agentsclimarketplace

Site capture audit

Skill sujanbhuiyan/Skills/skills/site-capture-audit

A curated, installable marketplace of production-grade Agent Skills for Claude — research, SEO/GEO, finance, deliverables & engineering, indexed by topic, lifecycle & domain.

Install
npx -y skills add sujanbhuiyan/Skills --skill site-capture-audit

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Captures a website (raw HTML), cleans pages to markdown, pulls sitemaps and robots.txt, then produces a prioritized 90-day quick-win SEO audit. Use when onboarding a new site, baselining technical SEO and on-page health, or building a ranked P0/P1/P2 fix list across indexability, on-page, content, and conversion.

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

3.7 KB, as published. Nobody here has run it

Site Capture & Audit

What this does

Captures a website to local files, then turns that capture into a prioritized, 90-day quick-win audit. The capture step saves raw HTML, a cleaned text/markdown version of each page, and the site's sitemap.xml + robots.txt. The audit step grades the site against a P0/P1/P2 rubric covering technical SEO (indexability, canonicals, redirects), on-page (titles, meta, headings, schema, internal links), content depth, Core Web Vitals, and conversion — and emits a ranked fix list with effort/impact.

When to use it

  • Onboarding a new client/site and you need a fast, evidence-based baseline.
  • Auditing technical + on-page SEO before a redesign, migration, or content push.
  • Producing a defensible "first 90 days" quick-win list for a stakeholder.

How to use it

  1. Capture the site with the bundled script: python scripts/capture.py https://example.com --max 25 --out ./capture It fetches the homepage, discovers more URLs via sitemap.xml / robots.txt, saves raw HTML to <out>/raw-html/, cleaned markdown to <out>/clean-md/, and <out>/sitemaps/ (sitemaps + robots.txt). See the script --help.
  2. Read the cleaned pages in <out>/clean-md/ and skim the raw HTML <head> blocks for titles, meta, canonical, robots directives, and JSON-LD.
  3. Work through references/audit-checklist.md in priority order — P0 indexability first, then P1 on-page, then P2 depth + CWV + conversion. Record evidence (the actual title, the missing canonical, the redirect chain) for each finding.
  4. For each finding assign: severity (P0/P1/P2), impact (high/med/low), effort (S/M/L), and the concrete fix.
  5. Write the audit as a ranked table grouped by priority, with a short executive summary and a 90-day sequencing (do P0 in weeks 1–2, etc.).
  6. For schema gaps found in the audit, hand off to a schema-org-sprint skill; for content gaps, hand off to a content-brief skill.

Inputs

  • A base URL (the site to capture and audit).
  • Optional: a max page count (--max) and output dir (--out).
  • Optional: access to Google Search Console / analytics to ground impact (the audit works from the capture alone, but real query/traffic data sharpens it).

Output

  • A capture folder: raw-html/, clean-md/, sitemaps/.
  • A prioritized 90-day quick-win audit (markdown): executive summary, ranked P0/P1/P2 findings with evidence + fix + effort/impact, and a week-by-week plan.

Notes & constraints

  • The capture script uses only the Python standard library (urllib) plus optional BeautifulSoup; if BeautifulSoup is absent it falls back to a regex tag-strip, so it runs on a stock Python 3. It sets a descriptive User-Agent and adds a small delay between requests to be polite.
  • It only fetches when you invoke it with a URL. Respect robots.txt, rate limits, and the site owner's terms; capture sites you're authorized to.
  • The audit is an expert review against the rubric, not a substitute for live Core Web Vitals field data or full crawl tooling — call those out where they'd change a conclusion.
  • Use forward-slash paths.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.