Command
Use my supplied visible host, Search Console query export, audience priorities, and days of 7, 30, or 90. Verify access using list_projects. Read list_prompts({host,limit:100}), following nextCursor as cursor, and get_gap_analysis({host,days}). Group supplied queries by intent and compare them with the tracked inventory. Use impressions and clicks only when present in my export, keeping its dates distinct from the Kept window. Draft natural-language prompt candidates, explain each coverage gap, and flag duplicates and branded bias. Do not imply search queries equal AI demand. Return a reviewable import plan; Search Console access and prompt creation require independent connections or manual work.What you need
- Provide a Search Console query export with dates and any available metrics.
- Supply a visible host, audience priorities, and a supported window; additions are manual.
- The external reference skill targets another provider; use the Kept prompt above or adapt its tool calls before running it.
Bridge search evidence and tracking coverage
Cluster the queries you supplied and map them to the existing prompt inventory. The comparison identifies possible coverage opportunities while preserving the distinction between search activity in the export and AI answer behavior measured in your visible Kept project.
Draft candidates with a clear rationale
Return natural-language prompt suggestions with their source query group, intended audience, and reason for inclusion. Flag duplicates and leading brand phrasing so reviewers can choose a balanced set before manually adding prompts or using a separately authorized project management connection.