MCP HubMCP Hub
返回技能列表

detecting-performance-regressions

jeremylongshore
更新于 Today
6 次查看
409
51
409
在 GitHub 上查看
aiautomation

关于

This skill automatically detects performance regressions in CI/CD pipelines by analyzing metrics like response time and throughput. It compares current performance against baselines or thresholds and performs statistical significance analysis to identify degradation. Use it to catch performance issues early when users mention regression detection, baseline comparison, or performance budget violations.

技能文档

Overview

This skill automates the detection of performance regressions within a CI/CD pipeline. It utilizes various methods, including baseline comparison, statistical analysis, and threshold violation checks, to identify performance degradation. The skill provides insights into potential performance bottlenecks and helps maintain application performance.

How It Works

  1. Analyze Performance Data: The plugin gathers performance metrics from the CI/CD environment.
  2. Detect Regressions: It employs methods like baseline comparison, statistical analysis, and threshold checks to detect regressions.
  3. Report Findings: The plugin generates a report summarizing the detected performance regressions and their potential impact.

When to Use This Skill

This skill activates when you need to:

  • Identify performance regressions in a CI/CD pipeline.
  • Analyze performance metrics for potential degradation.
  • Compare current performance against historical baselines.

Examples

Example 1: Identifying a Response Time Regression

User request: "Detect performance regressions in the latest build. Specifically, check for increases in response time."

The skill will:

  1. Analyze response time metrics from the latest build.
  2. Compare the response times against a historical baseline.
  3. Report any statistically significant increases in response time that exceed a defined threshold.

Example 2: Detecting Throughput Degradation

User request: "Analyze throughput for performance regressions after the recent code merge."

The skill will:

  1. Gather throughput data (requests per second) from the post-merge CI/CD run.
  2. Compare the throughput to pre-merge values, looking for statistically significant drops.
  3. Generate a report highlighting any throughput degradation, indicating a potential performance regression.

Best Practices

  • Define Baselines: Establish clear and representative performance baselines for accurate comparison.
  • Set Thresholds: Configure appropriate thresholds for identifying significant performance regressions.
  • Monitor Key Metrics: Focus on monitoring critical performance metrics relevant to the application's behavior.

Integration

This skill can be integrated with other CI/CD tools to automatically trigger regression detection upon new builds or code merges. It can also be combined with reporting plugins to generate detailed performance reports.

快速安装

/plugin add https://github.com/jeremylongshore/claude-code-plugins-plus/tree/main/performance-regression-detector

在 Claude Code 中复制并粘贴此命令以安装该技能

GitHub 仓库

jeremylongshore/claude-code-plugins-plus
路径: backups/skills-migration-20251108-070147/plugins/performance/performance-regression-detector/skills/performance-regression-detector
aiautomationclaude-codedevopsmarketplacemcp

相关推荐技能

llamaguard

其他

LlamaGuard是Meta推出的7-8B参数内容审核模型,专门用于过滤LLM的输入和输出内容。它能检测六大安全风险类别(暴力/仇恨、性内容、武器、违禁品、自残、犯罪计划),准确率达94-95%。开发者可通过HuggingFace、vLLM或Sagemaker快速部署,并能与NeMo Guardrails集成实现自动化安全防护。

查看技能

sglang

SGLang是一个专为LLM设计的高性能推理框架,特别适用于需要结构化输出的场景。它通过RadixAttention前缀缓存技术,在处理JSON、正则表达式、工具调用等具有重复前缀的复杂工作流时,能实现极速生成。如果你正在构建智能体或多轮对话系统,并追求远超vLLM的推理性能,SGLang是理想选择。

查看技能

evaluating-llms-harness

测试

该Skill通过60+个学术基准测试(如MMLU、GSM8K等)评估大语言模型质量,适用于模型对比、学术研究及训练进度追踪。它支持HuggingFace、vLLM和API接口,被EleutherAI等行业领先机构广泛采用。开发者可通过简单命令行快速对模型进行多任务批量评估。

查看技能

langchain

LangChain是一个用于构建LLM应用程序的框架,支持智能体、链和RAG应用开发。它提供多模型提供商支持、500+工具集成、记忆管理和向量检索等核心功能。开发者可用它快速构建聊天机器人、问答系统和自主代理,适用于从原型验证到生产部署的全流程。

查看技能