SkillClaw Model Card
这是可索引项目报告证据页：它保留 SkillClaw 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。
SkillClaw Model Card
| Field | Value |
|---|---|
| Repository | AMAP-ML/SkillClaw |
| Category | Collective Skill Evolution |
| Stars / forks snapshot | 1500 / 137 |
| Language | Python |
| License | MIT |
| Raw capture | raw-github/amap-ml_skillclaw.md |
| Updated by | hourly public metadata update, 2026-05-25 08:18 +0800 |
1. Role in Self Evolve
SkillClaw 直接回应“skill 如何自进化”：它从真实 agent 会话中抽取、去重、验证并共享可复用 skills，让单用户和多 agent 团队都能把经验沉淀为演化资产。
2. Working Principle
agent session data -> proxy capture -> skill dedup/improvement/verification -> shared evolve server -> reused skills
3. Evidence Path
web GitHub page observed MIT license, 1.5k stars, 137 forks, Python primary language, arXiv paper citation, 2026-04-10 open-source note, and 2026-04-20 Codex/Claude Code integration note. Shell GitHub API access remained blocked by DNS in this run, so this card treats the current snapshot as web-observed rather than API-verified.
4. Teaching Use
Use this card to explain Collective Skill Evolution in the raw -> classification -> project card -> site/report pipeline. The reading path is: raw capture -> classification row -> public site card -> project report -> aggregate GitHub analysis.
5. Limits
当前未克隆源码，未运行 benchmark、skill install flows、memory experiments、MCP servers、agent evolution loops、security scanners 或 production deployments；star/fork/commit/release 快照来自公开 GitHub 页面文本或可见页面片段。