How long does the integration take?

Change one base URL in an OpenAI-compatible client, or install the Python or TypeScript SDK; the quickstart routes a first request in a few minutes.

Will this lock us into another vendor?

No. The router is Apache-2.0 open source, your provider keys stay yours, and the API is OpenAI-compatible, so leaving is a change of base URL.

What happens when a provider fails?

The router falls back to the next model your policy allows and records the fallback in the trace, so an outage costs a retry, not an incident.

load-testing-expert - Skill

Name: load-testing-expert
Author: personamanagmentlayer

Imported from personamanagmentlayer/pcl (stdlib/qa/load-testing-expert/SKILL.md). Install upstream with npx skills add personamanagmentlayer/pcl --skill load-testing-expert. Copyright stays with the author.

Load Testing Expert

Core Concepts

Load Testing Types

Load Testing - Expected load performance
Stress Testing - Beyond capacity limits
Spike Testing - Sudden load increases
Soak Testing - Extended duration performance
Scalability Testing - System growth capacity
Volume Testing - Large data handling

Key Metrics

Response Time - Request completion time
Throughput - Requests per second (RPS)
Error Rate - Percentage of failed requests
Latency - Time to first byte (TTFB)
Concurrent Users - Simultaneous active users
Resource Utilization - CPU, memory, network

Tools & Frameworks

Apache JMeter - Java-based load testing
k6 - Modern JavaScript load testing
Gatling - Scala-based performance testing
Locust - Python-based load testing
Artillery - Node.js load testing
Vegeta - Go HTTP load testing

Best Practices

Test Design

Start with baseline performance
Define clear performance goals
Use realistic load patterns
Include think time between requests
Test with production-like data
Monitor system resources

Load Patterns

Ramp-up gradually
Test at sustained load
Include spike scenarios
Test soak/endurance
Validate under stress
Test auto-scaling

Metrics & Analysis

Define SLAs upfront
Monitor response times (p50, p95, p99)
Track error rates
Measure throughput
Monitor resource utilization
Analyze bottlenecks

Environment

Use dedicated test environment
Match production configuration
Isolate test traffic
Monitor database performance
Test with CDN if applicable
Include third-party dependencies

Anti-Patterns

Common Mistakes

Testing from single location only
Not ramping up load gradually
Ignoring think time
Testing without monitoring
No baseline for comparison
Testing in production

Test Design Issues

Unrealistic load patterns
Missing data variation
No error handling
Hard-coded test data
Ignoring cache effects
Not testing edge cases

Analysis Problems

Focusing only on averages
Ignoring outliers
Not correlating metrics
Missing resource bottlenecks
No root cause analysis
Incomplete reporting

Reference Documentation

Detailed material lives alongside this skill and is read on demand:

Implementation Examples — k6 Load Test Script, JMeter Test Plan (XML), Gatling Load Test (Scala), Python Locust Example

Resources

Official Documentation

k6 Documentation - Modern load testing
JMeter Manual - Complete guide
Gatling Documentation - Performance testing
Locust Documentation - Python load testing

Learning Resources

Performance Testing Guidance - Best practices
k6 YouTube - Video tutorials
JMeter Academy - Training
Awesome Load Testing - Curated resources

Tools & Platforms

BlazeMeter - JMeter cloud platform
k6 Cloud - k6 SaaS platform
Grafana Cloud k6 - Monitoring integration
Artillery Pro - Enterprise features

Community Resources

k6 Community - Forums
JMeter Forum - Mailing lists
Load Testing Slack - k6 Slack channel
Performance Testing Reddit - Discussions

Install load-testing-expert and it becomes one more candidate the router can pick - when it fits.

A listing is a routing target with a manifest: what it does, which domains it covers, what it costs and who publishes it. Once installed it sits beside your own models and tools, is scored like any other candidate for each request, and shows up in the trace when it wins. Ratings come from workspaces that installed it, one per account.

{ "ocm": "1", "id": "personamanagmentlayer-pcl-load-testing-expert", "kind": "skill", "name": "load-testing-expert", "description": "Expert in performance and load testing using JMeter, k6, Gatling, load patterns, metrics analysis, and performance optimization. Use when the user mentions performance, testing, QA, scalability, performance testing, or JMeter, or when the task involves Load Testing Types, Key Metrics, Tools & Frameworks, or k6 Load Test Script.", "publisher": "personamanagmentlayer", "version": "1.0.0", "capabilities": { "domains": [ "observability", "coding" ], "tags": [ "skill-md", "performance", "load-testing", "testing", "qa", "scalability", "performance-testing", "jmeter", "k6", "gatling" ], "languages": [ "en" ] }, "quality_prior": 0.6, "examples": [ "Expert in performance and load testing using JMeter, k6, Gatling, load patterns, metrics analysis, and performance optimization. Use when the user mentions performance, testing, QA, scalability, performance testing, or JMeter, or when the task involves Load Testing Types, Key Metrics, Tools & Frameworks, or k6 Load Test Script." ], "primary": false, "metadata": { "source": { "provider": "skills.sh", "repository": "https://github.com/personamanagmentlayer/pcl", "path": "stdlib/qa/load-testing-expert/SKILL.md", "ref": "HEAD", "url": "https://github.com/personamanagmentlayer/pcl/blob/HEAD/stdlib/qa/load-testing-expert/SKILL.md", "key": "personamanagmentlayer/pcl/stdlib/qa/load-testing-expert/SKILL.md" }, "allowed_tools": [ "Read", "Write", "Bash", "Grep", "Glob" ] }, "instructions": "# Load Testing Expert\n\n## Core Concepts\n\n### Load Testing Types\n\n- **Load Testing** - Expected load performance\n- **Stress Testing** - Beyond capacity limits\n- **Spike Testing** - Sudden load increases\n- **Soak Testing** - Extended duration performance\n- **Scalability Testing** - System growth capacity\n- **Volume Testing** - Large data handling\n\n### Key Metrics\n\n- **Response Time** - Request completion time\n- **Throughput** - Requests per second (RPS)\n- **Error Rate** - Percentage of failed requests\n- **Latency** - Time to first byte (TTFB)\n- **Concurrent Users** - Simultaneous active users\n- ", "cost": { "context_tokens": 944 } }

load-testing-expert

About