用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/seokan-jeong/team-shinchan --skill team-shinchanlearn命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
基于 SOC 职业分类
正在显示 SKILL.md
| name | team-shinchan:learn |
| description | Use when you want to manually add new learnings or rules to memory. |
| user-invocable | false |
Teach Team-Shinchan something to remember.
/team-shinchan:learn "Always use TypeScript strict mode in this project"
/team-shinchan:learn "User prefers functional components over class components"
| Keywords | Category | Tier |
|---|---|---|
| prefer, like, want | preference | preference |
| pattern, approach, way | pattern | procedural |
| always, never, rule, use | convention | procedural |
| error, mistake, avoid | mistake | procedural |
| decided, architecture | decision | procedural |
| other | insight | procedural |
Note: tool tier is set when user explicitly passes --tier tool or AI detects tool-specific phrasing (CLI flags, tool config patterns).
.shinchan-docs/learnings.md (create if missing):### [{category}] {title}
- **Date**: {YYYY-MM-DD}
- **Source**: manual
- **Tier**: {auto-mapped tier}
- **Confidence**: high
- **Tags**: #{auto-generated tags}
- **Insight**: {content}
---
🧠 [Learn] Saved to memory:
Category: {category}
Learning: "{content}"
📁 Location: .shinchan-docs/learnings.md
.shinchan-docs/)load-kb hook/team-shinchan:forget