Track how competitors refresh Facebook ad creatives.

I want to track how one competitor changed its Meta ads.

Competitor: True Classic
Exact domain: trueclassictees.com
Country (ISO-2): US
Lookback: 30 completed UTC days
Media: Video

Validate only the controlled field values above (lookback 30 or 90, an allowlisted country, media All media, Image or Video, a canonical registrable domain). Do not call Spytrend, browse the web or establish advertiser identity in this step. If a value or field relationship is invalid, return exactly `Invalid controlled input configuration`; otherwise return exactly `Configuration valid`.

Stop after Step 1.

Cost guard: before each paid Spytrend call, run the free `estimate_cost` with the same tool and arguments, estimate the cost and stay within the user's stated budget; if the user has not stated a budget, tell them the estimate before the first paid call. If pricing consent is required, show the returned terms and stop until the customer explicitly agrees; never accept on the customer's behalf. If a required call does not fit the stated budget, make no paid call and return `Budget limit prevents required research`; omit an optional call that does not fit and disclose reduced coverage. When opening ads with `get_ad`, skip multilang ads (`is_multilang`) and never call `get_media` after `get_ad`. Retry once only when `retryable=true`, after `retry_after_seconds`. Read `params.ignored` after every search; an ignored required filter never counts as applied. If a search reports `pagination.total_status=restricted` (a free-plan archive window; `meta.plan_window_cutoff` names the boundary), tell the user the result covers only ads that started on or before that date, and never report a zero from it as "no such ads exist". Cost notices are not part of the final customer report.
Using the validated configuration from Step 1, do not call get_advertiser. Calculate the window: the last 30 completed UTC days ending yesterday, and split it into three consecutive non-overlapping segments. Run exactly three Meta discovery searches sequentially in chronological order, one per segment, with landing_domain=trueclassictees.com, landing_domain_exact=true, countries=[US], media scope Video (media_type=image or video for Image or Video; omit it for All media), sort_by=date, sort_order=asc, dedupe=false, `limit=5` and no pagination. Do not add status_today. Retain full rows, segment provenance, arguments, outcomes, actual debit and remaining budget internally without printing them in progress/commentary. Return only completed/failed segment count, total returned rows before exact-ID deduplication, and whether enough chronological evidence exists. Hide calls, IDs, parameters and audit details.

Stop after Step 2.

Cost guard: before each paid Spytrend call, run the free `estimate_cost` with the same tool and arguments, estimate the cost and stay within the user's stated budget; if the user has not stated a budget, tell them the estimate before the first paid call. If pricing consent is required, show the returned terms and stop until the customer explicitly agrees; never accept on the customer's behalf. If a required call does not fit the stated budget, make no paid call and return `Budget limit prevents required research`; omit an optional call that does not fit and disclose reduced coverage. When opening ads with `get_ad`, skip multilang ads (`is_multilang`) and never call `get_media` after `get_ad`. Retry once only when `retryable=true`, after `retry_after_seconds`. Read `params.ignored` after every search; an ignored required filter never counts as applied. If a search reports `pagination.total_status=restricted` (a free-plan archive window; `meta.plan_window_cutoff` names the boundary), tell the user the result covers only ads that started on or before that date, and never report a zero from it as "no such ads exist". Cost notices are not part of the final customer report.
Using only retained discovery evidence and remaining budget, select up to two non-multilang ads from each non-empty segment, maximum six, so that 3–5 launch cohorts can survive verification. Before opening, take candidates only from the one Meta page_id that appears in the most non-empty segments (ties: most discovery rows), because one domain can run several Meta pages and they must not be mixed. Open sequentially only while each estimate fits. If a later estimate does not fit, stop without substitution and retain reduced coverage. The first opened ad whose original landing domain is exactly trueclassictees.com establishes the Meta page/advertiser; every other verified ad must match it. Verify domain identity, delivery-country evidence, media scope and start date inside its segment. A completed nonqualifying detail is valid exclusion evidence and does not make the run Partial. When media scope is All, a completed detail with no asset type is labelled `Not returned`; missing media or other optional fields do not make a completed run Partial. Build cohorts and compare product/offer, message, actual media, CTA and normalized landing route. Normalize media as Image, Video, Image and video, or Not returned; asset count does not prove carousel. Normalize URLs by lowercasing scheme/host, removing default ports/fragments and only `utm_*`, `fbclid`, `gclid`, `msclkid`, `ttclid`; preserve every other query key/value, including unknown or opaque keys. Different retained query values are distinct route variants but do not prove different audience, offer, funnel, experiment or page content. Do not treat a later launch as a refresh or conflate a cohort with an exact creative family. Retain evidence, outcomes, debit and remaining budget internally without printing progress details. Return only verified-ad count, cohort count, verified-change count and Complete/Partial. Hide IDs, calls and rejected-row details.

Stop after Step 3.

Using retained evidence only, render the final customer-facing dashboard titled `# Meta Creative Changes: True Classic in` followed by the full country name for US (for example United States, not US). Do not rerun Spytrend. Use no more than 300 words. Keep it table-free and chronological. Show up to four verified changes and up to three stable elements. Separate launch timing from current status. Never identify rejected candidates, expose signed media URLs, calls, IDs, quotas, tokens, timing or audit details. Put technical evidence only in a separate `Show audit details` response.

Stop after Step 4.