web performance regression detective

The problem

Manual performance audits often fail to distinguish between random network noise and actual code regressions. Developers waste hours chasing minor metric fluctuations while critical Core Web Vitals issues slip into production undetected.

What it does

• Analyzes metric snapshots to identify statistically significant performance regressions in LCP, FID, CLS, and TTI.
• Correlates performance drops with specific deployments, asset updates, or third-party script changes.
• Scores and ranks remediation tasks based on user impact, business risk, and implementation effort.
• Generates structured re-verification plans including target metrics and synthetic monitoring tasks.

Why this beats prompting it yourself

General-purpose prompts often treat every metric increase as a bug. This skill uses specific statistical thresholds and heuristic mapping to filter out noise, ensuring you only alert the team for regressions that actually impact SEO and conversion rates.

Use cases

• Post-deployment sanity checks to catch layout shifts or bundle size bloat.
• Third-party dependency monitoring to detect slow external scripts.
• Prioritizing technical debt by quantifying the conversion risk of specific performance bottlenecks.

Known limitations

Requires accurate input metrics, as missing data leads to false classifications. Root cause detection is heuristic-based and may require manual verification for complex infrastructure issues.