clawnote_skill

Overview

This skill provides a review-first workflow for creating and managing Xiaohongshu (Rednote) content inside OpenClaw. It organizes topic research, draft generation, Feishu-based review, exact-title approval, and structured publish packages. The default behavior favors human confirmation and explicit publish approval rather than automatic posting.

How this skill works

The skill sequences three main stages: research, draft creation, and publish assistance. Scripts build and save publish packages, capture memory entries, and optionally publish only after explicit approval via Feishu. Safety checks require exact-title confirmation and separate handling for deletion or archive-to-draft operations.

When to use it

• Building repeatable Xiaohongshu content workflows that require human review
• Daily AI news or OpenClaw practice drafts needing structured approval
• Coordinating Feishu-first review and feedback loops with editors
• Creating publish packages that must be validated before posting
• Running cron-driven content ops with memory-backed state

Best practices

• Follow the ordered workflow: research → draft → publish assist to keep iterations predictable
• Always require explicit user confirmation before publishing; do not auto-post
• Use provided scripts for package build/save/publish to avoid reimplementation errors
• Prefer exact-match title approval instead of fuzzy matching for publish safety
• Treat deletion as high-risk and handle it separately from normal publish steps

Example use cases

• Run a morning cron that compiles topic research, drafts a note, and queues it for Feishu review
• Use the draft stage to rewrite content into Xiaohongshu style before presenting to editors
• Save validated publish packages for scheduled posting after legal or brand sign-off
• Capture memory entries about past notes to inform future topic selection
• Approve a specific note title in Feishu and then trigger the explicit publish script