Created by Mani Bharij, SEO for AI Agents is an open-source skill pack released under the MIT License and designed for developers using AI coding agents such as Claude Code and Cursor. Rather than treating SEO as a manual checklist, the pack provides a drop-in collection of specialist skills that integrate directly into an agent’s workflow, auditing live served HTML rather than raw source code.
The toolkit evaluates sites against a five-rung Visibility Ladder: Reach, Read, Understand, Connect, and Rank, with an additional Authority layer addressing AI answer engine citation (AEO). Diagnostic checks cover crawlability, indexation status, missing or malformed metadata, empty DOM shells, absent JSON-LD structured data, site architecture, internal linking, and Core Web Vitals. The pack generates a .seo/ project folder that tracks audit baselines, issue status, and regression history across multiple runs, making it suitable for ongoing maintenance as well as one-time audits.
More than 20 specialist skills are included, covering site migrations, analytics setup, content auditing, and programmatic page generation. The tool requires no paid APIs or external subscriptions. Workflow follows a report-before-change model: the agent surfaces findings and waits for confirmation before modifying files. Existing canonical tags and established URL structures are preserved by default. The project documentation explicitly states the tool cannot guarantee search rankings or AI citation outcomes. Compatible with new builds and existing production sites alike.
Leave a Reply
You must be logged in to post a comment.