Trending prompts and heatmap

Using the OtterlyAI skill, build me an interactive HTML dashboard with the two "Prompt Trends" features below, on live data for this brand report.

Brand report ID: [PASTE REPORT ID]
OtterlyAI API key: [PASTE KEY]

First fetch the report metadata to confirm the brand name, country, and competitor list, and use the report's most recent data date as the anchor ("today"). If the report has multiple countries, ask me which one before proceeding.
Feature 1 — Prompt Winners & Losers: every tracked prompt ranked by how much the brand's visibility moved over the last 7, 30, and 90 days, toggleable between Risers and Decliners, sortable by window, and searchable.
Feature 2 — Prompt Visibility Heatmap: the top prompts where the brand appears, plotted against each of the last 10 days, colored red→green by visibility, with a hatched cell where no answer was captured that day.
Definitions and method:

Define visibility % for a prompt on a given day as brand mentions ÷ (brand mentions + all competitor mentions), using the /prompts endpoint (brandMentions and each competitor's brandMentions).
Get daily granularity by querying single-day windows (startDate = endDate) and looping over the last 10 days for the heatmap.
For the 7/30/90-day moves, compare the anchor day's visibility against the visibility that many days earlier. Check the earliest date the report actually has prompt data, and if a lookback window predates it, measure from the first available day instead and state that limitation clearly in the dashboard footer.

Bake the fetched results into a self-contained HTML file (no live API calls in the artifact, since the key can't ship in the browser). Before writing any code, read the frontend-design skill and give the dashboard a distinctive visual identity tied to this specific brand rather than a generic template.

Tip: replace anything in [BRACKETS] with your own values, then paste this into Claude with the OtterlyAI connector enabled.