xiaohongshu

小红书 (Xiaohongshu / RedNote) Posting Skill

Setup

Installation

If you haven't installed md2red yet, clone it into the xiaohongshu folder (parallel to this skill file) and install its dependencies:

cd /path/to/xiaohongshu
git clone https://github.com/nowork-studio/md2red
cd md2red
pip install -r requirements.txt

Configuration

Configure these before first use:

• md2red script path: path to your md2red.py install (see md2red)
• Browser: profile=openclaw only — already logged in. Never use Chrome relay (CDP disconnects on Xiaohongshu page navigation)
• Content plan: a markdown file in your workspace (e.g. xiaohongshu-content-plan.md)
• Account profile ID: your Xiaohongshu profile ID (visible in profile URL)

【1. Publishing a New Post — Publishing a Post】

Step 1 — Pick a topic (Pick a topic)

Read your content plan file, find the post scheduled for this time slot.

Step 2 — Write markdown (Write content)

• Short paragraphs, bold key sentences, natural call-to-action at the end
• ⚠️ No #tags in the markdown file — md2red renders markdown as images; hashtags will appear inside the image
• Tags go only in Step 7's body innerHTML

Step 3 — Generate images (Generate images)

Save your markdown to the md2red input file, then run:

cd /path/to/xiaohongshu/md2red && python3 md2red.py input.md --output <name>

Images output to output/<name>/ (1.png, 2.png, …)

Step 4 — Open upload page (Open upload page)

browser navigate profile=openclaw targetUrl=https://creator.xiaohongshu.com/publish/publish?from=tab_switch&target=image

Wait 3 seconds.

Step 5 — Upload images (Upload images)

• Take a snapshot → find the "上传图片" (Upload Image) button ref
• browser upload ref=<button-ref> paths=[1.png, 2.png, ...]
• ⚠️ Do not use selector="input[type=file]" — it's hidden and won't trigger the upload. Must use the button ref so Playwright intercepts the fileChooser event correctly
• Wait 5 seconds

Step 6 — Fill in title (Fill in title)

• Snapshot → find the title textbox → click + type
• ≤20 characters

Step 7 — Fill in body (Fill in body) — must use JS evaluate, not type

(function () {
  var el = document.querySelector(".tiptap.ProseMirror");
  if (!el) {
    var eds = document.querySelectorAll("[contenteditable=true]");
    el = eds[eds.length - 1];
  }
  el.focus();
  el.innerHTML = "<p>Your body text here</p><p><br></p><p>#tag1 #tag2</p>";
  el.dispatchEvent(new Event("input", { bubbles: true }));
  return "done";
})();

Step 8 — Publish (Publish)

• Click the "发布" button
• Success indicator: page returns to the initial video upload page

Step 9 — Mark as published (Mark as published)

• Update your content plan file, mark the post as published

---

【2. Engagement — Engagement】

Using profile=openclaw browser:

1. Open https://www.xiaohongshu.com/notification — review all notifications
2. For each comment/reply, write a quality response (show expertise, avoid templated replies)
3. For new follower notifications → visit their profile and follow back if relevant

---

【3. Stats Check — Stats Check】

Using profile=openclaw browser:

1. Open your profile page and note current follower count
2. Check the most recent post's likes and saves
3. Self-review: which content types perform better, how to adjust next time

---

【Report Format — Report Format】

At the end of each session, report:

• Title of post published this session
• Number of comments replied to
• Number of accounts followed back
• Current follower count
• Likes / saves on last post
• Content optimization thoughts for next time

---

⚠️ Content Moderation Guidelines

Xiaohongshu has strict automated review. Adapt these based on your niche, but general principles:

• Avoid sensitive financial/regulatory terms specific to your topic
• Use neutral / professional synonyms instead of platform-specific branded terms
• When in doubt, be vague rather than precise with anything compliance-adjacent
• Tags are also reviewed — avoid anything likely to trigger financial, political, or adult content filters