Hi, I'm reaching out because OpenClaw-RL sits close to training and improvement loops for OpenClaw-style agents.
Project:
https://github.com/AMAP-ML/SkillClaw
This is not meant as a generic wrapper pitch. SkillClaw focuses on a narrower ecosystem gap: after repeated usage, long-lived skill libraries often become noisy, duplicated, stale, and fragmented across agents or devices.
Its role is a post-task skill evolution loop that:
- deduplicates overlapping skills
- merges related skills
- improves skill quality over time
- shares evolved skills across agents / devices / teams
I thought this might be relevant as a complementary layer because your project already sits near the core runtime / workflow surface.
If useful, I can send a concise technical summary instead of promo copy.
Paper:
https://arxiv.org/abs/2604.08377
Hi, I'm reaching out because
OpenClaw-RLsits close to training and improvement loops for OpenClaw-style agents.Project:
https://github.com/AMAP-ML/SkillClaw
This is not meant as a generic wrapper pitch. SkillClaw focuses on a narrower ecosystem gap: after repeated usage, long-lived skill libraries often become noisy, duplicated, stale, and fragmented across agents or devices.
Its role is a post-task skill evolution loop that:
I thought this might be relevant as a complementary layer because your project already sits near the core runtime / workflow surface.
If useful, I can send a concise technical summary instead of promo copy.
Paper:
https://arxiv.org/abs/2604.08377