跳至正文
OpenAgentSkill
浏览已开放 · 目录记录日期 2026-11-21 · 自动更新和账户服务暂时关闭。
搜索与研究

cf-aeo-check

Post-publication AI-visibility check — re-probes target queries via web search, records whether the published piece is cited in Google AI Overviews, audits on-page extractability (definition block, dates, schema, structures) via web fetch, and appends every check to per-brand history so deltas are tracked over time. Triggers on \"/contentforge:cf-aeo-check\", \"is our article cited in AI Overviews\", \"did we earn AI citations\", \"check AEO for this URL\", \"why is a competitor cited instead of us\". Measures Google-observable signals only (no ChatGPT/Perplexity probing unless an AEO-tracking connector is present, and then labeled); closes the loop /contentforge:cf-brief opens and routes absent verdicts to /contentforge:content-refresh with evidence. Reports and records — it does not edit or republish the page.

indranilbanerjee · 已有 AI 审核记录

安装与使用

运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。

npx skills add indranilbanerjee/contentforge --skill cf-aeo-check

技能说明


name: cf-aeo-check description: "Post-publication AI-visibility check — re-probes target queries via web search, records whether the published piece is cited in Google AI Overviews, audits on-page extractability (definition block, dates, schema, structures) via web fetch, and appends every check to per-brand history so deltas are tracked over time. Triggers on "/contentforge:cf-aeo-check", "is our article cited in AI Overviews", "did we earn AI citations", "check AEO for this URL", "why is a competitor cited instead of us". Measures Google-observable signals only (no ChatGPT/Perplexity probing unless an AEO-tracking connector is present, and then labeled); closes the loop /contentforge:cf-brief opens and routes absent verdicts to /contentforge:content-refresh with evidence. Reports and records — it does not edit or republish the page." argument-hint: "[published URL or REQ-ID]" effort: medium

AEO Citation Check — Post-Publication

Close the loop that /contentforge:cf-brief opens. The brief checks AI Overview presence and citation patterns before production; this skill verifies — after publication — whether the piece actually earned citations, tracks the trend across re-checks, and routes losses to /contentforge:content-refresh with specific evidence.

Honest scope — read before promising anything

  • What this skill measures directly: Google SERP state for the target queries via web search (AI Overview present? which domains are cited/visible? does the piece rank?), and on-page extractability of the published URL via web fetch (definition block intact, dates visible, schema present, metrics tables surviving the CMS).
  • What it cannot measure directly: citations inside ChatGPT, Perplexity, Gemini or Copilot answers — there is no public API for "was I cited," and probing chat UIs is not reproducible. Cross-engine measurement needs a third-party tracker (Profound, Otterly, Conductor AgentStack, HubSpot AEO — same list the Phase 6 optimizer names). If such a connector is available in the tool list, use it and label the data's source; if not, say plainly that engine coverage is Google-observable-only. Never present an estimate as a measurement.

When to Use

Use /contentforge:cf-aeo-check when:

  • A piece published 2+ weeks ago and you want to know if AI engines are citing it (earlier checks mostly measure indexing lag, not merit)
  • You're deciding which content to refresh and want citation evidence, not hunches
  • A competitor appears in AI Overviews for your target query and you want the delta documented
  • You want a standing citation scoreboard per brand across re-checks

Not for: pre-production research (that's /contentforge:cf-brief), rank tracking as a discipline (this is a citation check, not a rank tracker), or measuring engines it cannot observe (see Honest scope).

Required Inputs

Minimum required (one of):

  • Published URL — the live page to check
  • Requirement ID (e.g., REQ-001) — resolved via the brand's tracking backend to the published URL recorded at completion; fails with a clear message if no URL was recorded

Optional:

  • Brand — brand profile (auto-detected from the tracking record when using REQ-ID)
  • Queries — comma-separated override of the queries to probe. Default: primary keyword + top 3 question keywords, recovered from the run's phase-6-seo.md (preferred) or the brief; if neither exists, derive 3-5 natural queries from the page's H1/H2s and say so. Also fold in any FAQ-disposition questions from the Phase 7 review report — the questions the reviewer found readers silently asking are the closest thing to real query data the pipeline produces, and they are exactly what answer engines get asked
  • --compare — show deltas vs the previous check (default when history exists)

How to Use

/contentforge:cf-aeo-check https://acme.com/blog/ai-in-healthcare-2026
/contentforge:cf-aeo-check REQ-001 --brand=AcmeMed
/contentforge:cf-aeo-check REQ-001 --queries="ai diagnostics accuracy,is AI better than radiologists"

What Happens

Step 1: Resolve target + queries (10-20 seconds)

  • Resolve REQ-ID → published URL via the tracking backend (local / google_sheets / airtable)
  • Assemble the query set (see Required Inputs); record which source the queries came from — scorecard honesty rule: derived queries are labeled derived, never passed off as the brief's targets

Step 2: SERP + AI Overview probe (per query, 1-3 minutes total)

For each query, via web search:

  1. AI Overview present? yes / no / unclear (record verbatim which sources are visibly cited when the search surface exposes them)
  2. Own-domain citation: is the published piece (or its domain) among the cited/visible sources? cited / visible-not-cited / absent
  3. Competitor presence: which Phase-1/Phase-2-era competitors (or new domains) hold the citations
  4. Organic position band for the published URL where observable: top-3 / top-10 / beyond / not-found — a band, not a fake precise rank

Step 3: On-page extractability audit (30-60 seconds)

Web-fetch the published URL and verify the machine-liftable elements survived the CMS:

  • Definition sentence within the first 150 words — intact?
  • Publication + last-updated dates visible?
  • Structured elements (tables, numbered steps, Q&A headers) from the phase-6 structure manifest — still present in the rendered page?
  • Article/Person schema in the page source — present? (CMSes routinely strip JSON-LD; this is the #1 silent AEO regression)
  • Author byline with credentials — rendered?

Step 4: Record + delta (10 seconds)

Append the check to ~/.claude-marketing/{brand-slug}/aeo/checks.json (same storage-resolution rules as every other artifact — $CLAUDE_MARKETING_HOME → $CLAUDE_PLUGIN_DATA → ~/.claude-marketing):

{
  "url": "...", "requirement_id": "REQ-001", "checked_at": "YYYY-MM-DD",
  "queries": [{"q": "...", 
阅读完整来源 ↗