token-optimization
Token economy and context optimization. Use when seeking to minimize token consumption or the context exceeds 60%.
来源信息
- 仓库
- dabydat/alpha-omega
- 最近来源活动
- 2026年9月25日 06:21
- 检测到的 SKILL.md 语言
- 英语
- 星标
- 1
- Forks
- 0
安装方式
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
检查来源文件
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
正在显示 SKILL.md
SKILL.md
来源说明 · 只读预览- name
- token-optimization
- description
- Token economy and context optimization. Use when seeking to minimize token consumption or the context exceeds 60%.
# neuron: token-optimization
- triggers_when: seeking to save tokens or the context exceeds 60% of the limit
- tier: T1
- budget: 300 tokens
- eval: context <=60% and no repeated context -> threshold 1
- reward: pheromone +1 if eval >= threshold
## Procedure
1. Never repeat context: use memory/STATE.md, AGENTS.md and diagrams.
2. Prefer diagrams over text (5x fewer tokens).
3. Use short prompts with file references.
4. Always update memory/STATE.md when finishing.
5. Compact when exceeding 60% or changing phase.
## Checklist
- [ ] Context in files, not repeated in the prompt
- [ ] Diagrams for complex flows
- [ ] One task per prompt
- [ ] Explicit response format; memory/STATE.md updated
## Pitfalls
- Pasting complete files when one section suffices
- Re-explaining context already in memory/STATE.md
- Not compacting on multi-day tasks
在 GitHub 查看