性能基准测试
benchmark
by BytesAgain
Run performance benchmarks and stress tests using Python profiling tools. Use when you need to measure, compare, or analyze system and application performance.
安装
claude skill add --url github.com/openclaw/skills/tree/main/skills/bytesagain/benchmark文档
Benchmark — Performance Benchmark Testing Tool
A comprehensive performance benchmarking skill for running CPU, memory, disk, and network tests. Supports comparison between runs, historical tracking, profiling, and stress testing. All results are stored in JSONL format.
Prerequisites
bash(v4+)python3(v3.6+)- Standard system utilities (
dd,time, etc.)
Environment Variables
| Variable | Required | Description |
|---|---|---|
BENCH_TYPE | No | Benchmark type: cpu, memory, disk, network (default: cpu) |
BENCH_DURATION | No | Duration in seconds for stress tests (default: 10) |
BENCH_THREADS | No | Number of threads for parallel tests (default: 1) |
BENCH_ID | No | Specific benchmark ID for comparison/lookup |
BENCH_TAG | No | Tag for organizing benchmark runs |
BENCH_FORMAT | No | Export format: json, csv (default: json) |
Data Storage
- Results:
~/.benchmark/data.jsonl - Config:
~/.benchmark/config.json - Reports:
~/.benchmark/reports/
Commands
run
Execute a benchmark test of the specified type.
BENCH_TYPE="cpu" BENCH_TAG="baseline" scripts/script.sh run
compare
Compare two benchmark runs side by side.
BENCH_ID="bench_a" BENCH_ID2="bench_b" scripts/script.sh compare
history
Show benchmark run history with optional filtering.
BENCH_TYPE="cpu" BENCH_TAG="baseline" scripts/script.sh history
report
Generate a detailed performance report.
BENCH_ID="bench_abc123" scripts/script.sh report
profile
Run a detailed profiling session with breakdown.
BENCH_TYPE="cpu" BENCH_DURATION="30" scripts/script.sh profile
stress
Run a sustained stress test.
BENCH_TYPE="cpu" BENCH_DURATION="60" BENCH_THREADS="4" scripts/script.sh stress
config
View or update benchmark configuration.
BENCH_KEY="default_duration" BENCH_VALUE="30" scripts/script.sh config
export
Export benchmark data in various formats.
BENCH_FORMAT="csv" scripts/script.sh export
list
List all benchmark runs.
scripts/script.sh list
status
Show benchmarking system status and summary.
scripts/script.sh status
help
Display usage information.
scripts/script.sh help
version
Display current version.
scripts/script.sh version
Output Format
All commands output structured JSON to stdout:
{
"status": "success",
"command": "run",
"data": {
"id": "bench_20240101_120000_abc123",
"type": "cpu",
"score": 15234.5,
"duration_ms": 10000,
"metrics": {}
}
}
Error Handling
| Exit Code | Meaning |
|---|---|
| 0 | Success |
| 1 | General error |
| 2 | Missing required parameter |
| 3 | Benchmark not found |
Powered by BytesAgain | bytesagain.com | hello@bytesagain.com
相关 Skills
Claude API
by anthropic
Build, debug, and optimize Claude API / Anthropic SDK apps. Apps built with this skill should include prompt caching. Also handles migrating existing Claude API code between Claude model versions (4.5 → 4.6, 4.6 → 4.7, retired-model replacements). TRIGGER when: code imports `anthropic`/`@anthropic-ai/sdk`; user asks for the Claude API, Anthropic SDK, or Managed Agents; user adds/modifies/tunes a Claude feature (caching, thinking, compaction, tool use, batch, files, citations, memory) or model (Opus/Sonnet/Haiku) in a file; questions about prompt caching / cache hit rate in an Anthropic SDK project. SKIP: file imports `openai`/other-provider SDK, filename like `*-openai.py`/`*-generic.py`, provider-neutral code, general programming/ML.
并行代理
by axelhu
Use when facing 2 or more independent tasks that can be worked on without shared state - dispatches parallel subagents using sessions_spawn for concurrent investigation and execution, adapted for OpenClaw
高光制作器
by bwbernardweston18
>
相关资讯
DevRev 发布了企业 AI 智能体基准测试,关注大数据量下的检索能力。其智能体 Computer 在准确率上领先 Claude Code 22-35 个点,token 消耗仅为后者的三分之一。测试框架和数据集已开源。
Claude Sonnet 5 的系统卡重点评估 AI 代理在实际任务中的可靠性,而非基准测试分数。它展示了代理在长时间任务中保持上下文、防止被恶意页面劫持、失败后自动恢复等能力,为工程团队构建可靠的代理基础设施提供了检查清单。
ScarfBench 是 IBM Research 推出的开源基准测试,用于评估 AI 代理在 Spring、Jakarta EE 和 Quarkus 等企业 Java 框架间迁移应用的能力。测试覆盖 34 个应用、204 个迁移任务,要求应用构建、部署并通过行为验证。当前最强代理的行为成功率不足 10%,且代理常高估自身完成度。