agentleFS
eunomia-bpfeunomia-bpf/eunomia.dev/.agents/skills/xiaohongshu-publisher/SKILL.md

xiaohongshu-publisher

Skill244 starsChanged 5 days agoMITView on GitHub

Prepare or publish eunomia.dev content for Xiaohongshu posts and visual notes. Use when asked to draft, adapt, paste, publish, QA, or record Xiaohongshu content for blog posts, tutorials, GitHub projects, papers, demos, talks, or public technical updates. Covers platform-native title/cover framing, carousel or single-note structure, browser-only observation scans, publication authorization, and media ledger updates.

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

11 rules. 0 are common, 7 are distinctive, and 4 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 7

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    If the source path is missing, inspect .github/publisher/posts_queue.txt, .github/publisher/media/not-published.md, recent …

    line 20
  2. Distinctive:

    Use a browser surface with the logged-in session when UI work is required.

    line 29
  3. Distinctive:

    Never bypass authentication with search results or alternate sources.

    line 30
  4. Distinctive:

    Do not directly access Xiaohongshu APIs, internal endpoints, background HTTP interfaces, or scraped datasets.

    line 40
  5. Distinctive:

    Read the source and identify the reader problem, visible artifact, practical takeaway, proof, and one low-friction next step.

    line 49
  6. Distinctive:

    Run a short platform-observation scan only when the user asks to improve native fit or when the topic has no recent local pattern.

    line 51
  7. Very specific:

    Keep GitHub/eunomia.dev/paper links as sources or next steps after the note already gives useful value.

    line 60
  8. Very specific:

    Before opening the Xiaohongshu editor, write or update the draft under draft/media/YYYY-MM-DD/<source-slug>/xiaohongshu.md using the local …

    line 65
  9. Distinctive:

    no confidential customer, partner, roadmap, pricing, or strategy information appears

    line 80
  10. Distinctive:

    After a confirmed publish, capture the public URL and monitor comments, saves, messages, and repeated reader questions only when the user …

    line 86
  11. Very specific:

    After confirmed publish, update .github/publisher/media/published.md with source path, Xiaohongshu URL, date, format, media, link target, …

    line 94

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Xiaohongshu Publisher

Prepare Xiaohongshu-native posts that make technical work approachable through clear titles, visual evidence, concise Chinese copy, and a useful takeaway. Publication authorization follows CLAUDE.md's Precedence Rule and Publishing section.

Inputs

  • Source Markdown path, project, repo URL, paper, screenshot set, or topic.
  • Intended reader, title direction, media assets, or posting account when specified.
  • Optional canonical link, GitHub link, tags, cover concept, or follow-up plan.

If the source path is missing, inspect .github/publisher/posts_queue.txt, .github/publisher/media/not-published.md, recent docs/blog/posts/, and relevant GitHub repos before asking the user.

Platform Entry Points

Use a browser surface with the logged-in session when UI work is required. Never bypass authentication with search results or alternate sources.

References

Load references/platform-preferences.md when choosing Xiaohongshu-native title, cover, note shape, carousel plan, body length, tags, link treatment, platform-observation scan, or promotion balance.

Browser-Only Platform Boundary

Do not directly access Xiaohongshu APIs, internal endpoints, background HTTP interfaces, or scraped datasets. All observation, drafting, QA, screenshots, and ledger evidence must come from normal browser interactions that a regular logged-in user can perform: navigating pages, searching, scrolling, opening rendered notes, reading visible content, using the editor UI, and capturing screenshots.

Draft Preparation

  1. Read the source and identify the reader problem, visible artifact, practical takeaway, proof, and one low-friction next step.
  2. Run a short platform-observation scan only when the user asks to improve native fit or when the topic has no recent local pattern. Keep it light: record the few observations that will change the draft, and do not copy wording or hooks from other creators.
  3. Choose the format:
    • single note for one sharp observation, screenshot, diagram, or result
    • carousel for a mini walkthrough, comparison, checklist, or debugging story
    • longer body only when the note needs context after the visual hook
  4. Draft from Chinese-native structure, not from a translated LinkedIn post.
  5. Keep GitHub/eunomia.dev/paper links as sources or next steps after the note already gives useful value.

Draft Archive

Before opening the Xiaohongshu editor, write or update the draft under draft/media/YYYY-MM-DD/<source-slug>/xiaohongshu.md using the local date. Include the paste-ready title, body, cover concept, image/carousel plan, source links, tags, first-screen check, browser-observation notes when used, and QA state.

Browser QA

  • title and cover communicate the concrete problem or result at first glance
  • first screen is understandable without clicking an external link
  • carousel order works as a visual story, not just cropped paragraphs
  • images, diagrams, code screenshots, and captions are legible on mobile
  • tags are relevant and minimal
  • account, visibility, and publish settings are intentional
  • no confidential customer, partner, roadmap, pricing, or strategy information appears
  • the final action matches the task: published when authorized, untouched for a draft-only task

Follow-Up

After a confirmed publish, capture the public URL and monitor comments, saves, messages, and repeated reader questions only when the user asks or follow-up was part of the task. Draft replies that add technical context, source links, or concrete next steps. Do not send private messages, likes, follows, comments, or reposts without explicit instruction.

Ledger Update

After confirmed publish, update .github/publisher/media/published.md with source path, Xiaohongshu URL, date, format, media, link target, tags, and follow-up notes. Record useful formatting or audience lessons in this skill or its references so the next Xiaohongshu launch starts from the learned pattern.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.