Skip to main content
Emit a technical recommendation against a specific URL: adds a recommendation card to the brand’s Action Engine and records the action for outcome measurement over time. Produces a recommendation ONLY (no CMS auto-apply). The user copies the suggestion and applies it in their own CMS. Common subtypes:
  • ‘schema_markup’ — add JSON-LD (Product / FAQ / Article / Organization)
  • ‘faq_block’ — add an FAQ section to capture FAQPage rich result
  • ‘tldr_summary’ — add a 2-3 sentence summary at the top of the page
  • ‘canonical_fix’ — set rel=canonical / consolidate duplicate URLs
  • ‘eeat_signal’ — author byline, credentials, last-updated date
  • ‘comparison_block’ — add a side-by-side comparison table vs cited competitors
Use this when:
  • get_topical_authority shows a specific URL is losing on prompts where competitors win, AND a structural fix would help.
  • The user asks “what should I change on /page-x to get cited more?”
Skip when:
  • The fix needs net-new content (use emit_content).
  • The fix needs a paid push (use emit_paid).
Inputs:
  • subtype (required): one of the categories above (free-form string ≤ 40 chars).
  • target_url (required): the page the recommendation applies to.
  • title (required, ≤ 200 chars): one-line summary of the change.
  • description (required): array of step-by-step instructions (each ≤ 500 chars).
  • effort_minutes (optional, default 30): estimated implementer effort.
  • priority (optional, ‘low’|‘medium’|‘high’|‘critical’, default ‘medium’).
  • target_prompt_ids (optional): tracked-prompt ids this recommendation expects to win.
Response: { actionId, taskId, deepLinkUrl }.

​
Inputs