More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    seo

    1

    This skill starts by measuring, then fixes only what the evidence points at, and says plainly what cannot be fixed.

    $19

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    seo

    seo

    Example session with this skill installed

    Our Search Console clicks are down about 8% over the last 28 days, and our main page, /invoice-generator, lost about 590 clicks (-18%). I've attached the Performance export with the comparison and a 3-month Dates export for that page. What happened, and what should we change?

    • Read your context and instructions
    • Compiled the seo

    I ran gsc_triage.py on both exports and checked the page live with audit_page.py.

    Answer in three sentences

    About 60% of the page's fall is the end of a seasonal peak in late September, which no page change can fix. The rest is real: CTR fell on "free" and "online" queries at an unchanged position, and the page lost ranking in India and Japan. One change is worth making now: the title and description.

    The numbers (observed)

    previous 28 dayslast 28 dayschange
    clicks3,3002,710-18%
    impressions27,80026,300-5%
    CTR11.9%10.3%-1.6 pts
    position5.15.4+0.3

    Weekly clicks were about 735 in August, peaked at 942 in the week of 24 September and settled near 670 by late October. Against the four weeks before the peak, the page is down 8%, not 18%.

    Where the real loss is

    CTR drop at the same position (fixable): "invoice generator free" 15.0% to 10.1% at position 2.9 to 3.0; "free invoice generator online" 12.7% to 8.3%. Competitors' titles promise "free" and "online"; this page's title does not.

    Ranking loss: India 5.4 to 7.3 and Japan 4.8 to 6.5; "gst invoice generator" 5.0 to 7.4. Check which URL ranks there now before changing anything.

    • Demand drop (not fixable): "invoice template" impressions -28% at position 4.4.

    Checked and ruled out

    200, indexable, self-canonical, hreflang includes itself; no manual action or security issue in Search Console.

    Change (one at a time)

    changepagemeasureread on
    Title: "Free Invoice Generator Online: PDF in 1 Minute | Tallyfox"; description states free, no sign-up, PDF/invoice-generatorCTR of the four "free/online" queries at unchanged position4 weeks after deploy

    Not recommended: llms.txt, FAQ schema or rewriting the page; none of them addresses a snippet problem.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Most SEO advice starts by changing pages. This skill starts by measuring, then fixes only what the evidence points at, and says plainly what cannot be fixed. It works in Claude Code, Codex and any agent that reads SKILL.md, and it was distilled from running search for a real product.

    What it does

    • Diagnoses traffic drops. Splits a Search Console comparison into demand, CTR and ranking effects, checks whether the "previous period" was a seasonal peak, and locates the loss (one page, a group, the whole site) before anyone rewrites anything.
    • Audits a site. Robots.txt with Google's real group rules, sitemaps and lastmod trust, a link crawl with hreflang reciprocity, soft 404s, http/https and www variants, canonicals, structured-data required fields, JavaScript rendering, Core Web Vitals field data, and hacked-site signals. Findings come with evidence, severity, the fix and a "not a problem" list.
    • Finds where to grow. Striking-distance queries, snippets that underperform for their position on your own site, question and modifier queries, and cannibalisation.
    • Plans pages that can rank. Intent from the live results, content briefs, and a gate for programmatic pages that keeps you clear of Google's scaled-content and doorway policies.
    • Implements fixes in code, with a Next.js reference for both routers, then verifies the live page.
    • Pulls live data when you want it. With Agensi's data tools connected, it can fetch live results pages, search volumes, keyword ideas, competitor keywords, the local pack and Google News. Each call is pay per use, priced up front and run only after you approve it. Without them, every workflow has a free fallback.

    12 scripts, standard library only

    gsc_triage.py (why traffic fell), gsc_opportunities.py (where to grow), audit_page.py (a page as crawlers see it, with required schema fields), crawl_site.py, robots_check.py, check_sitemap.py, site_checks.py (soft 404s, host variants, hacked-site signals), render_check.py (raw HTML versus headless Chrome), cwv_check.py (CrUX field data), template_similarity.py, verify_crawlers.py (real Googlebot and Bingbot in your logs) and links_audit.py (backlinks without "toxic scores").

    16 references with dated evidence

    Drop diagnosis, technical, on-page and CTR, structured data, content and programmatic pages, keyword research, international and hreflang, AI search, local SEO and Business Profile, links and disavow, e-commerce and merchant listings, Discover, News, images and video, Next.js, data access, report templates. Every rule is graded documented, observed, study or inference, with the source and the date it was checked (November 2026).

    What it will not do

    Invent ranking-factor weights or a health score, recommend llms.txt or FAQ schema as a fix, promise positions, or change Search Console settings, robots.txt or redirects in production without your go-ahead.


    New to Agensi? Create your account through my referral link to get this skill. Disclosure: I earn a commission on skills you buy in your first 6 months; it costs you nothing extra.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Marketing & SEO.

    No keyword volumes, rank tracking or backlink index of its own: it uses Search Console, Bing Webmaster Tools and your exports, plus Agensi's pay-per-use data tools or other paid tools only if you connect them and approve each call. The crawler suits small and medium sites (thousands of pages); use a desktop crawler beyond that. It will not promise rankings or change production robots.txt, redirects or Search Console settings without your go-ahead. Evidence was checked in November 2026 and search changes often, so the skill re-checks sources older than six months.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 8 days ago

    Accounts
    Google AccountBingApple
    Software
    Python 3.8+ (standard library only)
    Google Search Console access
    oran export (recommended)
    Dataforseo MCPNode.js
    Optional
    Chrome, Chromium
    orMicrosoft Edge
    CrUX API key

    Permissions

    Terminal / Shell
    Network Access
    Read Files
    Write Files
    Browser

    Allowed Hosts

    chromeuxreport.googleapis.com
    www.googleapis.com
    developers.google.com
    www.bing.com
    search.developer.apple.com

    Shell: runs the bundled Python scripts and your project's typecheck, lint and build. Read files: Search Console exports, logs, link exports and your site's code. Write files: reports in your repo and SEO fixes in your code when you ask. Browser: reads Search Console and live results pages in your signed-in browser; it never types passwords or solves CAPTCHAs. Network: plain HTTPS requests to the site you audit (pages, robots.txt, sitemaps), the CrUX and PageSpeed Insights APIs, and Google's, Bing's and Apple's published crawler IP lists; every endpoint is listed in SKILL.md. Environment: none read; the optional CrUX API key is passed as a flag.

    Listed8 days ago

    What's inside

    Frequently Asked Questions

    It diagnoses before it recommends. It separates seasonality, CTR and ranking in your own Search Console data, checks the live pages as crawlers see them, and grades every claim by evidence. It never outputs a made-up health score or ranking-factor weights.

    No. It runs on Search Console, Bing Webmaster Tools, your logs and its own scripts. For data Search Console cannot give (live results pages, search volumes, competitor keywords, the local pack, Google News), it can use Agensi's data tools: priced per call, quoted first and run only after you approve. Ahrefs, Semrush or DataForSEO work too if you already pay for them.

    Yes, as part of SEO: how each engine retrieves pages, which crawlers to allow, JavaScript rendering, and what the evidence says does and does not work. For a full AI-visibility program, the companion ai-search-optimization skill goes further.

    Yes, when you ask. It finds where titles, canonicals, hreflang and structured data are generated (with a Next.js reference for both routers), makes the change, runs your checks and verifies the page after deploy.

    Yes. It has references for Google Business Profile and local results, links and disavow (including a backlink-export audit), and e-commerce markup, Merchant Center, variants and faceted navigation.

    Ideally a Search Console export or access, and the site URL. It asks before any outward action such as editing robots.txt, redirects or Search Console settings.

    New in Marketing & SEO