Proven Xiaohongshu Title Formulas for Agents

Where this fits

Imagine a developer building a Xiaohongshu content workflow for a client who posts lifestyle articles. Every request is "帮我起个小红书标题," and a generic LLM invents a catchy title with no reasoning, no consistency, and no way to defend it. This skill plugs into that workflow as a formula matcher: it picks from 75 proven title formulas, tags each generated title with a formula ID, and explains why that formula fits—while deflecting video-title requests to ra-video-title.

Why agents benefit

• Every title is traceable. The agent can cite a formula number and psychological trigger, so output is auditable instead of a lucky guess.
• Boundary enforcement built in. The skill explicitly refuses video-title scenarios, preventing the agent from silently applying the wrong methodology.
• 12 trigger categories—curiosity gap, loss aversion, identity, and more—let the agent choose based on content type, not just keyword overlap.
• Works from Chinese or English triggers, so the same agent can serve multilingual production teams.
• Installation is standard and composable with other skills in the rnskill repo.

Practical scenarios

Breaking a misconception in a wellness article

A post debunks "drinking 8 glasses a day is mandatory." The agent picks a cognitive-conflict formula (IDs 1–6), generating a title that challenges the reader's existing belief while staying accurate.

Building a numbered travel checklist

For "10 things to prep before a solo trip," the agent uses a number-anchoring formula (IDs 26–32). The title reduces cognitive load and sets a clear expectation for the reader.

Targeting a specific audience for a product review

A skincare review for "sensitive skin in humid climates" maps to an identity formula (IDs 21–25). The title speaks directly to that niche, improving resonance and click-through.

Add it to your agent workflow

Install the skill with:

npx skills add Pluviobyte/rnskill --skill dbs-xhs-title

Then trigger it with /dbs-xhs-title or a natural-language request. For example, after receiving the article's key points, the agent responds with three title candidates, each tagged with a formula ID and a one-line rationale: "ID 18 (loss aversion)—because the audience is looking for pitfalls to avoid."

Compare before adopting

Before adding it, compare against a plain prompt or a generic title generator. Check three things:

• Quality signals: 1.4k GitHub stars and a maintained repo say the formula library is battle-tested, but verify recent commits and sample outputs.
• Alternatives: If your workflow is video-first, ra-video-title is the right fit; dbs-xhs-title is deliberately narrow.
• Workflow fit: Confirm the formula IDs and Chinese/English triggers match how your agent and users communicate.

Why it is worth tracking

This skill represents a useful pattern: domain-specific formula libraries that give agents constraints and explainability instead of open-ended generation. The formula collection is updated with community feedback, and the clear boundary between XHS article titles and video titles shows mature scoping. If you build content automation for Chinese social platforms, track this repo and evaluate it whenever your title quality plateaus. For a curated view of this skill and related design-creative tools, see its listing at https://www.openagentskill.com/skills/pluviobyte-dbs-xhs-title.