What does this skill do, and when should you use it?
geo-technical is one of 16 skills bundled in the geo-seo-claude repository, located at skills/geo-technical/SKILL.md. It audits a target site across 8 categories for a 100-point technical health score: crawlability, indexability, security, URL structure, mobile optimization, Core Web Vitals, server-side rendering, and page speed. Its distinguishing focus is GEO (Generative Engine Optimization): AI crawlers don't execute JavaScript, so missing SSR is fatal, and robots.txt rules that accidentally block GPTBot or PerplexityBot directly hurt AI citations. The output is a structured GEO-TECHNICAL-AUDIT.md report with per-category scores, prioritized issues, and non-scoring agent-readiness checks.
Given a homepage plus 2-3 key inner pages, it fetches raw HTML and HTTP headers via curl/WebFetch, then checks: robots.txt validity and access status for 11 AI crawlers (GPTBot, Google-Extended, ClaudeBot, PerplexityBot, Bytespider, etc.); XML sitemaps and crawl depth; canonicals, hreflang, duplicate content, and index bloat; HTTPS enforcement and 6 security headers; viewport, tap targets, and mobile content parity; LCP/INP/CLS thresholds against 2026 benchmarks (with estimation methods when CrUX data is unavailable); SSR status by comparing curl output to the rendered DOM, with framework-specific fix recommendations (Next.js, Nuxt, etc.); and TTFB, page weight, image optimization, compression, cache headers, and CDN usage. It produces GEO-TECHNICAL-AUDIT.md with score breakdowns, an AI crawler status table, prioritized issue lists, plus non-scoring checks for IndexNow, RFC 8288 Link headers, and Markdown content negotiation.
- An SEO consultant diagnosing whether a client's technical layer blocks AI crawlers before running a full AI-visibility assessment
- A team whose site is client-side rendered in React/Vue, needing to verify AI crawlers (which don't run JS) can read the actual content
- A site owner checking whether robots.txt is inadvertently blocking GPTBot, PerplexityBot, or ClaudeBot, causing exclusion from ChatGPT or Perplexity answers
- A dev/ops team wanting a scored, prioritized technical-debt list (missing security headers, high TTFB, oversized JS bundles)
- An e-commerce or SaaS site implementing IndexNow to speed Bing indexing and improve visibility in ChatGPT and Bing Copilot
How do you install this skill?
- This is a static source-only review; no audit workflow was executed and actual output quality is independently unverified.
- The skill issues network requests to target sites and uses Bash; confirm the target domain is legitimate and that you are authorized to audit it.
- Core Web Vitals and similar metrics are only estimated without field data; report scores should not be the sole decision basis.
- Skill and docs are English-only; Chinese users should assess fit themselves. README contains community-marketing content that should be separated from skill functionality.
- Shell / CLI
- Network access
- Local filesystem
curl/WebFetch for page fetching
Install the full skill collection (macOS/Linux): curl -fsSL https://raw.githubusercontent.com/zubair-trabzada/geo-seo-claude/main/install.sh | bash, or manually git clone https://github.com/zubair-trabzada/geo-seo-claude.git and run ./install.sh (on Windows, run install-win.sh in Git Bash). Requires Python 3.8+, Claude Code CLI, and Git; dependencies install into an isolated venv at ~/.claude/skills/geo/.venv/. Uninstall with ./uninstall.sh.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/zubair-trabzada/geo-seo-claude.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/skills/geo-technical" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
In Claude Code, trigger it with /geo technical <url>; or ask Claude to read skills/geo-technical/SKILL.md and follow it directly: supply the target URL (homepage + 2-3 key inner pages) and the skill fetches pages, runs all 8 audit categories, and writes GEO-TECHNICAL-AUDIT.md. The README documents /geo technical as the entry point; no standalone invocation method for this sub-skill alone is documented.
What are this skill's strengths and limitations?
- Transparent, concrete scoring: 8 categories, 100 points, with explicit point values and pass criteria per check
- Genuinely GEO-aware: captures nuances like AI crawlers not executing JS and the Google-Extended vs. Googlebot distinction
- All checks run with free tooling (curl, WebFetch) — no paid APIs or crawler services needed
- Structured output with prioritized fixes and actionable remedies (e.g., framework-specific SSR solutions)
- Forward-looking agent-readiness checks (RFC 8288, Markdown negotiation, IndexNow) are non-scoring, so absence isn't unfairly penalized
- Core Web Vitals assessment is estimated from page characteristics when field data is unavailable — approximate, not real-user data
- Mobile and rendering checks rely on a small sample (homepage + 2-3 pages), so site-wide conclusions are limited
- No test suite or benchmark evidence that scores correlate with actual search/AI visibility outcomes
- SKILL.md is version 1.0.0 and the repo has no topics; community validation is unclear
- You must install the whole repo for the environment; extracting just this skill means providing curl/WebFetch tooling yourself
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| GEO Technical SEO Audit Skill (geo-technical) this page | 54 · Use with care | ★ 11k | 20d ago | MIT |
| Technical SEO Checker | 60 · Recommended | ★ 3k | 11d ago | Apache-2.0 |
| Technical SEO Auditor | 49 · Use with care | ★ 19k | 21d ago | MIT |
| Full-Site SEO Audit | 53 · Use with care | ★ 19k | 21d ago | MIT |
| AI Search & GEO Optimizer | 52 · Use with care | ★ 19k | 21d ago | MIT |
Traditional technical SEO crawlers (e.g., Screaming Frog-style tools) cover similar crawlability/indexability/performance ground, but this skill's differentiators are AI crawler access analysis and SSR auditing — dimensions designed for AI search platforms like ChatGPT and Perplexity. Note: 'citability scoring, brand authority, schema markup, and PDF reports' mentioned in the repo description belong to sibling skills (geo-citability, geo-schema, geo-report-pdf), not this one.
How did FollowSkills review this skill?
Tool allowlist is explicit (Read/Grep/Glob/Bash/WebFetch/Write); behavior is mainly read-only fetching plus writing one audit report. No credential access, covert exfiltration, or destructive defaults. Deducted for: unconstrained Bash scope, no stated confirmation/isolation/rollback for fetching external sites or writing files, incomplete data-flow disclosure.
Eight audit categories, scoring rubrics, and output template are self-consistent with a clear executable path (curl/WebFetch), including degraded estimation when CrUX field data is unavailable. Deducted for: committed tests cover the shared fetch_page SSR heuristic, not this skill's own key paths; many checks rely on subjective judgment; abnormal-input failure feedback is undefined; static review cannot exceed 10.
Target scenario (technical SEO/GEO audit of any website), inputs, output format, and scoring boundaries are clear; scoring vs. non-scoring forward-looking checks are separated. Deducted for: English only with no Chinese-language support statement, trigger conditions rely on description matching, and non-fit ranges (e.g., login-gated sites) are undeclared.
Proper frontmatter, version 1.0.0, MIT license, uninstaller and install notes in the repo; readable structure. Deducted for: no per-skill changelog, no known-limitations section, author listed as the repo name (weak attribution), maintenance/update path only implicit, and README includes Skool community marketing.
Output format (score tables, AI crawler access table, prioritized issues) is directly usable by practitioners with clear marginal value over manual audits. Deducted for: static review cannot verify actual output quality; key metrics (measured CWV, TTFB distribution) are only roughly estimated without field data; no execution evidence of completeness.
Rubrics, crawler mapping tables, and referenced standards (RFC 8288, INP thresholds) are auditable in-source, partially corroborated by shared-script tests. Deducted for: no committed tests for this skill path, no third-party execution evidence, and README market statistics ($850M, +527%, etc.) lack citations; static cap of 5 applies.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →