iannuttall/dotagents

点代理人

一个规范的.agents文件夹，为你所有的AI工具提供动力。

简单设置 • 唯一可靠来源 • 随时可安全重启

快速启动

需求：包子1.3+。

运行引导CLI：

或者和Bun一起：

选择一个工作区（全局主页或项目文件夹），选择你想管理的客户端，然后按照提示操作。你可以随时重新运行，修复链接或撤销更改。

全局主页影响所有项目。项目文件夹只影响你当前运行dotagents的目录。

它的作用

• 保持.agents为真理来源。
• 根据你的选择，创建Claude、Codex、Factory、Cursor、OpenCode和Gemini的符号链接。
• 总是在覆盖前创建备份，确保更改可逆。

其关联（全球范围）

.agents/CLAUDE.md → ~/.claude/CLAUDE.md（如果存在）

.agents/AGENTS.md → ~/.claude/CLAUDE.md（无 CLAUDE.md 时退回）

.agents/GEMINI.md → ~/.gemini/GEMINI.md（如果存在）

.agents/AGENTS.md → ~/.gemini/GEMINI.md（无 GEMINI.md 时的备选）

.agents/commands → ~/.claude/commands

.agents/commands → ~/.factory/commands

.agents/commands → ~/.codex/prompts

.agents/commands → ‘~/.’

Quick Start

Requirements: Bun 1.3+.

Run the guided CLI:

npx @iannuttall/dotagents

Or with Bun:

bunx @iannuttall/dotagents

Choose a workspace (Global home or Project folder), select the clients you want to manage, and follow the prompts. You can run it again anytime to repair links or undo changes.

Global home affects all projects. Project folder only affects the current directory you run dotagents from.

What it does

• Keeps .agents as the source of truth.
• Creates symlinks for Claude, Codex, Factory, Cursor, OpenCode, and Gemini (based on your selection).
• Always creates a backup before any overwrite so changes are reversible.

Where it links (global scope)

.agents/CLAUDE.md → ~/.claude/CLAUDE.md (if present)

.agents/AGENTS.md → ~/.claude/CLAUDE.md (fallback when no CLAUDE.md)

.agents/GEMINI.md → ~/.gemini/GEMINI.md (if present)

.agents/AGENTS.md → ~/.gemini/GEMINI.md (fallback when no GEMINI.md)

.agents/commands → ~/.claude/commands

.agents/commands → ~/.factory/commands

.agents/commands → ~/.codex/prompts

.agents/commands → ~/.cursor/commands

.agents/commands → ~/.config/opencode/commands

.agents/commands → ~/.gemini/commands

.agents/hooks → ~/.claude/hooks

.agents/hooks → ~/.factory/hooks

.agents/AGENTS.md → ~/.factory/AGENTS.md

.agents/AGENTS.md → ~/.codex/AGENTS.md

.agents/AGENTS.md → ~/.config/opencode/AGENTS.md

.agents/skills → ~/.claude/skills

.agents/skills → ~/.factory/skills

.agents/skills → ~/.codex/skills

.agents/skills → ~/.cursor/skills

.agents/skills → ~/.config/opencode/skills

.agents/skills → ~/.gemini/skills

Project scope links only commands/hooks/skills into the project’s client folders (no AGENTS/CLAUDE/GEMINI rules).

Development

Run the CLI in dev mode:

bun run dev

Type-check:

bun run type-check

Run tests:

bun test

Build the CLI:

bun run build

Notes

• Cursor supports .claude/commands and .claude/skills (global or project). dotagents also links .agents/commands → .cursor/commands and .agents/skills → .cursor/skills.
• OpenCode uses ~/.config/opencode/AGENTS.md and prefers AGENTS.md over CLAUDE.md when both exist.
• Codex prompts always symlink to .agents/commands (canonical source).
• Skills require a valid SKILL.md with name + description frontmatter.
• Claude prompt precedence: if .agents/CLAUDE.md exists, it links to .claude/CLAUDE.md. Otherwise .agents/AGENTS.md is used. After adding or removing .agents/CLAUDE.md, re-run dotagents and apply/repair links to update the symlink. Factory/Codex always link to .agents/AGENTS.md.