agentscope-benchmark-runner

该 Skill 为 AgentScope 2.0 智能体和流水线生成基准场景文件与 Python 运行脚本，支持延迟、吞吐量和提示词/模型对比场景，并输出 JSON/CSV 结果用于复现实验；真实模型运行需配置凭据，生成配置使用占位 API Key。

热度 7.65

最近更新 2026-10-03 10:11:21

分类：开发依赖：代码执行 / 读写文件 / 第三方 API / 内部调用大模型应用场景：代码生成 / 自动化脚本 / 报告撰写

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add agentscope-benchmark-runner

Metadata

name: agentscope-benchmark-runner
description: Run benchmarks against AgentScope 2.0 agents and pipelines. Use this when the user wants to measure latency, throughput, token usage, or compare model/prompt performance in a reproducible way.
version: 0.1.0
metadata: {"openclaw":{"requires":{"env":["OPENAI_API_KEY"],"bins":["python3"]},"primaryEnv":"OPENAI_API_KEY","envVars":[{"name":"OPENAI_API_KEY","required":false,"description":"Optional API key for live benchmark runs."},{"name":"AGENTSCOPE_MODEL","required":false,"description":"Optional default model for benchmark configs."}],"emoji":"🏁","homepage":"https://github.com/agentscope-ai/agentscope"}}

This skill generates benchmark configs and a runner for AgentScope 2.0 agents. It focuses on reproducible performance measurement for:

Run the generator script from this skill folder:

python3 scripts/generate_benchmark.py --output /path/to/benchmarks --scenario latency

Supported scenarios:

After generation:

cd /path/to/benchmarks
python run_benchmark.py