Amazon Review Insights

Problem

Amazon review data for an ASIN is large and fragmented. Operators often have to open pages manually, filter negative reviews, and summarize recurring issues to understand product sentiment. This skill connects that workflow to AstrMap: it creates a task with an ASIN and marketplace site, lets an online collection endpoint fetch reviews, and then surfaces AI-generated insights, statistics, trends, and representative comments.

How it works

The skill requires CUSTOMER_INSIGHTS_API_KEY and sends requests to the fixed endpoint api.astrmap.com. Querying completed results only needs the API key and actions such as get_ai_insights, get_negative_reviews, and get_trend. Creating new tasks requires an online collection channel; the AstrMap desktop client is recommended, and check_device should return is_alive=true before task creation.

Tasks can run in two modes: --is-auto true continues from collection into PROCESSING and ANALYZING, while --is-auto false stops at COLLECTED. In that state, trigger_analysis can manually start analysis. Task creation and incremental fetching consume points, so the workflow should check balance and wait for confirmation before execution.

Boundaries

This skill fits teams that already maintain an AstrMap account and an online collection device. Browser extensions can also collect, but they depend on local browser state, Amazon login, and proxy configuration, which the Agent cannot reliably verify. Querying existing results does not consume points, but new tasks, incremental fetches, and manual analysis may. Errors such as insufficient points, invalid API keys, or non-COLLECTED task status must be handled explicitly.