‹ 首页

seo-core-web-vitals

@hainrixz · 收录于 1 周前

Audit Core Web Vitals & page performance — measure LCP, INP, and CLS against p75 field thresholds, diagnose render-blocking resources, unoptimized images, and layout-shift sources, and produce prioritized, advisory-only remediation guidance. Module M15. Feeds the Search SEO score (heavily) and the AI Visibility score (minimally).

场景网站性能seo优化页面加载

适合你，如果关心网站速度和搜索排名

/ 下载安装
seo-core-web-vitals.skill双击,或拖进 Claude 桌面版 / Cowork,即完成安装↓ .skill↓ .zip
用别的 agent?下载 .zip 解压,把文件夹放进它的技能目录
Claude Code~/.claude/skills/(项目级 .claude/skills/)
Codex CLI~/.codex/skills/
Cursor自动读取上面两处目录
其他工具见其文档的「skills」目录;两个下载是同一份文件,只是名字不同
/ 通过 npx 安装 校验哈希
npx oh-my-skill add hainrixz/claude-seo-ai/seo-core-web-vitals
/ 通过 bash 安装
curl -fsSL https://oh-my-skill.com/install.sh | bash -s -- hainrixz/claude-seo-ai/seo-core-web-vitals
/ 已经装过?验证本机副本,不用重装
npx oh-my-skill verify hainrixz/claude-seo-ai/seo-core-web-vitals
安装目标可用 --agent / --scope 或 --to 明确指定;省略时只会在唯一已存在的 agent 目录上自动选择,零命中或多命中会停止并提示。content_hash 缺失或不一致均拒装。
27GitHub stars
~1K上下文体积 · 单文件
镜像托管

怎么用

技能原文 SKILL.md作者撰写 · MIT · 788e7b4

seo-core-web-vitals (M15)

Core Web Vitals are a confirmed Google ranking signal and a proxy for page quality. Thresholds, field-vs-lab honesty rules, and LCP decomposition: references/cwv-thresholds.md (follow it exactly).

Audits

Working from the PageSnapshot (rendered_dom if present, else raw_html):

  1. Field CWV at p75 — LCP ≤ 2.5 s, INP ≤ 200 ms (INP replaced FID), CLS ≤ 0.1, all at the 75th percentile of real-user data. All three must pass for a "good" rating. Use the proactive-warning thresholds (LCP > 2.0 s, INP > 160 ms, CLS > 0.08) to warn before failing.
  2. Render-blocking resources — synchronous <script> in <head> (no defer/async/type=module), blocking <link rel="stylesheet">, and @import chains that delay first render (LCP/INP risk).
  3. Unoptimized images — missing width/height (or aspect-ratio), no loading="lazy" below the fold, no responsive srcset/sizes, legacy formats where AVIF/WebP would serve, and a non-preloaded LCP image.
  4. Layout-shift sources — images/iframes/ads/embeds without reserved dimensions, web-font swap without font-display/size-adjust, and content injected above existing content (CLS risk).
  5. LCP decomposition — attribute LCP to TTFB vs resource load delay vs load duration vs render delay (see references/cwv-thresholds.md) so each finding targets the dominant subpart.
Fixes (fixable: advisory)

Advisory only. Performance fixes touch build config, server, CDN, and runtime JavaScript — high breakage risk — so the tool diagnoses and prioritizes but never auto-edits performance code. For each failing metric, name the specific resource/element causing it and rank fixes by expected leverage (e.g. preload the LCP image, defer non-critical JS, reserve image dimensions). Give framework-specific direction where the stack is known. Never fabricate measured values — if field data is absent, ask the user to supply a PSI/CrUX key or leave a clearly-marked TODO placeholder rather than guessing a metric.

Verification
  • Field (Tier 1): node ${CLAUDE_SKILL_DIR}/../../scripts/psi-client.mjs --url <u> calls the PageSpeed Insights API (loadingExperience, method psi_api); CrUX History via the same client (crux_api) for p75 trends. Requires a free API key.
  • Tier 0 (no key): emit status: needs_api for every field-CWV finding, plus clearly-labeled lab heuristics (render-blocking count, missing image dimensions, bundle weight). Label lab signals "lab data — not what Google ranks on"; never present a heuristic as a measured field value.
  • When the required data tier is unavailable, status is needs_api — never a false pass.
Findings

Emit findings per schema/finding.schema.json. Examples:

  • M15.lcp.exceeds_p75 — field LCP > 2.5 s at p75 (status fail, severity 4, fixable: advisory, axis search, confidence established). evidence.observed quotes the measured p75 value and the LCP element; verification.reproduce runs the psi-client command.
  • M15.cls.unsized_image — image with no width/height shifting layout (status warn, severity 4, fixable: advisory, axis search, confidence directional). evidence.observed quotes the <img> tag.
  • M15.field.needs_api — no PSI/CrUX key, field CWV unverifiable (status needs_api, severity 4, fixable: advisory, axis search, confidence established).

Each finding: evidence.observed quotes the page/measurement; verification.reproduce is the runnable command; expected_impact is banded + confidence-tagged (no naked %).

Honesty
  • Lab ≠ field. A single local Lighthouse run is diagnostic only; Google ranks on field (CrUX) p75. Never let a lab number drive the Search score or masquerade as a field result.
  • CWV is a real but modest tie-breaker signal — it does not override relevance/content quality. Don't promise ranking jumps from a green score; report it as banded expected_impact, not a predicted percentage gain.
  • A "100" lab performance score is not a pass if field p75 fails; only field data settles a CWV finding.
按 MIT 许可原样转载,未经改动 · 在 GitHub 查看 →

评论

登录即可评论;带「已验证安装」的,是发布者名下有本店的安装或持有记录。
…