인터페이스는 현지화되어 있으며, Skill 소개와 설명 문서는 선택한 언어로 AI가 자동 번역합니다. 코드·명령어·경로·전문 용어는 원문을 유지하며 원작자 문서가 우선합니다.
Skills 둘러보기
디자인 및 창작 · SKILL.md

citlyze-visibility-report

citlyze-visibility-report

Build a window-over-window AI search visibility report from Citlyze data. Use when asked "how visible is my brand in AI search", "what changed this week/month", "build an AI visibility report", or for recurring exec reporting on ChatGPT/Claude/Perplexity/Gemini visibility. Requires the Citlyze MCP server.

Independent GitHub snapshot · Automated classification · Original documentation governs

Skill 문서

GitHub

name: citlyze-visibility-report description: Build a window-over-window AI search visibility report from Citlyze data. Use when asked "how visible is my brand in AI search", "what changed this week/month", "build an AI visibility report", or for recurring exec reporting on ChatGPT/Claude/Perplexity/Gemini visibility. Requires the Citlyze MCP server. license: MIT metadata: author: Citlyze (https://www.citlyze.com)

Citlyze visibility report

Produce an executive-ready report on how a brand shows up in AI search engines, comparing the latest completed measurement window to the previous one.

Prerequisites

The Citlyze MCP server must be connected (setup guide). All Citlyze tools are read-only. When a tool takes measurement_window_id and none is given, it defaults to the latest completed window.

Workflow

  1. Orient. Call get_workspace_overview. Note the workspace name, target brand, active_prompts, active_engines, tracked_competitors, and latest_completed_window.
  2. Pick windows. Call list_measurement_windows with status: "completed". Take the two most recent windows. If only one exists, report a single-window snapshot and say trend data needs a second window.
  3. Pull metrics for both windows. Call get_visibility_overview once per window. Each row in by_brand_engine carries: visibility_score, mention_rate, citation_rate, source_rate, avg_prominence, share_of_answer, and n_runs.
  4. Competitive standings. Call list_competitor_visibility for the latest window. by_brand is sorted by visibility_score; note the target brand's rank and the gap to the leader.
  5. Headline movers. Compute per-engine deltas between the two windows for the target brand. Flag any engine where visibility_score moved notably in either direction. Always sanity-check surprising swings against n_runs: a big move on very few runs is weak evidence, so say so.
  6. Write the report in this shape:
    • TL;DR - three sentences: overall direction, best engine, biggest risk.
    • Scoreboard table - one row per engine: previous score, current score, delta, mention rate, citation rate.
    • Competitive position - rank, gap to leader, any competitor overtakes since last window.
    • Watch items - low-n_runs caveats, engines with falling citation rate, anything needing a decision.

Interpretation rules

  • visibility_score is Citlyze's blended per-engine score; compare it across windows for the same brand rather than across different companies.
  • A citation (link shown) is not the same as a mention (brand named in the answer) or a source consulted (page the engine read). Report them as separate signals.
  • Percentages are rates across runs in the window, not absolute counts.

Out of scope

This skill reports; it does not change anything in the workspace (the MCP surface is read-only). For acting on findings, chain into citlyze-action-plan.

Learn more: Citlyze dashboard guide - citlyze.com

저장한 Skills