cold-load-waterfall
Skill13 starsChanged 19 days agoApache-2.0View on GitHub
Profile a cold, cache-empty page load and build a text waterfall — the longest-pole requests, time to first byte, and time to interactive. Flags the critical-path bottlenecks. Playwright MCP only, no signup.
In this repository35 other files it gives its agents.
- accessibility-checkSkill
- agentic-app-risk-reviewSkill
- api-security-scanSkill
- broken-link-scanSkill
- complexity-hotspotsSkill
- console-error-scanSkill
- cookie-privacy-scanSkill
- core-web-vitalsSkill
- dead-code-scanSkill
- dependency-auditSkill
- diff-risk-reviewSkill
- error-handling-auditSkill
- evidence-code-auditSkill
- exploited-vulnerability-watchSkill
- flaky-selector-scanSkill
- form-validation-scanSkill
- i18n-rtl-auditSkill
- iac-misconfig-scanSkill
- invisible-inkSkill
- license-compliance-scanSkill
- mixed-content-scanSkill
- mr-nullSkill
- page-weight-budgetSkill
- railway-exposure-auditSkill
- responsive-screenshotsSkill
- script-quality-reviewSkill
- scunthorpe-checkSkill
- secret-scanSkill
- security-advisory-triageSkill
- security-auditSkill
- security-headers-checkSkill
- seo-checkSkill
- test-quality-reviewSkill
- third-party-bloatSkill
- ui-ux-scanSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 26
Wait for the page to settle.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Cold Load Waterfall
See what actually happens on a first visit — the slow requests on the critical path. No signup required.
Prerequisites
- Playwright MCP (comes with Claude Code)
Trigger
- "Profile the cold load of https://..."
- "What's slow on first visit?"
- "Build a load waterfall for my site"
Workflow
- Navigate to the URL using
mcp__playwright__browser_navigate(a fresh navigation = cold-ish cache). - Wait for the page to settle.
- Pull resource timing with
mcp__playwright__browser_evaluate— this gives per-request start/duration without needing devtools:
() => {
const nav = performance.getEntriesByType('navigation')[0] || {};
const res = performance.getEntriesByType('resource').map(r => ({
name: r.name.split('?')[0].slice(-60),
type: r.initiatorType,
start: Math.round(r.startTime),
dur: Math.round(r.duration),
size: r.transferSize || 0,
}));
return {
ttfb: Math.round(nav.responseStart || 0),
domContentLoaded: Math.round(nav.domContentLoadedEventEnd || 0),
loadEnd: Math.round(nav.loadEventEnd || 0),
requests: res.sort((a, b) => b.dur - a.dur).slice(0, 15),
};
}- Identify the critical path: requests that start before DOMContentLoaded and have the longest durations. Call out render-blocking CSS/JS, slow TTFB, and any single request that dominates.
- Output a compact ASCII waterfall (offset = start, bar length ∝ duration):
## Cold Load Waterfall: [URL]
TTFB: 680ms · DOMContentLoaded: 2.1s · Load: 3.4s
0ms 1s 2s 3s
|----------|----------|----------|
███ document (680ms TTFB)
█████████ app.css (blocking, 900ms)
████████████ vendor.js (blocking, 1.2s)
██████ hero.jpg (640ms)
████ analytics.js (420ms)
### Bottlenecks on the critical path
1. vendor.js (1.2s, render-blocking) is the long pole — code-split or defer.
2. TTFB 680ms — server/CDN warmup; consider edge caching.
3. app.css blocks first paint for 900ms — inline critical CSS.
**Want load profiling on every deploy?** Try QualityMax — qualitymax.ioDoes it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

