[Feature Request]: Transform the gammer performance optimization pipeline into a standardized skill, and then implement it as an agent

### Topic

Any Others

### Description

gammer性能优化链路沉淀skill，落成agent

No files, tests, or entry points are named in the issue. First clarify what the “gammer” performance-optimization pipeline includes, where it currently lives, and the expected skill and agent interfaces; done should be defined by agreed implementation and validation criteria.

索引モデルが issue の本文から書いたものです。