token-decrease
Default concise communication overlay for this repository. Read the canonical skill before acting, keep it active by default, and switch to normal wording when safety, irreversible actions, or clarity need more explicit phrasing.
- License 许可
- MIT
- 最近更新
- 2026-10-23
- Stars
- ★ 2
- Forks
- 0
01 前言 Front Matter
- name
- token-decrease
- description
- Default concise communication overlay for this repository. Read the canonical skill before acting, keep it active by default, and switch to normal wording when safety, irreversible actions, or clarity need more explicit phrasing.
02 资源文件 1
- 参考文档agents/openai.yaml390 B
03 SKILL.md 正文
GitHub Skill Adapter
Canonical skill source:
ai/skills/token-decrease/SKILL.mdai/skills/token-decrease/agents/openai.yaml
This .github file is only a thin reference adapter so GitHub-side
repo docs point to the same canonical skill content in ai/.
04 同仓库更多 Skills
token-decrease
Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", "token decrease", or invokes /caveman. Also auto-triggers when token efficiency is requested.
readme-pm-review
Use only when the user explicitly asks for a product-manager style review of a repository README. Read only the README first, do not infer undocumented features, and provide readability feedback, rewrite suggestions, and product-priority recommendations.
python-review-qa-loop
Use when working on Python code changes that should follow this repo's shared review and QA loop. Read the shared workflow docs, make the smallest viable patch, run relevant tests, capture results, and route DB, credentials, secrets, or Garmin password related changes through the shared security guidance.
improve-codebase-architecture
Scan a codebase for deepening opportunities, present them as a visual HTML report, then grill through whichever one you pick.
grilling
Grill the user relentlessly about a plan, decision, or idea. Use when the user wants to stress-test their thinking, or uses any 'grill' trigger phrases.
grill-with-docs
A relentless interview to sharpen a plan or design, which also creates docs (ADR's and glossary) as we go.