Key takeaways
- Only tool in the category with broad multi-agent support — works with 35+ agents via skills.sh including Claude Code, Cursor, Codex, and Gemini CLI
- 2,092 stars and 291 forks as of June 2026 — up from 452 stars in March 2026. Apache-2.0 license, now primarily Python. 20 SEO and GEO skills with 5 commands
- Custom evaluation frameworks — CORE-EEAT Content Benchmark and CITE Domain Rating provide structured scoring methodologies
- Actively maintained — v9.9.10 shipped June 5, 2026, adding MCP integration, a HOT/WARM/COLD memory management system, and zero-dependency Markdown skill content
FAQ
What is seo-geo-claude-skills?
A library of 20 SEO and GEO skills distributed via skills.sh that works with 35+ AI coding agents including Claude Code, Cursor, Codex, and Gemini CLI. Covers keyword research, content writing, technical audits, and rank tracking.
What agents does seo-geo-claude-skills support?
It supports 35+ agents via the skills.sh registry — Claude Code, Cursor, Codex, Gemini CLI, Windsurf, and many more. This is the broadest agent support in the SEO/GEO skill category.
What are the CORE-EEAT and CITE frameworks?
CORE-EEAT is a content benchmark framework that scores content against Experience, Expertise, Authoritativeness, and Trustworthiness criteria. CITE is a domain rating system for evaluating site authority. Both are custom frameworks built into the skill.
How does it compare to Claude Code-only tools?
It trades deep integration with a single agent for broad compatibility. Claude Code tools like claude-seo and geo-seo-claude still go deeper on subagent orchestration, but seo-geo-claude-skills closed the MCP gap in 2026 and now matches the leaders on adoption at roughly 2,100 stars.
Overview
seo-geo-claude-skills is the multi-agent play in the SEO/GEO skill category — the only tool that works across 35+ AI coding agents via the skills.sh registry. While competitors like claude-seo and geo-seo-claude are locked to Claude Code, this library runs on Cursor, Codex, Gemini CLI, Windsurf, and dozens more. It is free and open source under Apache-2.0; optional paid data connectors (documented in CONNECTORS.md) are the only cost surface.
Key stats (as of June 2026): 2,092 stars, 291 forks, Apache-2.0 license, primarily Python (88%) with Shell (12%). Created December 18, 2025. Latest release v9.9.10 on June 5, 2026.
| Attribute | Value |
|---|---|
| Stars | 2,092 (June 2026) |
| Forks | 291 |
| License | Apache-2.0 |
| Language | Python (88%), Shell (12%) |
| Created | December 18, 2025 |
| Latest Release | v9.9.10 (June 5, 2026) |
| Skills | 20 |
| Commands | 5 |
| Agent Support | 35+ via skills.sh |
| Category | SEO/GEO Agent Skills |
How It Works
Installation happens through skills.sh — a universal skill registry that bridges the gap between different AI coding agents — or via Claude Code's plugin marketplace (/plugin marketplace add aaron-he-zhu/seo-geo-claude-skills). Once installed, the library exposes 5 commands (consolidated from the earlier 9) covering the full workflow. The 20 skills are organized across Research, Build, Optimize, and Monitor phases, plus 24 diagnostic scripts.
Architecture
The codebase shifted from Shell to primarily Python (88%) over the first half of 2026, while keeping the skill content itself as zero-dependency Markdown with a small Bash runner for Claude Code hooks. Every skill follows a "Skill Contract" operating model with consistent activation, handoff, and data source patterns, and a memory management system tracks projects across HOT/WARM/COLD states.
CORE-EEAT Framework
The CORE-EEAT Content Benchmark provides a structured scoring methodology for evaluating content quality against Google's E-E-A-T guidelines. Rather than a single score, it breaks down content performance across multiple dimensions with actionable improvement suggestions.
CITE Domain Rating
The CITE framework evaluates overall domain authority and trustworthiness — a site-level complement to the page-level CORE-EEAT analysis. Together, they provide a comprehensive quality assessment that covers both content and domain signals.
5 Commands
The command surface was consolidated from 9 commands at launch to 5 as of v9.9.10:
| Command | Purpose |
|---|---|
/aaron:auto | End-to-end automated SEO/GEO workflow |
/aaron:research | Keyword discovery, clustering, and competitor analysis |
/aaron:create | SEO-optimized content generation |
/aaron:audit | Technical, on-page, and content quality audits |
/aaron:track | Rank tracking and search visibility monitoring |
Competitive Position
seo-geo-claude-skills occupies a unique position in the SEO/GEO agent skills category. At 2,092 stars as of June 2026 — more than 4x its March 2026 count of 452 — it has moved from mid-tier to first-tier on adoption while remaining first-tier on compatibility. For teams using Cursor, Codex, Gemini CLI, or any agent that isn't Claude Code, this is effectively the only option.
The depth gap has narrowed. Claude Code-only tools like claude-seo (7 subagents) and geo-seo-claude (citability scoring, parallel analysis) still go deeper on subagent orchestration, but seo-geo-claude-skills added MCP integration support in its 2026 releases, removing what was previously its clearest feature deficit.
The Apache-2.0 license (vs MIT for most competitors) is slightly more permissive for enterprise use — a small but meaningful detail for commercial adoption.
Strengths
- 35+ agent support — the only tool that works outside Claude Code
- Strong adoption — 2,092 stars and 291 forks as of June 2026, matching the category leaders
- Active maintenance — v9.9.10 shipped June 5, 2026
- MCP integration — added in 2026 releases, closing its biggest earlier gap
- Custom frameworks — the 80-item CORE-EEAT benchmark and 40-item CITE domain evaluation provide structured scoring
- 20 skills — broad SEO and GEO coverage in a single package
- Apache-2.0 license — enterprise-friendly
Weaknesses
- No parallel subagents — lacks the concurrent analysis architecture of claude-seo and geo-seo-claude
- No citability scoring — misses the GEO-specific depth of geo-seo-claude
- Rapid churn — the move from Shell to Python and the 9-to-5 command consolidation mean docs and tutorials from early 2026 may be stale
- Single maintainer — the project remains driven by one author, a bus-factor risk despite the release cadence
What Developers Say
No verbatim, attributable developer testimonials for seo-geo-claude-skills were found in public forums, blog posts, or social media as of June 2026. The project is listed on skill marketplaces (Open Agent Skill, Claude Plugin Hub, LobeHub) but without user reviews. Adoption signals are limited to GitHub metrics: 2,092 stars and 291 forks.
Bottom Line
Recommended if you need SEO/GEO skills outside Claude Code — its multi-agent support via skills.sh remains unmatched in the category, and at 2,092 stars it is now also a first-tier choice on adoption. The CORE-EEAT and CITE frameworks add structure to quality assessment, and the 2026 MCP integration removed its clearest feature gap. For Claude Code users wanting parallel subagent orchestration, claude-seo and geo-seo-claude still go deeper.
Outlook: positive. A 4x star increase in three months, a June 2026 release, and a shift to a more capable Python codebase suggest sustained momentum; the main risks are single-maintainer dependency and interface churn.