Skills Best Practices Model Card
这是可索引项目报告证据页：它保留 Skills Best Practices Model Card 的源材料入口、机制线索和限制提醒；正文仍需 reader/editor 与 academic public-copy review 后才能当作最终结论引用。
Skills Best Practices Model Card
| Field | Value |
|---|---|
| Repository | mgechev/skills-best-practices |
| Category | Agent Skill Authoring Methodology |
| Stars / forks snapshot | 1900 / 131 |
| Language | Python / Markdown |
| License | not verified |
| Raw capture | raw-github/mgechev_skills-best-practices.md |
| Updated by | hourly public metadata update, 2026-05-26 03:41 +0800 |
1. Role in Self Evolve
Skills Best Practices is a methodology layer for agent self-evolution: it turns reusable know-how into discoverable, testable SKILL.md packages with progressive disclosure, deterministic helper scripts and validation loops.
2. Working Principle
skill need -> trigger-optimized frontmatter -> lean SKILL.md -> references/scripts/assets -> discovery/logic/edge-case validation -> regression-aware skill iteration
3. Evidence Path
web-observed GitHub and Skills Hub evidence showed about 1.9k stars, 131 forks, 5 commits, Python language signal, a guide for professional-grade agent skills, validation guidance, skillgrade linkage and a Skills Hub sync signal on 2026-05-24. Shell GitHub API freshness remained blocked by DNS and the local gh token was invalid, so this card treats the current metadata as web-observed rather than API-verified.
4. Teaching Use
Use this card to explain why the skill layer is not just a prompt library: good skills have routing metadata, a small executable contract, progressive disclosure, edge-case tests and regression checks.
5. Limits
The repository was not cloned in this iteration; no skill was installed, validated, scanned or benchmarked locally. Counts and claims are visible public-page or mirror-page signals unless independently revalidated later.