More screenshots
Works with the AI Agents you already use
Claude Code
Cursor
Codex CLI
Muse
Dots+22 moreClaude CodeCursorCodex CLIMuseDots
OpenClaw
Manus
Grok
Grok Bot+18 agentsGeoscope: Deterministic GEO & AI-SEO Audit
Real HTTP probes, not vibes. Audits AI-search readiness with 11 crawler user agents, robots.txt, JSON-LD and an SSR check - then scores only what it fetched.
$11.99
Geoscope: Deterministic GEO & AI-SEO Audit
Example session with this skill installed
Audit https://acme-analytics.io for AI search visibility. I keep hearing we need to 'do GEO' - tell me what's actually broken for ChatGPT, Perplexity and Claude, with evidence I can hand to our web team.
- Read your context and instructions
- Compiled the geoscope
- Generated the document
node scripts/geo-audit.mjs https://docs.anthropic.com
(real output, unedited)
geoscope · https://platform.claude.com/docs/en/home
GEO score: 71/100 - FAIR
OK Crawler access (training + citation) 25 / 25
robots.txt parsed: 1 agent group(s), 2 sitemap ref(s)
no AI crawler is blanket-blocked by robots.txt
sitemap refs: https://platform.claude.com/sitemap.xml, https://platform.claude.com/cookbook/sitemap.xml
FAIL Structured data 0 / 20
0 JSON-LD blocks - no machine-readable entity data
OK AI files (llms.txt + sitemap) 15 / 15
llms.txt has a # title
788 doc links listed
llms-full.txt also present
WARN Content citability 17.9 / 25
only 1 answer block(s) in the citable range
4/17 sentences carry stats, numbers or years (24% quotable density)
0 tables + 10 lists - pre-structured data engines cite readily
WARN Technical foundations 13.4 / 15
629 words of server-rendered text
no SPA shell markers - content appears in the initial payload
title "Documentation - Claude Platform Docs" (36 chars)
crawler probe:
200 training GPTBot
200 training ClaudeBot
200 training CCBot
200 training Google-Extended
200 training Meta-ExternalAgent
200 search OAI-SearchBot
200 search ChatGPT-User
200 search PerplexityBot
200 search Perplexity-User
200 search Claude-SearchBot
200 search Amazonbot
edge fingerprint:
Cloudflare
verify by hand (1-click searches):
- site:reddit.com "claude"
- site:en.wikipedia.org "claude"
- "claude" review
- site:youtube.com "claude"
1792253395849-bp5m.png
PDF · document
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
A GEO score your dev team can defend.
Most GEO audits invent platform scores like "ChatGPT readiness: 18/100" that change every run and fall apart the first time a senior engineer asks where the number came from. Geoscope scores only what it can fetch: your page, your robots.txt, your sitemap, your JSON-LD, your llms.txt. Every point links to a line of evidence. Run it twice, get the same number.
What the audit runs:
- 11 crawler user agents in two buckets. Search bots (OAI-SearchBot, ChatGPT-User, PerplexityBot, Perplexity-User, Claude-SearchBot, Amazonbot) vs training bots (GPTBot, CCBot, Google-Extended, ClaudeBot, Meta-ExternalAgent). Blocking training while allowing citation is a real strategy; most tools flatten it into one useless number.
- robots.txt parsing that names exactly which rule blocked which bot.
- SSR trap check: AI crawlers don't run JS at scale. If your SPA serves an empty shell, crawlers see nothing. The audit counts words in the raw HTML, not the rendered DOM.
- JSON-LD entity checks: Organization, WebSite, FAQPage, @id, knowsAbout, and it re-fetches every sameAs link because a dead sameAs is worse than none.
- Citability heuristics from the Princeton GEO paper: stat density, 40-167 word answer blocks after headings, tables and lists.
- llms.txt, llms-full.txt, sitemap.xml, x-robots-tag headers, meta robots directives (nosnippet, noai, max-snippet:0), and a WAF fingerprint so a 403 comes back named "Cloudflare Bot Fight Mode", not just "blocked".
What it refuses to do:
Wikipedia presence, Reddit mentions, review-site ratings, backlink authority: real GEO inputs, but no local script can prove them. Instead of inventing a number, the report emits 1-click verification searches so a human checks them in seconds. Uncheckable signals are labeled uncheckable.
Also in the package:
- llms-gen drafts an llms.txt from your sitemap or a saved page. Output is marked DRAFT, nothing auto-deploys.
- schema-gen emits Organization + WebSite + FAQPage JSON-LD prefilled from your page's real meta.
- geo-report turns the JSON audit into a standalone dark HTML report you can send to a client or print to PDF.
- 9 passing tests, a real example audit of docs.anthropic.com (71/100, FAIR) and two offline fixtures that run in 5 seconds.
Priced at Node 18+1.99 one-time - about 40% under the going rate for GEO audit suites, because an audit you can verify shouldn't cost agency money.
Node 18+, zero dependencies, no API keys, no LLM calls. The demo on this page is real output, not a mockup.
How to use
Skills on Agensi run inside your AI agent. Copy the prompt, paste it into your agent, and it connects to Agensi and installs this skill for you. If a payment is needed, your agent shows it to you first.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill passes our trust checks below.
Security scanned
Passed our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scanned
Verified clean 2 days ago
- Passed all security checks, Safe to install
Requires