wechat-article-writer

What to check before installing/use: - Understand file access: the skill will read any files you place in my-articles/ and will create/modify my-style.md and creation-log.md automatically (first-run analysis and ongoing updates). If you keep sensitive or private data in that folder, remove or sanitize it first. Back up important files before first run. - Automatic analysis: the SKILL.md indicates it may automatically analyze up to 3–5 sample articles on first use to generate my-style.md. If you prefer explicit consent, avoid adding files until you confirm the behavior in a controlled test. - Network actions and external tools: Phase 2 (research) requires web searches and fetching article content; Phase 7 (media) can call other image/video generation skills. Those operations may perform network requests and may involve third-party services that could receive article text or prompts. Verify which image/video skills are installed and what credentials they use. - Persistent logs: creation-log.md stores project state and will be read/written at the start/end of phases. Treat it as persistent metadata about your drafts and decisions. If you don't want persistent logs, remove or relocate this file and instruct the agent accordingly. - Privacy claim caution: the SKILL.md repeatedly claims "all analysis is local" — confirm in your agent/runtime whether web/search calls or media-generation calls actually keep data local in practice; some tool integrations will necessarily send data to external services. - Testing suggestion: first run with a non-sensitive sample article, observe what files are created/modified and whether any external tool is invoked, and check any prompts that ask to auto-analyze your articles. Overall: behavior is coherent for a writing assistant, but it has real privacy/persistence implications (file reads/writes and optional external media-generation). Review the files you give it and the other generator skills it may call before using on sensitive material.