seo-qa
Run NodeRoom's repeatable SEO and landing-flow QA workflow. Use when auditing or improving public SEO pages, metadata, sitemap/robots, Search Console metrics, Playwright landing journeys, Lighthouse/Core Web Vitals evidence, Gemini video judging, or Nebius-first text/JSON judging without scraping Google rankings.
适合你,如果需要系统化检查网站SEO和着陆页表现
/ 通过 npx 安装 校验哈希
npx oh-my-skill add homenshum/noderoom/seo-qa/ 通过 bash 安装
curl -fsSL https://oh-my-skill.com/install.sh | bash -s -- homenshum/noderoom/seo-qa/ 已经装过?验证本机副本,不用重装
npx oh-my-skill verify homenshum/noderoom/seo-qa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。
2GitHub stars
~718最小装载
~9.5K含声明引用
~9.5K文本包总量
索引托管
怎么用
技能原文 SKILL.md
NodeRoom SEO QA
Workflow
- Confirm the target surface.
- Public SEO pages are crawlable routes such as
/,/use-cases/,/compare/...,/learn/,/pricing/, and/faq/. - Private or app surfaces include room/demo/create URLs, hash-app routes, live rooms, backend state, and benchmark fixtures.
- Run the static audit:
npm run seo:audit
- Run browser journeys when UI behavior is in scope:
PLAYWRIGHT_RECORD_VIDEO=1 npm run test:journeys
- Compress a selected video before model judging:
- Run the local performance budget check:
npm run seo:perf
- Compress a selected video before model judging:
npm run seo:compress-video -- --input <video.webm> --mode review
- Judge the compressed video with Gemini when
GOOGLE_GENERATIVE_AI_API_KEYis available:
npm run seo:judge-video -- --input <video.mp4>
- Pull Search Console data only when credentials are available:
npm run seo:search-console
- Update
docs/seo/SEO_AUDIT.mdordocs/seo/JOURNEY_QA_REPORT.mdwith artifact links, before/after findings, and deferred fixes.
Hard Rules
- Do not automate mass Google searches, fake clicks, ranking scrapes, or link spam.
- Use Search Console for query, impression, CTR, and average-position data.
- Keep public pages crawlable and fast.
- Keep room/demo/create/app-scoped URLs out of the sitemap and protected with noindex behavior.
- Do not print or commit API keys.
- Do not rewrite backend, auth, realtime, or room engine behavior for an SEO change unless the task explicitly requires a compatibility fix.
- Structured data must match visible page content.
Scripts
scripts/audit-static.ts: check metadata, sitemap, robots, route files, headings, canonicals, and private-route indexing guards.scripts/search-console-report.ts: query Google Search Console using an access token or service-account JSON and write JSON/Markdown reports.scripts/performance-check.ts: collect lab Core Web Vitals-style metrics and write a performance report.scripts/compress-video.ts: create readable or low-cost MP4 review copies with ffmpeg.scripts/judge-video-gemini.ts: produce structured Gemini visual QA JSON and Markdown summaries from a video.
Browser QA Notes
- Use Playwright for repeatable journey tests, screenshots, traces, and videos.
- Use Chrome DevTools or Lighthouse separately for performance traces and Core Web Vitals evidence.
- The Google-origin journey is opt-in with
SEO_QA_ALLOW_GOOGLE_ORIGIN=1and performs one target phrase only; default CI should use direct landing URL fallback.
Model Policy
- Use Nebius first for text/JSON SEO judging, landing copy variants, app-side reasoning, fast chat, and code reasoning.
- Use Gemini first for video understanding.
- Use fallbacks only when the primary provider is unavailable or lacks the required capability.
- Record provider, model, routing reason, and missing-key blockers in the report without logging secret values.
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。
…