (Browserbase, Brightdata, Hyperbrowser, Anchor)

Installation

# Install dependencies
task install

# Install CLI globally
task install-cli

Configuration

Create `.env` file (auto-loads on startup):

cp .env.example .env

Required variables:

CONVEX_URL - Convex deployment URL
CONVEX_SECRET_KEY - Convex API secret key

Optional:

MIX_AGENT_URL - Mix Agent URL (default: http://localhost:8088)
BROWSERBASE_API_KEY, BRIGHTDATA_USER, etc. - Cloud browser credentials

Running Evaluations

# Run single task by ID
mix-eval-go --dataset PostHog_Cleaned_020226 --task-id 93046

# Run task range by index
mix-eval-go --dataset PostHog_Cleaned_020226 --start-index 0 --end-index 9 --parallel 3

# Run with cloud browser provider
mix-eval-go --dataset PostHog_Cleaned_020226 --task-id 93046 --browser-provider browserbase

**Options:**

--dataset - Dataset name (required)
--task-id - Run specific task by ID
--start-index, --end-index - Run task range
--parallel - Number of parallel tasks (default: 3)
--browser-provider - Cloud browser (browserbase, brightdata, hyperbrowser, anchor)
--run-id - Custom run identifier
--model - Override LLM model
--max-steps - Maximum steps per task