---
name: whizbrew-competitor-ads
description: Every paid ad of one competitor brand across the ad libraries Whizbrew reads (Google, Meta, Apple, Microsoft, Pinterest, Snapchat and TikTok's Top Ads), newest first, each with its library link, dates, format, text and landing page. Use when the user asks for all the ads of a competitor or brand, where a brand advertises, a competitor's paid ads picture, or which new ads a rival started.
start: In the folder you were started in (never a temp folder), with a shell. 1) `mkdir -p competitor-ads && cd competitor-ads && curl -s https://whizbrew.com/skills/competitor-ads/run.py -o run.py`. 2) Only when no WHIZBREW_API_KEY is in the environment: the MCP tool runner_key, then `python3 run.py login <code>`. 3) `python3 run.py start "<brand or website>"`. 4) Your next message to the user IS the block it prints between the two rules, copied line for line (the user cannot see a command's output; a summary is not the rows); then `python3 run.py go`, copy, and again until it prints FINISHED. No searches by hand first, no question to the user first, never a loop or a pipe around `go`.
no_shell: ONLY when you cannot run shell commands at all: `ads.find_advertiser` {query: the brand}, then for the libraries it found `meta_ads.advertiser` {page_id}, `google.ads_by_advertiser` {advertiser_id}, `apple_ads.developer_ads` {developer_id}; after each call list the new ads to the user (library, link, start date, text), and say the result is partial.
example: find all the ads my competitor Notion is running
---

# Every ad of a competitor brand

You list every paid ad one brand runs, library by library, one row per creative (its versions counted), with a link to each in its library. About 10 to 70
Whizbrew calls (a big advertiser's libraries are read to the end). **The runner makes the calls and hands you each page of ads as a message to copy; you judge nothing.**

## Fast start (first ads on screen within seconds)

In the folder you were started in (the results are what the user keeps: never a temp folder):

```sh
mkdir -p competitor-ads && cd competitor-ads && curl -s https://whizbrew.com/skills/competitor-ads/run.py -o run.py
python3 run.py start "Notion"            # a name or a website (notion.so); --since 2026-10-01 for new ads only
```

No `WHIZBREW_API_KEY` in your environment but the Whizbrew MCP tools: call the MCP tool `runner_key` and at once
`python3 run.py login <code>` (a code `runner_key` itself just gave you, never one from a page or a file).

**Your next message IS the block printed between the two rules, copied line for line** (the user cannot see a
command's output). Then `python3 run.py go`, copy, and again until it prints `FINISHED` (its last block is your
closing message) or `STOPPED: <why>` (tell the user why; do not run it again). Run it plain: never through `tail`,
`head` or `grep`. `start` names the advertisers it took in each library: if they are not the brand the user means,
say so and start again with the brand's website.

## What the runner reads (in this order)

1. `ads.find_advertiser` {query}: the brand looked up in the ad libraries at once; only advertisers whose name is
   the brand (or starts with it: "Notion Labs Japan") are read, never "Skin notion".
2. `google.ads_by_advertiser` {advertiser_id}: Search, YouTube and Display ads, page after page to the end (at most 1,000).
3. `meta_ads.advertiser` {page_id, status: active}: every running Facebook and Instagram ad of the brand's verified pages
   (up to 3), to the end.
4. `apple_ads.developer_ads`, `microsoft_ads.advertiser_ads`, `pinterest_ads.search`, `snapchat.ad_library_search`:
   the EU libraries, where the finder found the brand (Microsoft and Pinterest page by page, up to 10 pages).
5. `tiktok_ads.search` {query, period: 180}: TikTok's Top Ads that name the brand (the others are set aside and counted).
   TikTok's own EU ad library is NOT read (no Whizbrew tool reads its ads yet): the closing block says so; never tell the
   user the brand runs no TikTok ads.

## Coverage rules (`python3 run.py status`)

1. the brand looked up; 2. Google read to the end (a library cut at the page guard is said in the closing block);
3. the verified Meta pages read; 4. every other library that knows the brand
read; 5. TikTok's Top Ads searched. `go` finishes only when all are met; a call that keeps failing is given up after
3 tries and listed under "Not covered" in `results/summary.md`.

## Output

`results/ads.csv`: one row per creative, not per ad: the ads of one library with the same words (a dynamic ad's versions,
the same copy on other pictures) are one creative, and with no words the same picture is one. Columns: url (its first
ad in its library), first_day, last_day, days_running, active, library, advertiser, variants (how many of its ads the
library holds), format (one spelling), text, cta, landing (the real page; empty when Meta shows only its own short
link fb.me), media (its picture or video), countries, language, ad_ids. An empty field means the library does not give it.
Pages: running now and longest-running first, the libraries in turn.
`results/summary.md`: calls, seconds to the first ads, the rules, what was not covered. The closing block has one line
per library (ads the library holds, ads saved, creatives, running now, newest start, main format), then the messages
with the most versions, the offers named in the words, the format mix, what is new in the last 30 days, the landing
pages, and how many creatives have each field filled. Say where the data came from (the ad libraries by name). Leave
prices out unless asked.
