post-training
当你要通过更新模型权重来提升 Agent 能力(Mid-training 补知识、SFT 固化格式、RL 泛化策略、蒸馏降成本)时使用;当需要判断"某个问题该靠 prompt 解决还是值得训练"时使用。核心认识:数据和环境比算法更重要;SFT 记忆、RL 泛化。不适用于 Harness/上下文层的能力提升。
Source facts
- Repository
- GongLingRui/agent-dev-skills
- Last source activity
- August 27, 2026 at 01:47
- Detected SKILL.md language
- Chinese
- Stars
- 1
- Forks
- 0
Install options
The review-first prompt is selected by default. You can switch to a direct command or download a local copy.
Review the source files
Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.