Update to Sitewalk: SEO Audit & Crawler

Changed files

Choose a file to see its text changes. Binary or large files show metadata only.

plugin.json →

skills/get-started/SKILL.md →

plugin.json

− removed · + added

--- before
+++ after
@@ -1,7 +1,7 @@
 {
   "$schema": "https://agent-plugins.org/schemas/1.0.0/plugin.schema.json",
   "name": "sitewalk",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "description": "Crawl a website and get its technical SEO issues ranked by impact, with a plain fix for each.",
   "author": {
     "name": "Aditya Sharma",
@@ -23,22 +23,46 @@
     "sitemap",
     "website audit",
     "seo checker",
-    "hreflang"
+    "hreflang",
+    "site monitoring",
+    "ai search readiness",
+    "website audit tool",
+    "seo analyzer",
+    "seo report",
+    "broken link checker",
+    "core web vitals",
+    "page speed",
+    "structured data",
+    "schema checker",
+    "ai seo",
+    "llms.txt"
   ],
   "extensions": {
     "com.openai": {
       "interface": {
         "displayName": "Sitewalk: SEO Audit & Crawler",
         "shortDescription": "Crawl a site, fix SEO issues",
-        "longDescription": "Sitewalk crawls a website the way a search engine bot does and brings back its technical SEO problems ranked by impact. Each issue comes with the number of pages affected, example URLs and a fix written in plain English, so you can act on it yourself or hand it to a developer.\n\nWhat it checks:\n- Response codes: broken pages, server errors, redirects and redirect chains.\n- Titles and meta descriptions: missing, too long, too short or duplicated.\n- Headings and content: missing H1 and thin pages.\n- Indexability: noindex pages, soft 404s, canonical tags that are missing, chained or point at redirects or broken pages.\n- Links: broken internal links, links to localhost or private addresses, vague link text.\n- Images: missing alt text, broken images, missing width and height, insecure images on HTTPS pages.\n- URLs: spaces, double slashes, underscores, repeated segments, session ids, and calendar or filter traps.\n- Structure: pages buried deep in the site and pages nothing links to.\n- International: hreflang codes and return links.\n- Mobile, social and markup: viewport, language attribute, Open Graph and Twitter tags, structured data presence.\n- Security: pages still on HTTP and basic security headers on the home page.\n- Sitemap: sitemap URLs that redirect, error, are set to noindex or point elsewhere with a canonical.\n\nEvery report also says what was not checked (for example page speed, backlinks and rankings) and notes when the page limit was reached, so nothing is implied that the crawl did not see.\n\nSign in with your Sitewalk account to run an audit. A site you have not verified is crawled up to 50 pages; verify a site you own and it crawls up to 500 pages. The full report is also available on the web with CSV and PDF downloads. The crawler identifies itself as SitewalkBot, obeys robots.txt and keeps a slow, polite pace. It does not render JavaScript, so pages that build their content in the browser may show fewer issues than a visitor would see. Powered by LibreCrawl.",
+        "longDescription": "Sitewalk for ChatGPT crawls a website the way a search engine bot does and runs more than 200 checks in 18 groups. It brings back the technical SEO problems, ranked by how much they matter, with a site health score out of 100. Each issue comes with the number of pages affected, example addresses, why it matters and a fix in plain English, so you can act on it yourself or hand it to a developer.\n\nWho it is for: site owners, freelancers, agencies and developers who want a clear list of what to fix, with no SEO tool to learn.\n\nWhat it checks, in the main groups:\nResponse codes: broken pages, server errors, redirects and redirect chains.\nTitles and meta descriptions: missing, too long, too short or duplicated.\nHeadings and content: missing H1, thin pages, and pages whose text only appears after JavaScript runs.\nIndexability: noindex pages, soft 404s, and canonical tags that are missing, chained, or point at redirects or broken pages.\nLinks and structure: broken links, vague link text, links with no text or label, links with no usable address, pages nothing links to, dead-end pages and pages buried deep in the site.\nImages: missing alt text, broken images, missing width and height.\nURLs: spaces, double slashes, underscores, capital letters, addresses over 115 characters, repeated segments, session ids, and calendar or filter traps.\nInternational: hreflang codes and return links.\nPage basics: HTML sent uncompressed, HTML over Googlebot's 2 MB limit, and pages that do not declare their character encoding.\nMobile, social and markup: viewport, language attribute, Open Graph and Twitter tags.\nStructured data: markup that fails to read, is missing a required property, or is repeated or inconsistent across pages of the same kind.\nSecurity: pages still on HTTP, basic security headers, an expired or soon-to-expire certificate, and whether http, https, www and non-www all lead to one address.\nAI search readiness: whether robots.txt and headers let ChatGPT search and other AI bots in, your llms.txt file, whether long pages are easy for AI tools to read, and how many of these checks pass.\nSitemap: sitemap addresses that redirect, error, are set to noindex or point elsewhere with a canonical.\nPage speed: a mobile Lighthouse score for up to 3 pages in each audit, plus Google's real-user Core Web Vitals data where Google has it, always labelled a sample.\n\nEvery report ends with Sitewalk's read: the three fixes to do first and the steps for this week. It says when the page limit was reached and when a check could not run on your site, so nothing is implied that the crawl did not see. After you fix the top three, ask for another audit and Sitewalk shows how the health score moved. Ask Sitewalk to keep watching a site you have audited: it re-audits it every month and emails you only when something changes.\n\nHow it works: ask for an audit in chat. The first time, Sitewalk asks you to connect: it emails you a 6-digit code and you type it in, with no password. Without a confirmed email an audit covers up to 50 pages; once you are signed in with your confirmed email it covers up to 500. No domain verification is needed. The full report is on the web with CSV and PDF downloads.\n\nLimits: Sitewalk does not measure backlinks, keyword rankings, Google Search Console or analytics data, or writing. The crawler identifies itself as SitewalkBot, obeys robots.txt and keeps a slow, polite pace of at least one second between requests, so a small site usually takes a few minutes and a 500-page audit 20 to 60 minutes. It reads each page as the server delivers it and flags pages that look empty without scripts. Powered by LibreCrawl.",
         "developerName": "Aditya Sharma",
         "category": "Productivity",
         "capabilities": [
-          "Crawl a website and start a technical SEO audit",
-          "Track audit progress page by page",
-          "Rank issues by impact with a plain-English fix",
-          "Show example URLs and pages affected per issue",
-          "List what was checked and what was not"
+          "Technical SEO site audit",
+          "More than 200 checks in 18 groups, ranked worst first",
+          "Site health score out of 100 with a plain-English fix list",
+          "Broken links, redirect chains and server errors",
+          "Missing, long, short and duplicate titles and descriptions",
+          "Canonical, noindex and hreflang problems",
+          "Dead-end pages and pages with few internal links",
+          "Structured data (schema) checks",
+          "AI search readiness: robots.txt, llms.txt and AI bot access",
+          "Pages that need JavaScript to show their content",
+          "Expired SSL certificate and www or http redirect checks",
+          "HTML pages over Googlebot's 2 MB limit",
+          "Sitemap checks",
+          "Page speed sample: mobile Lighthouse score, 3 pages",
+          "Image alt text and broken images",
+          "Health trend across repeat audits",
+          "Monthly monitoring of a site you have audited, with an email only when something changes"
         ],
         "websiteURL": "https://audit.adityaarsharma.com",
         "supportURL": "https://audit.adityaarsharma.com/bot/",
@@ -46,14 +70,13 @@
         "termsOfServiceURL": "https://audit.adityaarsharma.com/terms",
         "defaultPrompt": [
           "Audit my website for technical SEO problems and tell me what to fix first.",
-          "Find broken links, redirect chains and duplicate titles on my site.",
-          "Turn my latest Sitewalk report into a task list for my developer."
+          "Is my site ready for AI search? Check what AI crawlers can read and what is missing.",
+          "Find broken links, redirect chains and duplicate titles on my site."
         ],
         "brandColor": "#E11D1D",
         "brandColorDark": "#FF6B6B",
         "composerIcon": "./assets/logo.svg",
         "logo": "./assets/logo.png",
-        "screenshots": [],
         "composerIconDark": "./assets/logo-dark.svg",
         "logoDark": "./assets/logo-dark.png"
       },
@@ -117,7 +140,7 @@
         "commerce": false
       },
       "publication": {
-        "release_notes": "Clearer name and description, a full list of what each audit checks, a report section that states what was not checked, better starter prompts, and a refreshed icon with a dark version."
+        "release_notes": "Description now matches the product: more than 200 checks in 18 groups, AI search readiness, structured data, a page speed sample, and monthly monitoring of a site you have audited. Corrected the page rule (50 pages without a confirmed email, 500 with one, no domain verification needed) and replaced a starter prompt that needed an earlier report."
       },
       "apps": "./.app.json"
     }