xhs-content-ops_skill

Overview

This skill is a compound content operations assistant for Xiaohongshu (RED). It combines search, detail retrieval, publishing, and interaction actions into end-to-end workflows for competitor analysis, trend tracking, content creation, and interaction management. Each workflow reports progress step-by-step and requires explicit user confirmation for publishing or posting comments.

How this skill works

The skill drives the project's CLI to run only approved subcommands (search-feeds, get-feed-detail, publish, post-comment, like-feed, favorite-feed, etc.). For a given task it selects the right workflow, executes searches, fetches feed details, synthesizes findings into structured reports (tables for comparisons), and pauses for user confirmation before any write action. It enforces rate and sequence constraints, spacing browsing and interactions to reduce limits.

When to use it

• Requesting competitor analysis or side-by-side note comparisons for a topic.
• Tracking hot topics and trend changes across recent and high-engagement posts.
• Assisting with researching, drafting, and publishing a Xiaohongshu note.
• Managing interactions: targeted commenting, liking, and favoriting with strategy.
• When you need step-by-step progress updates and confirmation before actions.

Best practices

• Provide clear targets: keywords and competitor accounts for precise searches.
• Limit write actions to confirmed steps; review generated drafts before publish.
• Use both 'latest' and 'most likes' search modes to capture trends and hits.
• Select 3–5 reference notes for analysis to form reliable content patterns.
• Keep interaction cadence: 30–60 seconds between actions and <20 comments/day.

Example use cases

• Competitor report: search a keyword, fetch 3–5 top notes, output a table comparing titles, cover style, structure, tags, and engagement.
• Trend scan: daily run on a set of keywords to rank heat, summarize recurring formats, and recommend 5 potential topics.
• Content creation: analyze reference notes, generate title (≤20 UTF-16 chars), body draft, and tags; ask user to confirm then publish via CLI.
• Interaction campaign: find medium-engagement notes in a niche, suggest comment drafts, ask approval, then post comments and optionally like/favorite with spacing.
• Failure handling routine: expand keywords on no-results, retry detail fetches for deleted notes, and increase spacing when rate limits hit.