社区维护的AI代理技能，帮助他们保存代币并向世界写信。

该仓库以代理为先：技能设计适用于任何具备能力的 AI 运行时，Automaton 是主要的参考实现。

技能的运作方式

技能是一个文件夹，里面有一个SKILL.md：

SKILL.md 使用 YAML 前置 + 标记指令：

播放时间说明

• 自动机加载来自~/.automaton/skills/ /SKILL.md的技能。
• 其他代理可以使用相同的SKILL.md模式，但加载路径和激活行为可能在运行时有所不同。

哲学

• 技能应当是安全、高度信任且默认安全的。
• 技能应使人工智能代理，尤其是自动机能够实现现实世界的行动。
• 技能应简洁实用，以减少执行时的代币使用。
• 技能应提升代理的自主性，同时保持明确的运营保护。

贡献

欢迎永久居民。

添加技能时：

• 创建 /SKILL.md。
• 保持指令具体、简洁且安全。
• 包含假设（例如，已经预设的内容）。
• 把技能加到 SKILLS.md 上。

Conway AI Automaton Agent Skills (WIP)

Community-maintained skills for AI agents to help them save tokens and write to the world.

This repo is agent-first: skills are designed for any capable AI runtime, with Automaton as the primary reference implementation.

How skills work

A skill is one folder with one SKILL.md:

/
  SKILL.md

SKILL.md uses YAML frontmatter + markdown instructions:

---
name: my-skill
description: One-line summary
auto-activate: true
---

# Instructions
...

Runtime notes

• Automaton loads skills from ~/.automaton/skills//SKILL.md.
• Other agents can use the same SKILL.md pattern, but loading paths and activation behavior may differ by runtime.

Philosophy

• Skills should be secure, high-trust, and safe by default.
• Skills should enable real-world action for AI agents, especially Automatons.
• Skills should be concise and practical to reduce token usage during execution.
• Skills should improve agent autonomy while preserving clear operational guardrails.

Contributing

PRs are welcome.

When adding a skill:

• Create /SKILL.md.
• Keep instructions concrete, concise, and safe.
• Include assumptions (for example, what is already provisioned).
• Add the skill to SKILLS.md.