agentleFS
SpytrendSpytrend/spytrend-mcp/skills/spytrend-keyword-search/SKILL.md

spytrend-keyword-search

Skill30 starsChanged 3 months agoMITView on GitHub

Search and analyze ads matching a keyword or phrase with SpyTrend MCP data, with country breakdowns and exact creative-reuse ranking inside the selection, from a database of 1B+ tracked Meta ads. Works like an ad library search by text when the user does not know who runs the ads. Use when the user wants to find ads mentioning a product, brand, slogan or offer, pull ad examples for a topic, or measure how many ads use a phrase by market. Triggers: "facebook ads spy", "ad spy", "find ads with the word X", "search the ad library for a phrase", "which ads use this slogan"; RU «найди объявления со словом…», «поиск по библиотеке рекламы по ключу», «спай рекламы по ключу»; ES «buscar anuncios por palabra clave»; PT «biblioteca de anúncios»; JA «広告ライブラリで検索». Not for a whole vertical (use spytrend-vertical-intel), a known advertiser (use spytrend-advertiser-overview) or a known domain (use spytrend-website-review).

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 3 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    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 20
  2. Distinctive:

    Ask whether the search is worldwide or country-specific when that choice is absent and materially affects the answer.

    line 26
  3. Very specific:

    Call get_usage, present maximum row depth and keep media off by default.

    line 28
  4. Very specific:

    Call search_ads with the exact query, selected filters and an explicit small limit.

    line 32
  5. Very specific:

    When a market breakdown is requested, bound the selection first: a free-text query needs concrete date_from/date_to (a week or two is …

    line 34
  6. Distinctive:

    When analytics are pending, render ready sections and their coverage.

    line 47
  7. Distinctive:

    Use one-row counter calls only for explicitly requested exact slices and name the total status.

    line 49
  8. Very specific:

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

    line 66

≈ “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 keyword advertising search

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

Scope and preflight

  • Required: keyword or short phrase.
  • Optional: country, vertical and inclusive concrete dates.
  • Ask whether the search is worldwide or country-specific when that choice is absent and materially affects the answer.
  • Call get_usage, present maximum row depth and keep media off by default.

Workflow

  1. Call search_ads with the exact query, selected filters and an explicit small limit. Keep response order and date axis.
  2. When a market breakdown is requested, bound the selection first: a free-text query needs concrete date_from/date_to (a week or two is enough) on top of the country. An open-ended text selection does not settle — its sections stay ready=false no matter how many times the call is re-issued, and the advertised retry_after_seconds will not change that. Say that the window was added and why. Then call get_ads_analytics with the same query/country/category/date filters and an explicit max_items_per_section. If that tool is disabled or the account cannot use it, continue with the bounded ad rows and report the returned access state.
  3. For each analytics section, read its own status, ready, available_count, returned_count and truncated. A section's available_count is bounded section inventory, not market breadth. Market breadth comes only from total with total_status.
  4. When analytics are pending, render ready sections and their coverage. Do not paid-poll after a response has delivered rows.
  5. Use one-row counter calls only for explicitly requested exact slices and name the total status. Missing or unavailable is not zero.
  6. For creative reuse inside the selected filters, use the server's selection reuse ordering and the response field that counts ads in that selection. Keep lifetime reuse as a separate labeled value when returned.
  7. A representative ad is one example placement from a creative group; its copy and destination do not describe every ad in that group.
  8. With media confirmation, call get_media for only the representative ads the user selected. Do not fetch broad creative media automatically.

Report

Show exact filters, axis, total status, returned rows, truncation, advertisers, networks, destinations and top reused creative groups in the selection. State which analytics sections were returned and which were unavailable. Do not call a bounded item count a market total.

Use references/report-template.md or the packaged renderer with references/report-template.html; otherwise provide the full report in chat.

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.