# Diagnose and Improve Full-Stack Performance

Full-stack teams often lack a repeatable way to measure performance and locate bottlenecks. This skill provides testing workflows, diagnostic commands, optimization guidance, and report structures.

## Install

```bash
npx skillstore add zl2023github/perf-fullstack-engineer
```

## Metadata

- Status: approved
- Slug: zl2023github-perf-fullstack-engineer
- Skillstore revision: r2
- Version status: missing
- Tree hash: a0e21b69662fcdab760eeeea83ee14cb98235ff75a7ddadb8ade701587f84971
- Author: zl2023github
- GitHub username: zl2023github
- License: MIT
- Repository: https://github.com/zl2023github/software-engineer-skills/tree/main/software-engineering/perf-fullstack-engineer
- Ref: 88a8e9a07f4c54ab105c1c41b6267c287146b07b
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands, network, env\_access
- Quality score: 77
- Quality tier: bronze
- Public page: https://skillstore.pages.dev/skills/zl2023github-perf-fullstack-engineer
- Manifest: https://skillstore.pages.dev/api/skills/zl2023github-perf-fullstack-engineer/manifest

## Capabilities

- Defines TPS, QPS, latency percentiles, concurrency, error rate, throughput, and resource targets.
- Designs baseline, load, stress, endurance, peak, and capacity tests.
- Provides a k6 example and selection guidance for JMeter, Locust, wrk, Vegeta, and Apache Bench.
- Guides diagnosis across hosts, JVMs, databases, applications, containers, and monitoring platforms.
- Covers frontend, backend, API, database, architecture, deployment, and operations workflows.
- Structures performance reports with environments, scenarios, trends, bottlenecks, recommendations, and conclusions.

## Use Cases

- Establish a Performance Baseline: Define service targets, select test scenarios, and produce a repeatable baseline before a release.
- Diagnose a Slow Application: Correlate latency, resource, runtime, database, and application signals to prioritize likely bottlenecks.
- Plan Capacity for Growth: Model peak traffic, identify system limits, and organize scaling recommendations with measurable acceptance criteria.

## Prompt Templates

### Define a Baseline

```
Create a performance baseline for [service]. Use [traffic profile], [latency target], [error target], and [environment]. List assumptions and required metrics.
```

### Design a Load Test

```
Design a [tool] load test for [workflow]. Include stages, data preparation, checks, thresholds, monitoring, cleanup, and a safe execution sequence.
```

### Analyze a Bottleneck

```
Analyze these performance results: [results]. Correlate application, host, database, runtime, and network evidence. Rank hypotheses and propose verification steps.
```

### Build a Capacity Plan

```
Create a capacity plan for [system] at [forecast load]. Estimate limits, identify architecture risks, compare scaling options, and define regression gates.
```

## Limitations

- The skill contains guidance only, with no executable files, integrations, or automated result parser.
- Commands, URLs, thresholds, and test data are examples that require project-specific adaptation.
- Results depend on representative environments, workloads, monitoring coverage, and repeated measurements.
- Third-party tools must be installed, and infrastructure access must be authorized.

## Best Practices

- Confirm target ownership, isolate the test environment, and obtain approval before generating load.
- Warm the system, monitor every layer, increase load gradually, and repeat comparable tests.
- Base conclusions on correlated metrics and document assumptions, environment differences, and data limitations.

## Anti Patterns

- Do not run stress tests against production or third-party systems without explicit authorization.
- Do not optimize from averages alone; inspect percentiles, errors, saturation, and resource trends.
- Do not change multiple variables between baseline and comparison runs.

## Security Audit

- Audited at: 2026-07-23T23:51:31.261\+00:00
- Summary: The skill is readable Markdown guidance with no executable package, prompt injection, credential collection, obfuscation, or covert data transfer. Static alerts primarily misread Markdown backticks, Chinese text, placeholders, localhost URLs, and documented developer commands as dangerous behavior.

## Stats

- Views: 2
- Downloads: 7
- Favorites: 0
- Popularity score: 0