Skip to content
Skills Library
Programming · Analysis

performance-observability

/performance-observability

Use to diagnose performance with baselines, profiling, system measurements, and safe observability practices, and to verify changes against evidence.

Licence
MIT
Last repository change
05.11.2026

Assessment and security

100/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content35 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
1
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 05.11.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What it does and when to use it

Use to diagnose performance with baselines, profiling, system measurements, and safe observability practices, and to verify changes against evidence.

Original author description

Use to diagnose performance with baselines, profiling, system measurements, and safe observability practices, and to verify changes against evidence.

How it works

  1. Scope and handoffs
  2. Measurement workflow
  3. Completion checklist
Original SKILL.md headings · 29 lines, 464 words
SKILL.md preview
---
name: performance-observability
description: Use to diagnose performance with baselines, profiling, system measurements, and safe observability practices, and to verify changes against evidence.
---

# Performance and Observability

## Scope and handoffs

Own measurement-led performance diagnosis, profiling, instrumentation choices, and regression verification. Begin with [fullstack-core](../fullstack-core/SKILL.md) for cross-layer task coordination. Hand UI implementation to [react-typescript](../react-typescript/SKILL.md), database/query implementation to [database-prisma](../database-prisma/SKILL.md), service implementation to [node-api](../node-api/SKILL.md), and architecture tradeoffs to [system-design](../system-design/SKILL.md). Use [debugging-root-cause](../debugging-root-cause/SKILL.md) when the task is primarily fault isolation.

## Measurement workflow

1. **Inspect first.** Read repository instructions and worktree state; identify the affected user journey, runtime, existing telemetry/profilers, performance tests, deployment shape, and relevant dependency boundaries. Use only tools and environments actually available.
2. **Set a target and baseline.** Define a user- or system-relevant measurable target, workload, environment, and time window before optimization. Measure an appropriate baseline, such as latency percentiles, throughput, error rate, CPU/memory/network/database use, and user-visible rendering or interaction delay. Record sample size and conditions; distinguish measurements from estimates.
3. **Profile before editing.** Reproduce the representative workload and use an available profiler, trace, query plan, browser/network inspection, or other appropriate diagnostic to locate the bottleneck. Consider frontend render and network behavior, API execution, database queries, cache behavior, serialization, and downstream dependencies as applicable. Do not optimize by intuition alone.
4. **Instrument carefully.** Prefer existing structured logs, metrics, traces, and correlation/request identifiers. Add only signals needed to answer a concrete diagnostic question. Check metric label cardinality, sensitive data exposure, instrumentation overhead, sampling, and retention implications. Never put credentials, raw sensitive payloads, or high-cardinality user identifiers into metric labels.
5. **Make a focused change.** Address the measured bottleneck with the smallest viable change. Avoid adding monitoring infrastructure, caches, dependencies, or broad refactors absent evidence and need. Check that correctness, resource bounds, and failure behavior remain intact.
6. **Compare fairly.** Repeat the baseline workload under comparable conditions and compare the same metric, percentiles, sample, and time window. Check errors and resource costs as well as latency; report measured results with methodology and limitations. Separate observed outcomes from hypotheses about other workloads.
7. **Guard against regressions.** Add or adapt an existing repeatable benchmark, performance assertion, or operational signal only when it meaningfully detects regression without flaky timing assumptions. Use [testing-quality](../testing-quality/SKILL.md) for test strategy and CI integration.
8. **Report and hand off.** State the measured bottleneck, change, before/after evidence, unavailable checks, and residual uncertainty. If a production symptom remains, coordinate with [cloud-production-ops](../cloud-production-ops/SKILL.md); do not claim an unobserved production improvement.

## Completion checklist

- A target and comparable baseline were established before optimization.
- Profiling or equivalent evidence identified the bottleneck being changed.
- Relevant user-visible, latency, throughput, error, and resource dimensions were considered.
- Instrumentation protects sensitive data and accounts for cardinality, overhead, and retention.
- Improvements and tradeoffs are supported by comparable measurements; estimates and assumptions are labeled.
- Regression verification and exact check outcomes or limitations are reported.

This skill does not install monitoring infrastructure or guarantee performance gains without measurement.

Versions

The database has no other skills named performance-observability.

Database snapshot: 13.11.2026 · package version 9a2473323a66 (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 100performance
    Profiling-driven performance optimization: caching, query tuning, N+1 fixes, memory, and load testing. Use to diagnose a bottleneck or set performance budgets.
  • 100integra-performance-audit
    Diagnose frontend loading, responsiveness, interaction latency, rendering and animation cost using reproducible measurements. Use when performance is slow or needs verification; do not optimize or claim gains without comparable evidence.
  • 90performance-analysis
    Measurement approaches, profiling patterns, bottleneck identification, and optimization guidance. Use when diagnosing performance issues, establishing baselines, identifying bottlenecks, or planning for scale. Always measure before optimizing.
  • 100application-performance-performance-optimization
    Optimize end-to-end application performance with profiling, observability, and backend/frontend tuning. Use when coordinating performance optimization across the stack.
  • 100application-performance-performance-optimization
    Optimize end-to-end application performance with profiling, observability, and backend/frontend tuning. Use when coordinating performance optimization across the stack.
  • 90application-performance-performance-optimization
    Optimize end-to-end application performance with profiling, observability, and backend/frontend tuning. Use when coordinating performance optimization across the stack.

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.