What is seo-llm-skill-cluster?

This repository is a collection of Codex skills, meaning packaged instruction sets an AI coding agent can load, aimed at building and auditing websites so they work well for human visitors, search engines, and AI agents that crawl and cite web content. Rather than one long prompt trying to cover all of SEO at once, the author split the work into thirteen specialized skills coordinated by a central orchestrator skill. The core skills cover mapping out a site's topics and search intent, planning its URL structure and canonical rules, designing its internal linking, auditing its technical metadata and schema.org markup, and checking whether its pages are readable and citable by LLM agents through files like llms.txt. A further set of companion skills handles editorial quality checks, mapping reader journeys, analyzing server logs for crawler behavior, monitoring whether LLMs cite the site, scouting for legitimate external authority opportunities, and validating backlink quality. Each skill is meant to produce a specific written output, such as a link graph, an audit report, or an improvement backlog. The cluster grew out of the author's own site, mlllm.io, which is used throughout the README as a worked example, with real sample reports for each skill included in the repository. The project ships with validation tooling, including a linter for the skill cluster, an evaluation verifier, and a continuous integration workflow that checks JSON and YAML syntax and scans for sensitive patterns before anything is published. The README is explicit that the approach is meant to be white hat only. It states there should be no hidden content shown only to bots, no duplicate content farms, no fake rankings or citations, no link farms or spam, and no outreach or external posting without explicit authorization, and that any authority placement skill only scouts and validates opportunities in a dry run mode rather than acting on its own. Work is tracked through an external task plan dashboard so a person can review progress without watching every step.