seo-crawl-audit 0.16.1 → 0.17.0

package/README.md

@@ -1,14 +1,42 @@
 # SEO Crawl Audit
 
-Free, open-source, local-first SEO crawler and regression audit tool.
+**Find technical SEO issues. Keep the evidence. Catch regressions.**
 
-`seo-audit` crawls server-rendered pages, saves a versioned JSON snapshot,
-finds concrete SEO problems, and detects regressions after a deployment. It
-needs no account or API key. Page HTML is processed in memory; normalized
-results stay in local JSON, checkpoint, CSV, and self-contained HTML files.
+A free, open-source, local-first crawler with HTML reports, versioned JSON
+snapshots and eight MCP tools for Codex, Claude Code and Claude Desktop. Plan a
+bounded crawl, inspect concrete URL findings and compare runs after a deployment.
+No SEO service account or SEO API key is required. Scans default to 100 pages
+and respect robots.txt.
 
-[Open the interactive example report](https://melnixdev.github.io/seo-crawl-audit/)
-or [view its source file](https://github.com/MelnixDev/seo-crawl-audit/blob/main/examples/quotes-toscrape-report.html).
+[Explore the example report](https://melnixdev.github.io/seo-crawl-audit/) ·
+[Choose your installation](https://github.com/MelnixDev/seo-crawl-audit/blob/main/docs/native-plugins.md) ·
+[CLI quick start](#requirements-and-installation)
+
+## Choose your client
+
+| Client | Package and workspace |
+| --- | --- |
+| **Codex** | ZIP plugin and local marketplace. Configure one explicit, non-overlapping audit workspace before importing. Calls must name that workspace. [Setup](https://github.com/MelnixDev/seo-crawl-audit/blob/main/docs/native-plugins.md#codex-desktop-configure-then-install) |
+| **Claude Code** | Same ZIP, loaded from outside the audited project. Uses Claude Code’s project directory. [Setup](https://github.com/MelnixDev/seo-crawl-audit/blob/main/docs/native-plugins.md#claude-code) |
+| **Claude Desktop** | MCPB extension with a required existing-directory picker. [Setup](https://github.com/MelnixDev/seo-crawl-audit/blob/main/docs/native-plugins.md#claude-desktop-chat) |
+
+Native packages are available with **0.17.0**. See the
+[release notes](https://github.com/MelnixDev/seo-crawl-audit/blob/main/docs/releases/0.17.0.md) for exact verification and limitations.
+No public directory listing or client endorsement is implied. The CLI remains
+available through npm; its [quick start](#requirements-and-installation) is below.
+
+The crawler sends HTTP requests to your chosen site and keeps artifacts in your
+configured workspace. Your AI client receives tool results under its own data
+policies. A crawl measures observed pages; it does not establish Google indexing,
+rankings, traffic or complete site coverage.
+
+### A real bundled-MCP report
+
+![Report produced by the bundled MCP server from a three-page local fixture](https://raw.githubusercontent.com/MelnixDev/seo-crawl-audit/main/docs/site/media/native-report-pages.png)
+
+This report was generated through the plugin’s bundled STDIO MCP server against
+a deliberately small local fixture. It demonstrates actual report output, not an
+installed AI client interface. [Watch the report walkthrough](https://github.com/MelnixDev/seo-crawl-audit/blob/main/docs/site/media/native-workflow.mp4).
 
 ## Report preview
 
@@ -579,7 +607,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
-      - uses: MelnixDev/seo-crawl-audit@v0
+      - uses: MelnixDev/seo-crawl-audit@v0.17.0
         id: seo
         with:
           url: https://preview.example.com/

package/agent-skill/seo-crawl-audit/SKILL.md

@@ -9,7 +9,7 @@ Use this skill when a user asks to audit a website, check SEO regressions, inspe
 
 ## Workflow
 
-1. Start with `seo_audit_plan` when using MCP. Confirm the URL, sitemap mode, page limit, and robots policy before requesting pages.
+1. Read the bound workspace in MCP initialization and ensure it matches the user's current project. If tools declare a required `workspace` argument, pass the current project's absolute real path; stop on a mismatch and reconfigure the plugin rather than scanning another project. Start with `seo_audit_plan` when using MCP. Confirm the URL, sitemap mode, page limit, and robots policy before requesting pages.
 2. Run `seo_audit_scan` for a baseline, or `seo_audit_check` with a saved baseline for a regression comparison. Keep the default 100-page limit unless the user explicitly asks for more.
 3. Use `seo_audit_compare` for two saved snapshots, `seo_audit_history` for a compact read-only trend over saved runs, or `seo_audit_issues` with filters and pagination to inspect concrete findings. Prioritize errors, then warnings, then informational context. Never treat a lower count across an incomplete or policy-changed trend as proven improvement.
 4. Use `seo_audit_report` when a human-readable HTML artifact is needed. Return artifact paths rather than embedding a large report in chat.