← Back to Skills Marketplace
329
Downloads
1
Stars
0
Active Installs
11
Versions
Install in OpenClaw
/install xhs-ts
Description
Automate Xiaohongshu (小红书/RedNote) operations via Playwright CLI — search notes, publish image/video posts, interact (like/collect/comment/follow), scrape da...
Usage Guidance
This skill appears to be what it claims: a Playwright-based Xiaohongshu automation CLI. Before installing, consider: 1) It manages and stores cookies, fingerprints, and profile data under the skill directory (users/{user}/), and cleanup operations can delete those files — back up any important account data. 2) Automation and anti-detection techniques increase the risk of account suspension or TOS violations on Xiaohongshu; use throwaway or test accounts if you need to experiment. 3) Playwright will install browser engines and the skill may run browsers with --no-sandbox or other flags when run as root or inside containers — running unprivileged is safer. 4) The .env supports optional PROXY settings; if you plan to use a proxy, ensure you trust the proxy provider. 5) If you must be certain there is no hidden data exfiltration, review the code paths that save/load profile.json, cookies.json, and any network requests the Playwright browser will make (the automation interacts with the website via the browser). Overall the package is coherent and proportional to its stated purpose; treat it like any browser automation tool and protect stored cookies/credentials accordingly.
Capability Analysis
Type: OpenClaw Skill
Name: xhs-ts
Version: 0.1.1
The xhs-ts skill bundle is a comprehensive automation framework for Xiaohongshu (RedNote) using Playwright. It features robust multi-user profile management, anti-detection mechanisms (stealth scripts in scripts/core/browser/stealth/), and structured data extraction. The code logic across the actions/ and core/ directories is strictly aligned with the stated automation purposes (search, publish, interact, scrape). No evidence of data exfiltration, malicious remote execution, or prompt injection was found; sensitive data like cookies and fingerprints are stored locally within the users/ directory as intended for session persistence.
Capability Tags
Capability Assessment
Purpose & Capability
Name/description match the code and runtime instructions: the project is a Playwright CLI that manages multiple Xiaohongshu accounts, handles cookies, fingerprints, anti-detection, publishing, scraping, and interaction. Required binaries (node, npx) and listed node packages (playwright, tsx, commander, dotenv) are appropriate and expected for this functionality.
Instruction Scope
SKILL.md tells the agent to run the CLI (npm run <command>) and to prompt the user before destructive actions. The codebase contains migration and cleanup routines that move/copy/delete files under the skill directory (cookies.json, users/*, tmp/). This behavior is coherent for multi-account management but means the skill will read and write local files (cookies, fingerprints, profile.json) and may remove user-data when cleanup is performed — the SKILL.md documents prompting the user before cleanup.
Install Mechanism
Install spec uses node package installation (playwright, tsx, commander, dotenv). No remote arbitrary downloads or obscure URLs are used in the install metadata; Playwright will install browser engines (standard behavior). This is a typical and proportionate install mechanism for a Node/Playwright CLI.
Credentials
No secrets or external credentials are requested. The code detects environment (DISPLAY, container markers, /proc/1/cgroup) and decides flags like --no-sandbox or --disable-dev-shm-usage accordingly — these checks are reasonable for running headless browsers in containers or as root. The .env template contains optional PROXY/HEADLESS/BROWSER_PATH settings, which match expected needs. Data written (cookies, fingerprints, profile.json) are expected for browser-based automation and are stored under the skill's users/ directory.
Persistence & Privilege
The skill is not always:true and does not request system-wide persistence beyond its own project directory. It stores profiles, cookies, and fingerprints under users/, and may save browser connection info in profile.json; these are normal internal files for this functionality. It does not modify other skills' configs or ask for broad platform privileges.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install xhs-ts - After installation, invoke the skill by name or use
/xhs-ts - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.1
xhs-ts v0.1.1
- Improved documentation for error recovery and multi-user management, including steps to clean up corrupted user data and edge cases when user data is locked.
- Added new login methods to documentation: SMS login with phone, and cookie string direct login.
- Updated and clarified option tables for all commands (login, search, publish, user management).
- Expanded gotchas and agent workflow instructions to improve reliability and user guidance.
- Minor corrections to parameter defaults and CLI usage explanations.
v0.1.0
xhs-ts 0.1.0 is a major refactor with new file structure, user data model, and enhanced commands.
- Fully reorganized file structure and codebase for better scalability and maintainability
- Multi-account data storage updated: cookies and meta merged into unified user-data and profile file layouts (version 3)
- Added new browser management commands: `browser --start`, `browser --status`, `browser --stop`
- All scrape commands split to `scrape-note` and `scrape-user` scripts
- SKILL.md and user-facing docs updated to reflect new commands, options, and directory structure
- Many obsolete files and logic removed; all core features re-implemented under the new structure
v0.0.9
**Enhanced anti-detection and browser simulation capabilities in 0.0.9.**
- Added mouse trajectory simulation and scroll physics for improved anti-detection.
- Introduced browser fingerprint presets and user fingerprint management.
- Refactored and updated multiple scripts for robustness and realism.
- Expanded helper utilities for probability and anti-detect features.
- Updated documentation and configuration for new features.
v0.0.8
xhs-ts v0.0.8
- Bump version number to 0.0.8 in SKILL.md and package.json
- Minor documentation updates in SKILL.md: add/clarify anti-detection info, update formatting, fix metadata version
- No changes to core features or commands
v0.0.7
xhs-ts v0.0.7
- Refactored authentication and login detection logic for better reliability and maintainability.
- Introduced modular extraction and navigation utilities for the search functionality.
- Updated and expanded documentation on channel integration and output formats (Feishu, WeChat).
- Added new scripts for authentication and search data extraction.
- Removed deprecated scripts related to authentication checks.
- Incremented version to 0.0.7 and refreshed usage documentation.
v0.0.6
**Major update: Adds full interact (collect, comment, follow) and scraping capability to Xiaohongshu automation.**
- Added collect, comment, and follow commands for note/user interaction.
- Introduced scraping commands to extract note and user data, including support for comments and recent notes.
- Expanded and updated output formatting and error/info messages, especially for actions requiring valid URLs and phone binding for comments.
- Enhanced anti-detection measures (random delays, mouse movement, rate limiting, captcha handling).
- Documentation and help texts revised to cover new commands, usage, and requirements.
- Internal structure changes: new scripts added for each feature; existing browser/session/interaction logic improved and refactored.
v0.0.5
**Adds "like" automation for Xiaohongshu notes, including batch support.**
- New `like` command: automate liking single or multiple notes by URL (requires valid `xsec_token` in URL; batch input supported).
- Added new internal modules for interaction automation (`interact/`).
- Updated docs and CLI for new batch "like" usage details and limitations.
- Added `.env` for environment configuration.
- Removed legacy `users/users.json` (user management now fully handled via new structure).
- Minor type and utility improvements for authentication and CLI.
v0.0.4
xhs-ts 0.0.4 — Multi-account (user) management and migration
- Adds full multi-user support: isolated cookies, temp files, and per-user directories (--user <name>)
- Introduces `users/` directory with auto-migration from old `cookies.json` and `tmp/` on upgrade
- Implements new `user` management commands for listing, switching, and setting default/current user
- All CLI commands (login, search, publish, etc.) now accept `--user <name>` argument for account selection
- SKILL.md, CLI help, and docs updated to reflect new multi-user workflow and migration steps
v0.0.3
**Summary:**
This version improves agent integration, refines the output protocol, and adds documentation for multi-channel adaptation.
- toAgent output protocol expanded with clearer agent handling instructions and channel adaptation requirements.
- Output format descriptions updated and documented for agent developers in [Channel Integration Guide](references/channel-integration.md).
- Response example and field reference for toAgent now emphasize dynamic field substitution and channel-specific message formatting.
- Improved documentation clarity in SKILL.md; added reference to new channel-integration documentation.
- No breaking changes to task commands, core automation, or error codes.
v0.0.2
xhs-ts 0.0.2
- Version updated to 0.0.2
- Added README.md file
- Improved and updated documentation in SKILL.md
- Multiple core script updates for browser stealth, main entrypoint, and search result extraction
- Updated package.json dependencies and metadata
v0.0.1
Initial release of Xiaohongshu (xhs-ts) automation skill.
- Supports searching notes, publishing posts (image/video), and data scraping for Xiaohongshu.
- Implements login (QR code, SMS, creator mode), search (with filtering and sorting), and publish commands.
- Outputs all responses in standardized JSON format for agent handling, with clear success/error formats and actionable instructions.
- Includes detailed error codes, anti-detection tips, and best-practice recommendations.
- Notifies users about not implemented commands (like, collect, comment, follow, scrape).
- Provides comprehensive documentation in SKILL.md for setup, usage, and troubleshooting.
Metadata
Frequently Asked Questions
What is xhs 多用户?
Automate Xiaohongshu (小红书/RedNote) operations via Playwright CLI — search notes, publish image/video posts, interact (like/collect/comment/follow), scrape da... It is an AI Agent Skill for Claude Code / OpenClaw, with 329 downloads so far.
How do I install xhs 多用户?
Run "/install xhs-ts" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is xhs 多用户 free?
Yes, xhs 多用户 is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does xhs 多用户 support?
xhs 多用户 is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created xhs 多用户?
It is built and maintained by 海刚 (@lv-saharan); the current version is v0.1.1.
More Skills