agentleFS
Sign inStart free
dougwithseismicdougwithseismic/geo-kit/skills/geo-measure/SKILL.md

geo-measure

Skill2 starsChanged 9 days agoMITView on GitHub

Measure AI-search visibility and citations — pull Google Search Console data, check AI-referral and AI-bot telemetry, and run manual citation spot-checks against a baseline. Use when asked "are we getting cited by AI", "track our GEO/AI-search performance", or to report progress after GEO work.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Classic Search Console deltas: Run the kit's pull script and compare to the recorded baseline:

    line 18
  2. Very specific:

    If no baseline exists, capture one now (today's date) and note that future runs compare against it.

    line 67

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

GEO Measure

Report AI-search visibility. Most AI exposure is zero-click, so no single metric is enough — triangulate the six signals below and compare to a baseline.

The six signals

  1. Google generative-AI report (impressions). https://search.google.com/search-console/performance/search-analytics/ai?resource_id=<PROPERTY> — impressions in AI Overviews / AI Mode (no click data). Eligibility is gated by AI-impression volume, so small sites may not see it yet.
  2. Classic Search Console deltas. Run the kit's pull script and compare to the recorded baseline:
    node src/measurement/gsc-pull.mjs      # writes dated JSON to data/, prints deltas
    Auth: Application Default Credentials with the webmasters.readonly scope (gcloud auth application-default login --scopes=…/auth/webmasters.readonly,…/auth/cloud-platform). Watch the target-intent positions (page-2 → page-1 movement) and the new pillar pages appearing.
  3. AI-referral events (analytics). In GA4/PostHog, count sessions/events from AI-assistant referrers (chatgpt.com, perplexity.ai, claude.ai, gemini.google.com, …). The kit's AiReferralTracker emits an ai_referral event; segment by ai_source. Caveat: misses app traffic and noreferrer AI Mode.
  4. AI-bot fetch telemetry (server/middleware). Count requests whose UA matches the AI bots (kit's matchAiBot). ChatGPT-User / Perplexity-User hits are the closest proxy for being used in a live answer (vs GPTBot = training). Bing Webmaster Tools' AI Performance report adds real citation counts + grounding queries — check it if the site is verified there.
  5. Manual citation spot-checks (ground truth). Monthly, ask each engine the canonical prompts and record C=cited / M=mentioned / –=absent. Use the templates/ai-citation-checks.md runbook. Run in the target locale (location matters — engines localise).
  6. Fan-out coverage (monthly). Re-run the site's fixed term list through the agents and score the funnel per engine:
    node src/fanout/cli.ts report --in fanout-data/<date>-<site>
node src/fanout/cli.ts coverage --in fanout-data/<date>-<site> --site <domain> --brand "<Brand>" --gsc Queries.csv
    Track seen / opened / cited out of N runs per engine, and the "opened instead" domains. Use the same term list every month, or the numbers don't compare. The geo-fanout skill covers the first run.

Report format

  • Headline: direction of travel vs baseline (impressions, target-intent positions, AI-referral count, bot-fetch count, citation spot-check score, fan-out funnel per engine).
  • What moved and why (tie changes to shipped work).
  • Realistic framing: LLM referrals are ~0.1% of organic sessions even for top publishers today — report the trend and the citation presence, not just raw referral volume.
  • Next actions, ranked by leverage.

Baseline

If no baseline exists, capture one now (today's date) and note that future runs compare against it. Store baselines in the plan doc or templates/ai-citation-checks.md.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free