agentleFS
nucliwebnucliweb/webperf-snippets/.claude/skills/webperf-core-web-vitals/SKILL.md

webperf-core-web-vitals

Skill1.5k starsChanged 8 months agoMITView on GitHub

Intelligent Core Web Vitals analysis with automated workflows and decision trees. Measures LCP, CLS, INP with guided debugging that automatically determines follow-up analysis based on results. Includes workflows for LCP deep dive (5 phases), CLS investigation (loading vs interaction), INP debugging (latency breakdown + attribution), and cross-skill integration with loading, interaction, and media skills. Use when the user asks about Core Web Vitals, LCP optimization, layout shifts, or interaction responsiveness. Compatible with Chrome DevTools MCP.

  • Installs packages

We split this skill into the rules it gives an agent

18 rules. About 0 are common, 11 are distinctive, and 7 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈11

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 7

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Run the script with mcp__chrome-devtools__evaluate_script and read the object it returns.

    line 19
  2. Very specific:

    When status is "tracking", the script keeps observing the page.

    line 20
  3. Very specific:

    When status is "error" or "unsupported", read error: the page may not have loaded yet, or the browser lacks the API.

    line 21
  4. Distinctive:

    When the user asks for a comprehensive Core Web Vitals analysis or "audit CWV":

    line 40
  5. Distinctive:

    When LCP is slow or the user asks "debug LCP" or "why is LCP slow":

    line 50
  6. Very specific:

    LCP-Image-Entropy.js: - Check if LCP image has visual complexity issues

    line 55
  7. Distinctive:

    When layout shifts are detected or the user asks "debug CLS" or "layout shift issues":

    line 60
  8. Distinctive:

    When interactions feel slow or the user asks "debug INP" or "slow interactions":

    line 70
  9. Very specific:

    When LCP is a video element (detected by LCP-Video-Candidate.js):

    line 80
  10. Distinctive:

    When LCP is an image (most common case):

    line 91
  11. Distinctive:

    Use this decision tree to automatically run follow-up snippets based on results:

    line 105
  12. Distinctive:

    If early in load: wait and re-run the script

    line 177
  13. Distinctive:

    If page is an SPA: user may need to navigate to the target route first

    line 178
  14. Distinctive:

    When multiple CWV metrics are poor, prioritize investigation:

    line 223
  15. Distinctive:

    Check for late-loading content pushing LCP element

    line 226
  16. Very specific:

    Run LCP-Trail.js to see LCP candidate changes

    line 227
  17. Very specific:

    Run Layout-Shift-Loading-and-Interaction.js to correlate timing

    line 228
  18. Distinctive:

    Check if shifts occur during/after interactions

    line 238

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

WebPerf: Core Web Vitals

JavaScript snippets for measuring web performance in Chrome DevTools.

Running a script

  1. Run the script with mcp__chrome-devtools__evaluate_script and read the object it returns. The scripts have no console output, so the returned object is the result: script, status and the metric data (value, rating, details, items, issues), as described in references/schema.md.
  2. When status is "tracking", the script keeps observing the page. After the interactions, call the function named in getDataFn with another mcp__chrome-devtools__evaluate_script call, for example getINP(), and read the object it returns.
  3. When status is "error" or "unsupported", read error: the page may not have loaded yet, or the browser lacks the API.

Without Chrome DevTools MCP, run the same snippet with the webperf-snippets CLI, which loads the page in headless Chromium and prints the same result objects: npx webperf-snippets <url> --snippet <Name> --json, where <Name> is the script name without .js. It needs Playwright and Chromium (npm install --save-dev webperf-snippets playwright, then npx playwright install chromium), and tracking snippets take --interact-script <file> to drive the interactions.

Scripts

  • scripts/CLS.js — Cumulative layout shift (CLS)
  • scripts/INP.js — Interaction to next paint (INP)
  • scripts/LCP-Image-Entropy.js — LCP image entropy
  • scripts/LCP-Subparts.js — LCP subparts
  • scripts/LCP-Trail.js — LCP trail
  • scripts/LCP-Video-Candidate.js — LCP video candidate
  • scripts/LCP.js — Largest contentful paint (LCP)

Common Workflows

Complete Core Web Vitals Audit

When the user asks for a comprehensive Core Web Vitals analysis or "audit CWV":

  1. LCP.js - Measure Largest Contentful Paint
  2. CLS.js - Measure Cumulative Layout Shift
  3. INP.js - Measure Interaction to Next Paint
  4. LCP-Subparts.js - Break down LCP timing phases
  5. LCP-Trail.js - Track LCP candidate evolution

LCP Deep Dive

When LCP is slow or the user asks "debug LCP" or "why is LCP slow":

  1. LCP.js - Establish baseline LCP value
  2. LCP-Subparts.js - Break down into TTFB, resource load, render delay
  3. LCP-Trail.js - Identify all LCP candidates and changes
  4. LCP-Image-Entropy.js - Check if LCP image has visual complexity issues
  5. LCP-Video-Candidate.js - Detect if LCP is a video (poster or video element)

CLS Investigation

When layout shifts are detected or the user asks "debug CLS" or "layout shift issues":

  1. CLS.js - Measure overall CLS score
  2. Layout-Shift-Loading-and-Interaction.js (pending — available in webperf-interaction skill)
  3. Cross-reference with webperf-loading skill:
    • Find-Above-The-Fold-Lazy-Loaded-Images.js (lazy images causing shifts)
    • Fonts-Preloaded-Loaded-and-used-above-the-fold.js (font swap causing shifts)

INP Debugging

When interactions feel slow or the user asks "debug INP" or "slow interactions":

  1. INP.js - Measure overall INP value; call getINP() after interactions, getINPDetails() for the 15 slowest interactions
  2. Interactions.js (pending — available in webperf-interaction skill)
  3. Input-Latency-Breakdown.js (pending — available in webperf-interaction skill)
  4. Long-Animation-Frames.js (pending — available in webperf-interaction skill)
  5. Long-Animation-Frames-Script-Attribution.js (pending — available in webperf-interaction skill)

Video as LCP Investigation

When LCP is a video element (detected by LCP-Video-Candidate.js):

  1. LCP-Video-Candidate.js - Identify video as LCP candidate
  2. Video-Element-Audit.js (from Media skill) - Audit video loading strategy
  3. LCP-Subparts.js - Analyze video loading phases
  4. Cross-reference with webperf-loading skill:
    • Resource-Hints-Validation.js (check for video preload)
    • Priority-Hints-Audit.js (check for fetchpriority on video)

Image as LCP Investigation

When LCP is an image (most common case):

  1. LCP.js - Measure LCP timing
  2. LCP-Subparts.js - Break down timing phases
  3. LCP-Image-Entropy.js - Analyze image complexity
  4. Cross-reference with webperf-media skill:
    • Image-Element-Audit.js (check format, dimensions, lazy loading)
  5. Cross-reference with webperf-loading skill:
    • Find-Above-The-Fold-Lazy-Loaded-Images.js (check if incorrectly lazy)
    • Priority-Hints-Audit.js (check for fetchpriority="high")
    • Resource-Hints-Validation.js (check for preload)

Decision Tree

Use this decision tree to automatically run follow-up snippets based on results:

After LCP.js

  • If LCP > 2.5s → Run LCP-Subparts.js to diagnose which phase is slow
  • If LCP > 4.0s (poor) → Run full LCP deep dive workflow (5 snippets)
  • If LCP candidate is an image → Run LCP-Image-Entropy.js and webperf-media:Image-Element-Audit.js
  • If LCP candidate is a video → Run LCP-Video-Candidate.js and webperf-media:Video-Element-Audit.js
  • Always run → LCP-Trail.js to understand candidate evolution

After LCP-Subparts.js

  • If TTFB phase > 600ms → Switch to webperf-loading skill and run TTFB-Sub-Parts.js
  • If Resource Load Time > 1500ms → Run:
    1. webperf-loading:Resource-Hints-Validation.js (check for preload/preconnect)
    2. webperf-loading:Priority-Hints-Audit.js (check fetchpriority)
    3. webperf-loading:Find-render-blocking-resources.js (competing resources)
  • If Render Delay > 200ms → Run:
    1. webperf-loading:Find-render-blocking-resources.js (blocking CSS/JS)
    2. webperf-loading:Script-Loading.js (parser-blocking scripts)
    3. webperf-interaction:Long-Animation-Frames.js (main thread blocking)

After LCP-Trail.js

  • If many LCP candidate changes (>3) → This causes visual instability, investigate:
    1. webperf-loading:Find-Above-The-Fold-Lazy-Loaded-Images.js (late-loading images)
    2. webperf-loading:Fonts-Preloaded-Loaded-and-used-above-the-fold.js (font swaps)
    3. CLS.js (layout shifts contributing to LCP changes)
  • If final LCP candidate appears late → Run webperf-loading:Resource-Hints-Validation.js
  • If early candidate was replaced → Understand why initial content was pushed down (likely CLS issue)

After LCP-Image-Entropy.js

  • If entropy is very high → Image is visually complex, recommend:
    • Modern formats (WebP, AVIF)
    • Appropriate compression
    • Potentially a placeholder strategy
  • If entropy is low → Image may be over-optimized or placeholder-like
  • If large file size detected → Run webperf-media:Image-Element-Audit.js for format/sizing analysis

After LCP-Video-Candidate.js

  • If video is LCP → Run:
    1. webperf-media:Video-Element-Audit.js (check poster, preload, formats)
    2. webperf-loading:Priority-Hints-Audit.js (check fetchpriority on poster)
    3. LCP-Subparts.js (analyze video loading phases)
  • If poster image is LCP → Treat as image LCP (run image workflows)

After CLS.js

  • If CLS > 0.1 → Run webperf-interaction:Layout-Shift-Loading-and-Interaction.js (pending — available in webperf-interaction skill)
  • If CLS > 0.25 (poor) → Run comprehensive shift investigation:
    1. webperf-loading:Find-Above-The-Fold-Lazy-Loaded-Images.js (images without dimensions)
    2. webperf-loading:Fonts-Preloaded-Loaded-and-used-above-the-fold.js (font loading strategy)
    3. webperf-loading:Critical-CSS-Detection.js (late-loading styles)
    4. webperf-media:Image-Element-Audit.js (missing width/height)
  • If CLS = 0 → Confirm with multiple page loads (might be timing-dependent)

After INP.js

  • If INP > 200ms → Run webperf-interaction:Interactions.js to identify slow interactions
  • If INP > 500ms (poor) → Run full INP debugging workflow:
    1. webperf-interaction:Interactions.js (list all interactions)
    2. webperf-interaction:Input-Latency-Breakdown.js (phase breakdown)
    3. webperf-interaction:Long-Animation-Frames.js (blocking frames)
    4. webperf-interaction:Long-Animation-Frames-Script-Attribution.js (culprit scripts)
  • If specific interaction type is slow (e.g., keyboard) → Focus analysis on that interaction type
  • After INP data is collected → call getINPDetails() for the 15 slowest interactions, slowest first (useful for identifying patterns across multiple slow interactions)

Error Recovery

  • If any script returns status: "error" → Check if the page has finished loading:
    • If early in load: wait and re-run the script
    • If page is an SPA: user may need to navigate to the target route first
  • If LCP.js / LCP-Subparts.js returns status: "error" → Tell the user: "LCP data is not available yet. Please ensure the page has fully loaded, then run the analysis again."
  • If INP.js getINP() returns status: "error" → The getDataFn: "getINP" field signals the agent can retry after user interaction. Prompt the user to click, type, or scroll, then call getINP() again.

Cross-Skill Triggers

Context fork note: This skill runs with context: fork. Cross-skill triggers below are recommendations to report back to the parent agent, not direct calls this subagent can execute. When a cross-skill trigger fires, tell the user which skill and script to run next. Scripts marked (pending) are not yet available — skip them and note the limitation.

These triggers recommend using snippets from other skills:

From LCP to Loading Skill

  • If LCP > 2.5s and TTFB phase is dominant → Use webperf-loading skill:
    • TTFB.js, TTFB-Sub-Parts.js, Service-Worker-Analysis.js
  • If LCP image is lazy-loaded → Use webperf-loading skill:
    • Find-Above-The-Fold-Lazy-Loaded-Images.js
  • If LCP has no fetchpriority → Use webperf-loading skill:
    • Priority-Hints-Audit.js

From CLS to Loading Skill

  • If CLS caused by fonts → Use webperf-loading skill:
    • Fonts-Preloaded-Loaded-and-used-above-the-fold.js
    • Resource-Hints-Validation.js (for font preload)
  • If CLS caused by images → Use webperf-media skill:
    • Image-Element-Audit.js (check for width/height attributes)

From INP to Interaction Skill

  • If INP > 200ms → Use webperf-interaction skill for full debugging:
    • Interactions.js, Input-Latency-Breakdown.js
    • Long-Animation-Frames.js, Long-Animation-Frames-Script-Attribution.js
    • LongTask.js (if pre-interaction blocking suspected)

From LCP/INP to Interaction Skill

  • If render delay or interaction delay is high → Use webperf-interaction skill:
    • Long-Animation-Frames.js (main thread blocking)
    • LongTask.js (long tasks delaying rendering)

Multi-Metric Correlation

When multiple CWV metrics are poor, prioritize investigation:

  • If LCP > 2.5s AND CLS > 0.1 → Likely shared root cause:
    1. Check for late-loading content pushing LCP element
    2. Run LCP-Trail.js to see LCP candidate changes
    3. Run Layout-Shift-Loading-and-Interaction.js to correlate timing
  • If LCP > 2.5s AND INP > 200ms → Main thread congestion:
    1. Run Long-Animation-Frames.js
    2. Run webperf-loading:Script-Loading.js
    3. Run webperf-loading:JS-Execution-Time-Breakdown.js
  • If CLS > 0.1 AND INP > 200ms → Layout thrashing or interaction-triggered shifts:
    1. Run Layout-Shift-Loading-and-Interaction.js
    2. Run Interactions.js
    3. Check if shifts occur during/after interactions

References

  • references/snippets.md — Descriptions and thresholds for each script
  • references/schema.md — Return value schema for interpreting script output

Download now (ZIP)

Trending

More talk lately.about the whole repository

Dec 2025: 0 mentionsJan 2026: 0 mentionsFeb 2026: 0 mentionsMar 2026: 0 mentionsApr 2026: 0 mentionsMay 2026: 0 mentionsJun 2026: 0 mentionsJul 2026: 0 mentionsAug 2026: 0 mentionsSep 2026: 1 mentionOct 2026: 16 mentionsNov 2026: 0 mentions
Dec 2025Jun 2026Nov 2026
Mentions a month of the nucliweb/webperf-snippets repository that our weekly crawl found. There is no adoption data for this file yet.

Who's talking about nucliweb/webperf-snippets

The biggest 5 of 24 mentions linking to the repository that our crawl found. Each line is their own headline.

See all 24

Talked about on GitHub, Bluesky and Hacker News (24 times).

On GitHub · 19

14 more on GitHub

On Hacker News · 2

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

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.