# xiaohongshu-skill full LLM reference

Project:

- Name: xiaohongshu-skill
- Category: AgentSkill and Python Playwright CLI
- Domain: Xiaohongshu, RedNote, 小红书, xhs, RED
- Repository: https://github.com/DeliciousBuding/xiaohongshu-skill
- Pages: https://deliciousbuding.github.io/xiaohongshu-skill/
- skills.sh: https://www.skills.sh/deliciousbuding/xiaohongshu-skill
- License: MIT

One-sentence summary:

xiaohongshu-skill lets AI agents search Xiaohongshu and RedNote, read note details, prepare publish workflows, inspect selectors, and return JSON through a Python Playwright browser client.

Install commands:

```bash
npx skills add DeliciousBuding/xiaohongshu-skill
clawhub install xiaohongshu-skill
pip install git+https://github.com/DeliciousBuding/xiaohongshu-skill.git
playwright install chromium
xiaohongshu-skill qrcode --headless=false
```

Agent platforms:

- Claude Code
- Codex
- OpenClaw
- ClawHub
- skills.sh
- Hermes Agent
- AgentSkills-compatible runners
- Shell scripts, Cursor, Cline, n8n, and CI jobs that can call a local CLI

Trigger phrases:

- "帮我搜下小红书"
- "看看小红书上这个帖子"
- "发到小红书"
- "小红书上怎么说的"
- "find RedNote posts about this"
- "search Xiaohongshu for..."
- "xhs note analysis"
- "RedNote automation"
- "小红书运营"

Read-only commands:

```bash
xiaohongshu-skill check-login
xiaohongshu-skill profiles
xiaohongshu-skill selectors
xiaohongshu-skill selectors --owner=publish
xiaohongshu-skill contracts
xiaohongshu-skill contracts --command=search
xiaohongshu-skill search "关键词" --limit=5
xiaohongshu-skill feed <id> <xsec_token>
xiaohongshu-skill feed <id> <xsec_token> --load-comments --max-comments=20
xiaohongshu-skill user <user_id> [xsec_token]
xiaohongshu-skill me
xiaohongshu-skill explore --limit=10
```

Write commands that require user approval:

```bash
xiaohongshu-skill publish --title="标题" --content="正文" --images="img.jpg"
xiaohongshu-skill publish-video --title="标题" --content="正文" --video="video.mp4"
xiaohongshu-skill publish-md --title="标题" --file=article.md
xiaohongshu-skill publish-longform --title="标题" --content="正文"
xiaohongshu-skill comment <id> <token> --content="内容"
xiaohongshu-skill reply <id> <token> --comment-id=<cid> --reply-user-id=<uid> --content="内容"
xiaohongshu-skill like <id> <token>
xiaohongshu-skill collect <id> <token>
```

Safety rules:

- Ask the user before publishing, commenting, replying, liking, collecting, or scheduling.
- Stop when Xiaohongshu shows captcha, QR login, or verification.
- Use headed mode for login and manual verification.
- Keep pacing enabled.
- Use test accounts for live tests.
- Do not expose cookies, QR codes, profile links, screenshots, account names, or local paths in public issues.

Output contract:

- Commands return JSON where possible.
- `search` returns `count` and `results`.
- `publish` returns status fields and stops before final publish unless the user approved the action.
- `selectors` and `contracts` expose stable data for agent wiring.

Public docs:

- Install: https://github.com/DeliciousBuding/xiaohongshu-skill/blob/main/docs/INSTALL.md
- Commands: https://github.com/DeliciousBuding/xiaohongshu-skill/blob/main/docs/API.md
- Integrations: https://github.com/DeliciousBuding/xiaohongshu-skill/blob/main/docs/INTEGRATIONS.md
- Security: https://github.com/DeliciousBuding/xiaohongshu-skill/blob/main/docs/SECURITY.md
- Reference map: https://github.com/DeliciousBuding/xiaohongshu-skill/blob/main/docs/REFERENCE.md
- Roadmap: https://github.com/DeliciousBuding/xiaohongshu-skill/blob/main/docs/ROADMAP.md
- Examples: https://github.com/DeliciousBuding/xiaohongshu-skill/tree/main/examples
- Demo outputs: https://deliciousbuding.github.io/xiaohongshu-skill/demo.html

Search and catalog terms:

Xiaohongshu Skill, RedNote Skill, 小红书 Skill, xiaohongshu-skill, xhs automation, 小红书 AI Agent, AI Agent Skill, AgentSkill, Python Playwright, browser automation, Claude Code Skill, Codex Skill, OpenClaw Skill, ClawHub, skills.sh.