keyword-research

Pull keyword data from GSC, Ahrefs Webmaster, and DataForSEO. Returns a clustered keyword brief.

SKILL.md

GitHub

/keyword-research

Usage: /keyword-research <domain | topic | competitor-domain> [--for=<product-slug>]

Examples:

  • /keyword-research linkedup.io --for=linkedup — pull GSC + AWT data for own site
  • /keyword-research "form builder for indie hackers" --for=formpad — topic search via DataForSEO
  • /keyword-research typeform.com --for=formpad — competitor's ranking keywords via DataForSEO

What this does

Invokes agents/launch/keyword-researcher/SKILL.md. Output saved to products/<slug>/launch/keyword-briefs/YYYY-MM-DD-<topic>.md.

What you do as the assistant

  1. Resolve <for> to a product slug. If omitted and unambiguous from context, infer; else ask.
  2. Determine which API surface to call based on the input:
    • Looks like the product's own domain → GSC API + Ahrefs Webmaster
    • Looks like a competitor → DataForSEO domain queries
    • Looks like a topic (with quotes / spaces) → DataForSEO keyword suggestions
  3. If API credentials aren't set yet (check .env.local for DATAFORSEO_LOGIN, GSC OAuth, etc.), tell the user what's missing and provide the setup link. Don't fail silently.
  4. Execute the agent's protocol.
  5. Save output, print location, summarize top 3 clusters in chat.

Cost discipline

Per brief, target < $0.50 in DataForSEO API costs. If a query batch would exceed $2, confirm with user before running.