ai-skills Model Card
这是可索引项目报告证据页：它保留 ai-skills 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。
ai-skills Model Card
| Field | Value |
|---|---|
| Repository | iliaal/ai-skills |
| Category | Agent Process Skill Library |
| Stars / forks snapshot | 13 / 3 |
| Language | Shell |
| License | MIT |
| Raw capture | raw-github/iliaal_ai-skills.md |
| Updated by | hourly public metadata update, 2026-05-26 06:45 +0800 |
1. Role in Self Evolve
ai-skills is a small behavioral skill library for coding agents. It matters to self-evolution because it treats planning, debugging, verification and multi-agent coordination as reusable process assets rather than one-off prompt instructions.
2. Working Principle
agent task -> triggered behavioral skill -> process discipline for planning, testing, review, debugging or verification -> safer reusable execution habit
3. Evidence Path
Web-observed GitHub evidence showed 13 stars, 3 forks, 61 commits, MIT licensing, Shell-dominant repository metadata, and a v4.0.3 release dated May 16, 2026. The README says the skills work across Codex, Claude Code, Cursor, Gemini CLI, OpenCode, OpenClaw and related agent environments.
4. Teaching Use
Use this card to explain the difference between capability and process: many coding agents can write code, but a skill layer can make verification, planning and root-cause debugging persistent across context resets.
5. Limits
The repository was not cloned or executed locally. Metadata is web-observed from the public GitHub page, not authenticated GitHub API data.