Marketing Foundation
Compact two-skill starter: clarify positioning and choose a lead magnet. Use Marketing Launch for the broader eight-skill go-to-market workflow.
The tunnel uses the same cards as the catalogue. Browse only as deep as needed — or load a broad bundle immediately.
SEO, Sales, Agents or another broad area → one bundle call → work.
Read-only access to published skills. Default 8, maximum 10 skills / 120,000 characters.
Compact two-skill starter: clarify positioning and choose a lead magnet. Use Marketing Launch for the broader eight-skill go-to-market workflow.
Build an evidence-led marketing plan from ICP and competition through positioning, campaigns, growth and measurement.
Diagnose architecture and context, plan agent-team responsibilities, then organize project context and session handoffs. Memory and cost-runtime reviews remain outside this pack.
Review the journey from landing page and lead capture through registration, first value and transparent upgrades.
Plan a campaign, draft its channel content and review the work against actual brand guidance.
Prioritize an editorial roadmap and plan how to launch and distribute it across suitable channels.
Understand customer needs, compare competitors and plan a community around real member value.
Choose a relevant lead magnet, then draft a permission-based nurture journey with entry, suppression and exit rules.
Define the API contract, then plan how to observe its latency, failures and retries. Guidance and checklist; no production changes.
Profile a dataset, choose and interpret statistical methods, then validate calculations and conclusions before sharing.
Define the target account, prioritize buying signals, plan a human LinkedIn engagement routine and prepare evidence-led responses to buyer concerns.
Research public Meta ad-library samples and optionally business contact information through HyperFX; use Competitor Ad Collection for GoMarble.
---
name: meta-ads-library
description: Research competitor Facebook and Instagram ads from the Meta Ads Library via the Hyper MCP — search by keyword, pull full ad creative and metadata, enrich with page contact info for lead generation, and surface structured ad-intelligence summaries in chat. Use when the user wants to scrape the Meta Ads Library, spy on competitor ads, monitor new ads in a category, build a lead list from advertisers, or surface creative trends across an industry.
requires_toolkits:
- meta_ads_library
icon: meta_ads_library
short_description: Search the Meta Ads Library for competitor ads and surface structured ad intelligence.
---
# Meta Ads Library
Guide for searching the Meta Ads Library and producing structured competitor ad intelligence.
The skill's job is to turn raw scraped ads into useful summaries: top advertisers, common CTAs, recurring hooks, recently launched creatives, and (optionally) enriched lead lists. All output is presented inline in chat — there is no database or persistence layer.
## Out of scope — defer to other skills
| Request | Send them to |
| --- | --- |
| Multi-source competitor research (site, social, search rank, etc., not just ads) | [`competitor-intel`](../competitor-intel) |
| Generating *new* ad creative based on what you found | [`ad-creative-generation`](../ad-creative-generation) |
## Requirements
- **Hyper MCP installed and connected.** [https://app.hyperfx.ai/mcp](https://app.hyperfx.ai/mcp)
- **Apify integration connected** at [https://app.hyperfx.ai/apps](https://app.hyperfx.ai/apps) — the Meta Ads Library tools run via Apify under the hood.
If `search("meta_ad_library_ads_search")` does not find `meta_ad_library_ads_search`, stop and tell the user to enable Hyper MCP and connect Apify.
### How to run the tools in this skill
Every tool in this skill is named by its canonical tool name. Run it with the call your surface gives you:
| Surface | Find a tool | Run it |
| --- | --- | --- |
| MCP client (Claude, Cursor, Codex, ChatGPT) | `search("<what you want to do>")`, then `describe("<name>")` | `call("<name>", {...})` |
| Hyper CLI | `hyperai search "<what you want to do>"`, then `hyperai describe <name>` | `hyperai call <name> --json '{...}'` |
If a tool is not found, its integration is not connected or not enabled for the workspace: stop and tell the user which integration to connect.
## Tool surface
| Tool | Purpose |
| --- | --- |
| `meta_ad_library_ads_search` | Search the Meta Ads Library by keyword. Returns compact results (title, body, CTA, link, page name, dates, platforms). Max 40 per call. |
| `meta_ad_library_ads_get` | Get full details for a specific ad. Requires both `ad_archive_id` **and** `page_id` — both come from `meta_ad_library_ads_search` results. |
| `meta_ad_library_ads_search_enriched` | Search + enrich each result with page contact info (email, phone, website). Slower (multiple API calls per result). Max 20 per call. |
| `meta_ad_library_pages_search` | Search Facebook pages by category + location (not by keyword). Useful for building a lead list from a vertical. |
| `meta_ad_library_pages_scrape` | Scrape detailed data from specific Facebook page URLs. |
## Critical rules
1. **Public-only data.** The Meta Ads Library is public. Don't attempt to bypass any access control or scrape private content.
2. **Count limits differ between tools.** `meta_ad_library_ads_search` allows `count` up to 40. `meta_ad_library_ads_search_enriched` caps at 20 — exceeding this returns an error.
3. **`meta_ad_library_ads_get` needs two IDs.** Both `ad_archive_id` and `page_id` are required. Both are returned in every `meta_ad_library_ads_search` result row — pass them through together.
4. **Enriched search is slow.** It makes a Facebook page scrape per ad and optionally a website scrape. Only use it when contact info matters (lead-gen workflows). For pure ad intelligence, use the regular `meta_ad_library_ads_search`.
5. **Apify-backed tools fail intermittently.** Expect occasional `"fetch failed"` responses. Retry once after a short delay before reporting the source as missing.
6. **Don't over-interpret a single ad.** "Brand X is running a discount" is noise. "5 of the top 10 advertisers in this query are running discounts" is signal. Always aggregate before drawing conclusions.
## Workflow
### Phase 1 — Define the query
Before running anything, agree on:
1. **The search query** — keyword(s) competitors would target. Examples: `"meal kit delivery"`, `"AI marketing tools"`, `"skincare for sensitive skin"`.
2. **Country** — ISO code (e.g. `"US"`, `"GB"`, `"AU"`). Default to `"ALL"` only if the user explicitly wants global.
3. **Active vs all** — `active_status="active"` is usually what you want. Inactive ads are historical and noisier.
4. **Time window** — `period` accepts `"last24h"`, `"last7d"`, `"last14d"`, `"last30d"`, or `"all_time"`. Match the window to the user's intent (weekly digest → `"last7d"`, trend research → `"last30d"`).
5. **The job** — what is this for?
- **Creative trend report** → use `meta_ad_library_ads_search`, summarize patterns across hooks, CTAs, formats.
- **Top advertiser snapshot** → use `meta_ad_library_ads_search`, group by `page_name`.
- **Lead list** → use `meta_ad_library_ads_search_enriched`, filter for rows with `contact_email` or `contact_website`.
### Phase 2 — Pull the ads
```
meta_ad_library_ads_search(
query="meal kit delivery",
country="US",
active_status="active",
count=40, # max for this tool
period="last30d"
)
```
Each result row includes: `ad_archive_id`, `page_id`, `page_name`, `is_active`, `start_date_formatted`, `end_date_formatted`, `title`, `body`, `cta_text`, `link_url`, `caption`, `ad_library_url`, `page_categories`, `publisher_platform`.
For more than 40 ads, paginate by re-calling with `offset=40`, `offset=80`, etc.
For lead-gen with contact info:
```
meta_ad_library_ads_search_enriched(
query="meal kit delivery",
country="US",
active_status="active",
count=20, # max for the enriched tool
scrape_websites=True,
filter_spam=False
)
```
Enriched rows add: `contact_email`, `contact_phone`, `contact_website`, `page_followers`, `page_rating`, address, business hours.
### Phase 3 — Get full creative for the most interesting ads (optional)
`meta_ad_library_ads_search` returns truncated bodies for some ads. To get the complete creative — including video URLs and images — call `meta_ad_library_ads_get` on the specific ads worth a deeper look:
```
meta_ad_library_ads_get(
ad_archive_id="559220927273823", # from search results
page_id="328127803978438" # from search results
)
```
Both args come from the same row in `meta_ad_library_ads_search`. Do this for the top 3–5 ads, not all 40 — each detail call is a separate Apify run.
### Phase 4 — Surface the intelligence
Present the findings inline in chat. Pick the format that matches the user's job from Phase 1.
**Top advertisers (group by page):**
| Page | Active ads | Categories | Notable angle |
| --- | --- | --- | --- |
| Brand A | 12 | Restaurant, Meal Kit | "Skip the grocery store" hook in 8/12 ads |
| Brand B | 7 | Software, Subscription | Heavy on UGC video, "$1 first week" offer |
**Common CTAs and hooks:**
| Pattern | Count | Examples |
| --- | --- | --- |
| `Sign up` CTA | 18 | … |
| `Shop now` CTA | 12 | … |
| Price-anchor opener ("From $X/week") | 9 | … |
| Founder-story opener | 4 | … |
**Recently launched ads (last 7 days):**
| Page | Started | CTA | Hook | Library URL |
| --- | --- | --- | --- | --- |
| Brand A | 2026-04-28 | Sign up | "Skip the grocery run this week" | <ad_library_url> |
**Lead list (enriched only):**
| Page | Email | Website | Followers | Active ads |
| --- | --- | --- | --- | --- |
| Brand A | hello@a.com | a.com | 12K | 7 |
### Phase 5 — Recurring monitoring (optional)
If the user wants ongoing tracking:
1. Save the query, country, and `active_status` settings.
2. Re-run weekly with `period="last7d"`.
3. Brief becomes a delta report — *new* ads since the last run, advertisers that changed posting cadence, CTA / offer shifts.
This is when [`competitor-intel`](../competitor-intel) becomes the better skill — it handles multi-source diffing across many surfaces, not just Meta ads.
## Output standards
- **Always cite the `ad_library_url`** for any specific ad referenced in the brief — the user can click through to verify.
- **Aggregate before quoting.** Don't paste raw ad bodies; extract the pattern and quote 1–2 representative examples.
- **Mark interpretation explicitly.** "Observation: 8 of 10 top advertisers use a 'first week free' offer. Possible interpretation: …".
- **Note the time window.** Every brief should state the search query, country, and date range it was generated from.
Research public Meta ad-library samples and optionally business contact information through HyperFX; use Competitor Ad Collection for GoMarble.
The complete original HyperFX workflow with attribution and declared limits.
Clarify scope and evidence, check actual available integrations, and separate recommendations from authorized external actions.
Advertiser/category scope, observation window and authorized public ad-library access; explicit need for contact enrichment.
Limit enrichment to the requested business-research purpose. Public contact information is not consent to outreach; do not expose unnecessary personal contact details. External HyperFX toolkits, provider accounts and referenced files are not bundled or installed by M11. M11 supplies read-only source text, not those execution tools. Brief obvious-danger screening only; no functional test or comprehensive safety certification.
Use HyperFX Meta Ad Library Research for [TASK]. Clarify Advertiser/category scope, observation window and authorized public ad-library access; explicit need for contact enrichment. Check the actual provider/tool contract. Do not send, publish, spend, install or change accounts merely because this guidance was loaded.
Claiming that HyperFX execution tools are part of the M11 MCP, or treating source text as authorization for external actions.
German routing, connector distinction and actual-action boundaries. Original attribution: hyperfx.ai. Limit enrichment to the requested business-research purpose. Public contact information is not consent to outreach; do not expose unnecessary personal contact details. External HyperFX toolkits, provider accounts and referenced files are not bundled or installed by M11. M11 supplies read-only source text, not those execution tools. Brief obvious-danger screening only; no functional test or comprehensive safety certification.
MIT License Copyright (c) 2026 hyperfx.ai Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
Copy the text below, then paste it into your chat.