whywontai-scan-and-fix/SKILL.md
Setup
- MCP server URL:
https://mcp.whywontai.com/ check_robotsandcheck_agent_listingswork without sign-in. The other tools use the user's whywontai.com account through OAuth.- Claude Code:
claude mcp add --transport http whywontai https://mcp.whywontai.com/ - Cursor and VS Code: add the URL as a remote MCP server in the editor's MCP settings. See whywontai.com/integrations for the exact steps.
Tools
check_robots: read robots.txt rules for AI and search crawlers.scan_site: start a full scan. It returns ascan_id. A scan takes 1 to 5 minutes and counts against the plan quota.get_scan: status and summary. Passwait_seconds(up to 25) and call again until the status is complete.list_issues: issues, most severe first. Filter bymin_severityorcategory.get_internal_links,get_long_questions,get_page_two,get_index_status,get_bing_ai_citations: data that needs Search Console or Bing connected.
Workflow
- Ask for the site URL if the user did not give it.
- Run
check_robotsfirst. If a wanted crawler is blocked, report it before the full scan. - Run
scan_site, thenget_scanwithwait_seconds: 25until complete. - Run
list_issueswithmin_severity: "high". Group issues by category. - For each issue, open the matching files in the repo and fix it. Use the related skill: client-side-rendering-fix, ai-crawler-robots-txt, title-tags, meta-descriptions, canonical-tags, organization-schema, internal-linking.
- Show the user a diff per fix. Do not deploy without the user's approval.
- After deploy, run
scan_siteagain and compare the score and the issue count.
Output
Score before and after, a table of issues fixed (issue, page, file changed), and the issues left with a reason.