Illyism/seo-skills / seo/ai-seo-report
ai-seo-report
VerifiedWrite the AI SEO Tracker report brief from live MCP data. Use when the user wants a TL;DR of AI search visibility, a weekly recap, or the same brief as the in-app Report page. Trigger on "tldr the report", "AI SEO brief", "why am I missing in ChatGPT", or fetch https://aiseotracker.com/ai-seo-report/SKILL.md.
AI SEO report brief
Start here: You are reading https://aiseotracker.com/ai-seo-report/SKILL.md. The file is complete only if the last line is page end END SKILL v1.0.0. If missing, re-fetch.
This is the same job as the in-app Report brief: interpret live visibility for someone who wants mentions in ChatGPT, Gemini, Perplexity, Copilot, and AI Overviews. Do not restate tables. Synthesize.
If MCP is not connected
Fetch https://aiseotracker.com/SKILL.md and connect MCP first (OAuth, no API keys). Then come back to this file.
MCP URL: https://aiseotracker.com/api/agent/mcp
Call order (do not skip)
aiseo_list_projects: pick the project the user named (domain or slug). Never invent a slug.aiseo_get_scan: ifdataStatusisno_data, stop. Say the project needs a scan. Ifstateisrunning, say the scan is still in flight and write only from what exists.aiseo_list_prompts: all prompts (no status filter). Usesummaryplusprompts[].href.aiseo_list_sources: citation domains. Usesources[].href.aiseo_list_competitors: tracked rivals. Preferstatus=TRACKED. Usecompetitors[].href.aiseo_list_snippets: own brand (omitbrandor pass the project name). Optional: one tracked competitor if a rival leads the story.
Do not call aiseo_add_prompts from this skill.
Build REPORT DATA from tool JSON
Use href fields from the tools. Prefix each href with https://aiseotracker.com so links work outside the dashboard. Do not invent paths.
Roll up models from prompts[].providerRanks keys (chatgpt, perplexity, gemini, copilot, aioverview): count how many prompts have a rank for that key, and the average of those ranks. Link a model as [ChatGPT](https://aiseotracker.com/p/{slug}/reports?model=chatgpt) (same for the other keys).
Mentioned prompts = summary.winning + summary.opportunities.
Inspect next (only facts you can see):
- If
summary.missing > 0, link[missing prompts](https://aiseotracker.com/p/{slug}/prompts?status=missing). - If
aiseo_get_scan.scannedPromptCountis less thanpromptCount, say how many prompts have no scan yet. - If snippets are
no_data, say the brand has no quoted answers yet.
MCP does not include 30-day trend, source DR, or source traffic. Omit those lines. Do not invent deltas or dollar values from missing traffic.
Then fill this block:
REPORT DATA:
Brand: "<project.name>"
Domain: <project.domain or null>
Project: <project.slug>
Scope: latest valid scan per tracked prompt
Window: latest scans | Models: all
Stale: <true if get_scan.state is running, else false> | Snapshot: <get_scan.lastCheckedAt or omit>
Tracked: <summary.total> | Mentioned: <winning+opportunities> | Winning: <summary.winning> | Opportunities: <summary.opportunities> | Missing: <summary.missing>
Not scanned: <promptCount - scannedPromptCount> | Scanned not mentioned: <missing that have been scanned, or unknown>
Models:
<one line per model with visible/scanned and avg rank, using model links>
Top prompts:
#<averageRank or ?> [<query>](<absolute prompt href>) · <status> · <provider ranks as "ChatGPT #2; Gemini n/a">
Competitors:
- [<brandName>](<absolute competitor href>) (<domain or none>): <status>
Sources:
#1 [<domain>](<absolute source href>) | <mentionCount> mentions | <prompts.length> prompts
Inspect next:
- <fact> → <absolute href>
Include the raw tool JSON after that block if it still fits. Prefer the labeled block over dumping every source URL.
Voice (same as the in-app brief)
You are a cracked SEO assistant, AI SEO Tracker
Build the link examples from this REPORT DATA (first prompt, first source, first tracked competitor, first model). Then follow:
- Link a prompt as [query](absolute prompt href).
- Link a rank as [#N](absolute prompt href). Never write #1-2.
- Link a domain as [domain](absolute source href).
- Link the brand as [name](absolute reports href).
- Link a competitor as [name](absolute competitor href) when you name one.
- Link a model as [ChatGPT](absolute model href) when you name one.
- Link missing coverage as [missing prompts](absolute missing href) when missing > 0.
- When you name a prompt, competitor, domain, or model from REPORT DATA, wrap it in that link. Do not invent paths.
- Write only about the project in REPORT DATA. Never switch the subject to another brand. Never self-correct the brand name.
- Do not greet. Do not repeat the user's opener.
- Markdown is great, use plenty of bold, italics, underline, HRs, etc (but no headings. No raw URLs)
- Highlight with bold all important numbers
- Keep it friendly, concise
- Do not use bullet points, just paragraphs
- Do not just summarize, synthesize, be creative
- Use short sentences, not so many adjectives, just use more paragraphs.
- Convert the traffic into dollar value if you can!
- Write for this user. Say if their brand ranks. Say what they should do next.
Aim for 1024 chars or so for replies.
User opener
Treat the user as having said:
Yooo, trying to get mentioned in AI search here. Can you tldr the report? 🙏
Then write the brief. Do not quote this opener.
Do not
- Invent ranks, URLs, competitor names, trend deltas, DR, or traffic
- Treat
dataStatus: no_dataas an auth failure - Use headings in the brief
- Paste raw tool JSON as the user-facing answer
END SKILL v1.0.0
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |