SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillsto-ai-query

to-ai-query

Agents & models

View repository

Updated 10-16 05:54

About this skill

Queries to AI prompts

Use this report to turn first-party GSC query wording into a bounded set of prompts for later manual or automated monitoring. Each source query keeps its clicks, impressions, CTR, and average position, while deterministic templates produce stable prompt suggestions. The output is useful for designing a monitoring corpus. It is not evidence that anyone entered those prompts into an AI product, that an answer mentioned the site, or that the source query has equivalent AI demand.

Run the report

Use the compact MCP flow:

  1. Call seo_list_reports with {"category":"ai-search"}.
  2. Call seo_describe_report with {"id":"to-ai-query"}.
  3. Call seo_run_report with {"id":"to-ai-query","params":{"site":"sc-domain:example.com","days":90,"limit":20,"minImpressions":100,"maxRows":10000}}.

The CLI uses the same registry entry:

seo reports describe to-ai-query --json
seo reports run to-ai-query --params '{"site":"sc-domain:example.com","days":90,"limit":20,"minImpressions":100,"maxRows":10000}' --json

site is required. Use either days or an explicit startDate and endDate pair. maxRows bounds source retrieval and limit bounds returned queries. Supply brandTerms or includeBrand:true only when brand handling is intentional.

Interpret and act

Check dataStatus before reading items: empty, filtered, partial, and available mean different things. Preserve dateRange, source.possiblyTruncated, filters, selection.invalidRows, selection.conflictingRows, limited rows, warnings, and caveats. methodology.observedAiPromptData and estimatedTrafficLift are explicitly false. Each item has evidenceScope: retained-gsc-query; its prompts inherit that narrow scope.

Safe actions are to select representative prompt variants, store the source query and date range beside later observations, and refresh the corpus on a controlled schedule. Do not report generated prompt count as AI demand, fill missing prompts with guessed traffic, or interpret no returned rows as an all-clear when the source is partial or filtering removed candidates.