القائمة
Run Google PageSpeed Insights against every post URL plus the homepage, capturing Core Web Vitals (LCP, CLS, INP) and failing audit IDs for both mobile and desktop.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Detect winners, decaying posts, striking-distance keywords, and surprise wins. Computes week-over-week deltas from GSC + Fathom + Kit snapshots across the current and previous weeks.
Orchestrator. Runs all 6 fetchers and 3 analyzers in dependency order, assembles the full markdown report, creates a Google Doc in a configured Drive folder, persists everything to the snapshot store. Returns the doc URL.
Pull 90-day pageviews, uniques, top referrers, and tracked events from Fathom Analytics, aggregated by ISO week and by page path.
Pull primary-tag audience size, weekly new-signup history, and per-post attribution counts from Kit (ConvertKit) v4. CC4M stores subscribers by tag, not form, so this skill is tag-driven.
Surface technical, indexation, and on-page SEO issues as a prioritized fix list. Reads all fetcher snapshots; produces a sorted list of issues with severity, effort estimate, and recommended action.
Generate the punchy Sunday-morning email body from the week's insights, send via Resend with a link to the full Google Doc. The thing you'll actually read.
| name | fetch-pagespeed |
| description | Run Google PageSpeed Insights against every post URL plus the homepage, capturing Core Web Vitals (LCP, CLS, INP) and failing audit IDs for both mobile and desktop. |
Runs PSI per URL × strategy (mobile + desktop). Writes reports/{week}/pagespeed.json.
python .claude/skills/fetch-pagespeed/scripts/fetch.py [--week 2026-W18]
PSI_API_KEY — free, generous quota (25K queries/day)posts.jsonreports/{week}/pagespeed.json keyed by URL, with mobile and desktop sub-objects each containing performance, lcp, cls, inp, and an issues list of failing audit IDs.
unused-javascript), not full audit text. The analyzer maps IDs to human-readable issues.