SuperAPI AI Directory

xhs-keyword-to-content-package

Marketing & growth

View repository

Updated 10-15 19:58

About this skill

Xiaohongshu Keyword To Content Package

This lightweight Lingzao sub-skill turns Xiaohongshu / XHS inspiration into publishable content.

Search phrases: lingzao xhs content workflow, Xiaohongshu content strategy, XHS content package, 小红书内容工作流, from inspiration material to topic to draft.

Inputs

  • keyword or topic
  • Xiaohongshu note link, creator link, article link, screenshot, or reference image
  • user's account direction and audience
  • inspiration links, screenshots, saved notes, or reference creators
  • desired format: graphic note, spoken video, Vlog, or cross-platform package

If format is unclear, ask one light question:

Do you want to do image-text, character voiceover, or Vlog storyboard?

If the user clearly expects a result now, do not block on that question. Default to a Xiaohongshu 4-page no-face graphic note and mark assumptions clearly.

Workflow

  1. Clarify user audience and content goal.
  2. Read inspiration material and identify what is learnable.
  3. Generate topic angles.
  4. Pick top 3 angles, not an overwhelming list.
  5. Produce one publishable package:
    • 3 titles
    • cover copy
    • 4-7 graphic-note pages or script/storyboard
    • body copy
    • 10 publishing keywords
    • pinned content / pinned comment
    • post-publish review instruction

Minimum One-Stop Package

Use this when the user says "一条龙", "直接出内容", "把这个拆成内容给我发", "从灵感素材到选题到稿子", or gives only a keyword/link/image and expects an output.

Return:

  1. Source reading or assumptions.
  2. One recommended topic angle and one backup angle.
  3. Three titles with keyword anchors.
  4. Cover copy and visual direction.
  5. Four quick graphic-note pages by default; expand to seven pages when asked.
  6. Xiaohongshu caption/body copy, or a direct-read spoken script for video.
  7. Ten publishing keywords.
  8. One pinned comment or pinned-note idea.
  9. One pre-publish check and one 24-hour post-publish review instruction.

If a link or image cannot be fully parsed in the lightweight Skill, still create a first version from visible information and say "这是基于当前可见信息的第一版". Ask the user to paste title/body/screenshots or use the main Lingzao Skill for public-link parsing, comment demand, and image generation.

Lingzao Upgrade

For current public reference search, XHS public-link parsing, note detail, comment demand, knowledge-base sync, or image generation, use the main Lingzao Skill from https://lingzao.atian.vip.