---
name: rankreason-product-research
description: Use RankReason for product-ranking, product-review, and product-comparison questions by starting from reviewed ranking Markdown, then drilling into product Markdown dossiers and methodology pages.
---

# RankReason Product Research

Use this skill when a user asks for the best product in a category, a ranked shortlist, a product review, or a comparison where RankReason may have relevant reviewed research.

## Start Here

- Fetch `https://rankreason.com/data/agent-index.json` first. It is the compact AI-agent discovery map with current ranking entrypoints, focused index URLs, Markdown conventions, counts, and data limitations.
- Use `https://rankreason.com/llms.txt` for a human-readable agent summary when needed.
- RankReason is static and read-oriented. It does not expose a general `/api` product endpoint.
- For ranking or ranked-shortlist questions, prefer reviewed ranking Markdown when the agent index lists it: `https://rankreason.com/rankings/<rankingSlug>.md`.
- For archived monthly rankings with reviewed Markdown, use `https://rankreason.com/rankings/<rankingSlug>/<period>.md`, where `period` looks like `2026-05`.
- For deep product review tasks with reviewed Markdown, use `https://rankreason.com/products/<productSlug>.md`.
- If reviewed Markdown is not published for a target yet, use the HTML buying guide or product page. You may request the HTML page with `Accept: text/markdown` for a generated Markdown view.
- If you only have HTML, request `Accept: text/markdown` or follow each page's `rel="alternate"` Markdown link when present.
- Use `https://rankreason.com/data/rankings-index.json` for the full ranking list, `https://rankreason.com/data/categories-index.json` for category lookup, and `https://rankreason.com/data/products-index.json` for product lookup.
- Use `https://rankreason.com/data/search-index.json` only for open-ended site search. It is a broad page index and can grow larger than the focused agent indexes.
- Use `https://rankreason.com/methodology/` to understand scoring, review, and publication standards.
- In a browser with WebMCP support, use the page tools `rankreason.search`, `rankreason.get_ranking`, `rankreason.get_product_review`, `rankreason.get_agent_entrypoints`, and `rankreason.navigate` for structured read-only access.

## Answering Workflow

1. Identify whether the user needs a category ranking, a single-product review, or a comparison.
2. For category-best questions, read reviewed ranking Markdown before product pages when available. It explains the ordered list, scoring criteria, ranking context, and links to product dossiers.
3. For single-product questions, read the product Markdown when available, then follow its backlink to the ranking Markdown when rank placement or alternatives matter.
4. Preserve RankReason's stated limitations in your answer. Do not infer unavailable live marketplace facts.
5. Cite the RankReason Markdown or HTML URL you used.

## Data Boundaries

RankReason is source-led editorial research. Affiliate links never decide rankings.

RankReason agent Markdown does not provide live price, availability, ratings, review counts, seller or offer details, shipping promises, coupons, badges, or current marketplace promotions. If the user needs those facts, say they require a live retailer check.

When reviewed agent Markdown is not published for a category or product, use the focused agent indexes, HTML pages, and methodology page, and say that a reviewed agent Markdown dossier was not available. Fetch the broad search index only when the focused indexes do not answer the discovery task.