Get the SEO and AI visibility skill for free.

What you download

The public download is Three-Surface Search, a portable skill folder with one SKILL.md entrypoint. Keep the ZIP zipped when you upload it. The package is free. No Bloomiro account is required to use it.

A skill is reusable instruction, not autonomous monitoring, not model training, and not a live connection to your website. After you import it, you still start a task in Claude, ChatGPT, or another agent and give it your product, competitors, customer language, or Search Console export.

• ZIP package: three-surface-search.zip, for Claude, ChatGPT, and other agents that support skill uploads.
• Entrypoint: SKILL.md, with triggers, inputs, stages 1–10, output contract, failure handling, and examples.
• The skill tells the agent to work only from data you actually have or can connect. It does not invent scrapers or tools that are not available.
• No API keys, tracking scripts, customer data, or runtime code are included.

What the skill does

Three-Surface Search turns real customer language into a self-reinforcing content engine. The same buyer questions are treated as three extraction surfaces: classic Google search, LLM answers (ChatGPT, Claude, Gemini, Perplexity), and direct-answer boxes (featured snippets, People Also Ask, voice).

• Start from real buyer wording you already have: tickets, call notes, pasted reviews, surveys.
• Turn problem clusters into search opportunities and the right asset type per intent.
• Build pillar plus supporting clusters with internal links mapped before writing.
• Write answer-first pages humans and LLMs can extract from.
• Wire a publish loop an automation agent can run, with Search Console as the weekly feedback stage.
• Prefer updating near-page-1 pages before adding new posts, then check GEO/AEO citations on the same queries.

How it works with Bloomiro

Inside Bloomiro Actions, Three-Surface Search is already installed for every project. Users do not upload, view, or delete it. When the agent uses it, you see a tool-style progress line such as Reading Three-Surface Search skill.

The agent calls read_skill, which returns a catalog first, then loads only the section ids it needs (usually overview plus one or two stages). It combines those instructions with Bloomiro evidence tools: Search Console, AI presence, sitemap, competitors, and CMS when connected.

• Ask in chat: use Three-Surface Search, use the skill, or describe an organic content / GEO / cluster workflow.
• Bloomiro supplies measured data and write paths. The skill supplies the research → cluster → write → measure → update loop.
• The agent must not invent scrapers or language miners. It uses connectors and exports that actually exist for the project.

How it works without Bloomiro

Outside Bloomiro, the same stages run with exports and manual checks. You provide product and ICP context, competitor URLs, real customer language you already have, a Search Console export or CSV, and CMS or repo access if you want publishing help.

• No connected GSC: paste query reports so stage 6 can still propose page actions.
• No CMS API: the agent can still produce cluster maps, briefs, and drafts for you to paste or PR.
• GEO/AEO checks: run the same high-intent queries in ChatGPT, Claude, Gemini, and Perplexity and log mentions and citations by hand.
• The workflow still refuses to invent topic clusters from AI topic lists alone. Real customer language comes first.

Selective loading

In Bloomiro Actions, the agent loads only the stages the current task needs so it does not dump the entire playbook into context. Outside Bloomiro, tell your agent which stages matter for the request.

• Overview / what you need: decide which stages apply and confirm available data.
• One stage: for a narrow ask (for example only Search Console feedback).
• Two or three stages: for a bounded plan (for example language + clusters + writing).
• Full loop: only when you asked for the complete organic channel workflow.

Privacy and limits

The download contains no private customer data. Evidence you paste into a chat stays subject to that agent's policies. Project data used through Bloomiro stays inside the authorized account.

This skill does not guarantee rankings, citations, or traffic. Native skill imports, file context, web access, MCP connections, and write permissions still depend on each agent host, plan, workspace, and administrator settings.

Example tasks to try after install

Give the agent a real product description and real customer language you already have.

• Build a content cluster from these support tickets and competitor review excerpts.
• We publish AI blogs with no traffic. Diagnose and reset the workflow.
• Use this Search Console export to choose which existing pages to update before writing anything new.
• For our top integration query, check ChatGPT, Claude, Gemini, and Perplexity and tell me what our page is missing to be citable.
• Draft a pillar brief with internal links mapped before any writing starts.