Weibo Content Research

Problem

When researching Weibo topics or comparing competitor content, raw hot-search and keyword search results are scattered. Manual collection makes it hard to build a reviewable basis for judgment. This skill turns Weibo content research into a repeatable read-only workflow: given a keyword or topic direction, it returns ranked signals, heat indicators, titles, authors, links, or content IDs for follow-up analysis.

How it works

It prefers the SocialDataX direct CLI and also maps to MCP tools weibo_get_hot_search_list and weibo_search_posts. Hot-search uses weibo hot-search without --keyword; keyword search uses weibo search --keyword, with --max-items to cap sample size and --pages plus --page-token to continue the same pagination chain. Responses are JSON. Multi-page search merges items into data.items, adding page_count, item_count, and next-page markers.

Output should separate ranking signals, related samples, and next follow-up angles. When traceability matters, preserve post_id, post_url, author, interaction metrics, and publish time.

Boundaries

This is a read-only capability: it does not log in to Weibo, post, like, comment, modify accounts, or read local browser data. It relies on the SOCIALDATAX_API_KEY environment variable. If the error is insufficient balance, follow the returned recharge link rather than retrying blindly. Search has no stable complete boundary and does not support --all; an empty items page does not by itself mean there are no more results.