Agent Skills
› heyron-ai/agent-library
› human-writing
human-writing
GitHub提供人类化写作规范,指导撰写博客、营销文案等对外内容。通过避免AI常见特征(如过度正式、词汇循环)并检查反AI模式清单,确保内容自然真实。适用于起草、编辑和审查面向外部受众的文本。
Trigger Scenarios
起草或编辑博客文章、新闻通讯、营销文案
审查内容以消除AI生成痕迹
撰写社交媒体帖子或客户邮件
Install
npx skills add heyron-ai/agent-library --skill human-writing -g -y
SKILL.md
Frontmatter
{
"name": "human-writing",
"description": "Write content that reads as naturally human — no AI tells, no corporate fluff. Use when drafting, editing, or reviewing any content meant for publication or external audiences: blog posts, articles, newsletters, marketing copy, social media threads, client-facing documents, website copy, emails to lists. NOT for: casual chat, code, internal notes, or quick replies."
}
Human Writing
Apply these principles when drafting or editing content for external consumption.
Core rules
- Be specific over general — concrete facts beat vague praise
- Use simple verbs — is, has, was, did — not "serves as," "boasts," "showcases"
- No cheerleading — state facts, skip "this is important because…"
- Repeat words comfortably — humans reuse words; don't cycle synonyms
- Short sentences are fine — not everything needs three clauses
- Attribute opinions specifically — "Roger Ebert wrote…" not "Critics have noted…"
- Skip forced significance — not everything "reflects broader trends"
- Use lowercase headings — title case screams AI
- Bold sparingly — not every other phrase
- Use contractions — "it's," "don't," "won't" sound human
Before publishing
Scan the draft against the full anti-AI patterns reference for red flags:
read references/anti-ai-patterns.md
Check for:
- Clusters of AI vocabulary (delve, pivotal, tapestry, etc.)
- "-ing" phrases tacked onto sentence ends
- "Despite challenges…" formulas
- Vague attributions ("experts say")
- Significance inflation
- Synonym cycling
- Rule-of-three lists
- Overly formal verbs
Fix anything that trips the checklist, then re-read once more for natural flow.
Version History
- fcecd5d Current 2026-07-25 05:36


