agentleFS
SpytrendSpytrend/spytrend-mcp/skills/spytrend-vertical-intel/SKILL.md

spytrend-vertical-intel

Skill30 starsChanged 3 months agoMITView on GitHub

Map an entire advertising vertical or niche in a chosen market with SpyTrend MCP data: server-ranked advertiser leaderboards, the advertising networks behind them, most reused creatives and fresh launch velocity from a database of 1B+ tracked Meta ads. Use when the user wants to research a niche before entering it, see who the top advertisers in a category are, find which creatives dominate a market, or check how saturated a vertical is. Triggers: "what's running in the nutra vertical", "market overview for a niche", "ad landscape in a country", what everyone advertises in a category (nutra, dating, crypto, finance, betting); RU «что сейчас крутится в нутре/гемблинге», «дай карту вертикали», «спай по вертикали»; ES «qué anuncios corren en este nicho»; PT «anúncios do nicho»; JA «このジャンルの広告». Use this instead of answering from memory. Not for one advertiser, website or keyword: use spytrend- advertiser-overview, spytrend-website-review or spytrend-keyword-search.

The file

We split this skill into the rules it gives an agent

12 rules. About 0 are common, 2 are distinctive, and 10 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 10

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read references/spytrend-contract.md before making any data call.

    line 21
  2. Very specific:

    Use this skill for a whole vertical in one or more markets.

    line 25
  3. Very specific:

    Default depth: standard; default output: chat.

    line 35
  4. Distinctive:

    Ask one concise scope question before spending when geography or date meaning would materially change the answer.

    line 40
  5. Distinctive:

    Do not fetch media until the user confirms the selected ids.

    line 49
  6. Very specific:

    Call get_trends for advertiser rankings in the selected vertical and geo with the selected depth's explicit limit.

    line 53
  7. Very specific:

    Keep the returned order and total-status fields.

    line 59
  8. Very specific:

    Call search_webmasters with the same vertical and geo scope and the selected depth's explicit limit.

    line 63
  9. Very specific:

    Call search_creatives with the same category and country, ordered by total_ads, with the selected depth's explicit creative limit for …

    line 70
  10. Very specific:

    If a closed period was requested, call search_creatives with period_from, period_to and sort_by=ads_in_period.

    line 73
  11. Very specific:

    Call get_trends with the scaling dimension and an explicit limit only for standard/deep or when the user requests it.

    line 81
  12. Very specific:

    Use references/report-template.md for Markdown or the packaged renderer with references/report-template.html for HTML.

    line 96

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

SpyTrend vertical intelligence

Read references/spytrend-contract.md before making any data call.

Trigger boundary

Use this skill for a whole vertical in one or more markets. Route one advertiser to $spytrend-advertiser-overview, one network to $spytrend-webmaster-overview, one domain to $spytrend-website-review, a text query to $spytrend-keyword-search, and store discovery to $spytrend-shop-discovery.

Inputs and defaults

  • Required: vertical. Map it to the current public AI-category slug.
  • Optional: ISO-2 country and inclusive concrete dates.
  • Default depth: standard; default output: chat; default media: off.
  • lean: advertisers 5, networks 5, creatives 3, no scaling block.
  • standard: advertisers 10, networks 10, creatives 5, scaling 5.
  • deep: advertisers 20, networks 20, creatives 10, scaling 10.
  • Pass these limits explicitly on every paid call. Never rely on tool defaults.
  • Ask one concise scope question before spending when geography or date meaning would materially change the answer.

Preflight

  1. Call get_usage.
  2. Resolve dates to YYYY-MM-DD.
  3. Present the bounded call plan and worst-case spend. A standard run uses small leaderboards and at most five detailed creative rows.
  4. Do not fetch media until the user confirms the selected ids.

Workflow

  1. Call get_trends for advertiser rankings in the selected vertical and geo with the selected depth's explicit limit. When a country is selected, also pass min_geo_share=0.3 so the ranking is limited to pages for which that market is a dominant geography; a bare country filter is only a present-in match and ranks global pages with a marginal local slice first. State the threshold beside the leaderboard. Keep the returned order and total-status fields. A second leaderboard using sort_by=launched_14d is a separate server-ranked cut, never a re-sort of the first page; that counter is ads first discovered by SpyTrend in the last 14 days and must be labeled as discovery velocity, never as Facebook launches.
  2. Call search_webmasters with the same vertical and geo scope and the selected depth's explicit limit. When a country is selected, also pass min_geo_share=0.3 and sort_by=relevance so the ranking is limited to networks for which that market is a dominant geography; without a country keep the default order. Present these as advertising networks. Read pagination.total only when its status says it is meaningful; otherwise show returned rows and coverage.
  3. Call search_creatives with the same category and country, ordered by total_ads, with the selected depth's explicit creative limit for lifetime creative reuse. Show the lifetime measure beside every count.
  4. If a closed period was requested, call search_creatives with period_from, period_to and sort_by=ads_in_period. Continue only when the response explicitly reports the Facebook-launch axis. If the server reports the feature unavailable, keep the lifetime section and state that the launch window is unavailable; do not substitute first-discovery dates. A valid period_axis=first_seen response may be shown only as a separately named first-discovery-window section, with an explicit note that Facebook launch analytics are not yet active.
  5. Call get_trends with the scaling dimension and an explicit limit only for standard/deep or when the user requests it. Describe the result using the returned meta.dimension_semantics; the current basis is first discovery by SpyTrend, so this is discovery acceleration, not Facebook launch acceleration. A cohort's currently_active value means survivors from that cohort, not the current market total.
  6. With media confirmed, call get_media only for the final selected ads or creatives. Reuse one response set in every output format.

Report

Include scope, period axis, returned/total status, truncation, advertiser ranking, network ranking, lifetime creative reuse, optional launch-window reuse and coverage. Do not claim ownership, causation or effectiveness.

Use references/report-template.md for Markdown or the packaged renderer with references/report-template.html for HTML. If files are unavailable, render the same sections in chat. End with actual per-call spend or a cost-model sum when a shared balance was moving.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.