Rosetta Agent工作流 是 AI Skill Hub 本期精选Agent工作流之一。综合评分 5.8 分，整体质量稳定。我们值得关注将其纳入你的 AI 工具库，帮助提升工作效率。
Rosetta Agent工作流 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排，将复杂的多步骤任务拆解为清晰的自动化流程，实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成，适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
Rosetta Agent工作流 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排，将复杂的多步骤任务拆解为清晰的自动化流程，实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成，适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
# 克隆仓库 git clone https://github.com/RajanChavada/Rosetta cd Rosetta # 查看安装说明 cat README.md # 按 README 完成环境依赖安装后即可使用
# 查看帮助 rosetta --help # 基本运行 rosetta [options] <input> # 详细使用说明请查阅文档 # https://github.com/RajanChavada/Rosetta
# rosetta 配置说明 # 查看配置选项 rosetta --config-example > config.yml # 常见配置项 # output_dir: ./output # log_level: info # workers: 4 # 环境变量（覆盖配置文件） export ROSETTA_CONFIG="/path/to/config.yml"
██████╗ ██████╗ ███████╗███████╗████████╗████████╗ █████╗
██╔══██╗██╔═══██╗██╔════╝██╔════╝╚══██╔══╝╚══██╔══╝██╔══██╗
██████╔╝██║ ██║███████╗█████╗ ██║ ██║ ███████║
██╔══██╗██║ ██║╚════██║██╔══╝ ██║ ██║ ██╔══██║
██║ ██║╚██████╔╝███████║███████╗ ██║ ██║ ██║ ██║
╚═╝ ╚═╝ ╚═════╝ ╚══════╝╚══════╝ ╚═╝ ╚═╝ ╚═╝ ╚═╝| Feature | Description |
|---|---|
| **3-Layer Memory** | Project decisions, heuristics, and daily logs |
| **Skill Catalog** | Browse, search, and install pre-built skills from the catalog |
| **Multi-Source Skills** | Local, global (~/.rosetta), or git-sourced skills |
| **Auto-Detection** | Automatically detects project type and tech stack |
| **Format Translation** | Convert configs between any supported IDE format |
| **YAML-First Config** | Strongly typed, schema-validated .rosetta configuration |
| **Agent / Persona** | Scaffolding tools for interactive sub-agents and conventions |
| **Post-Scaffold Hooks** | Run scripts automatically after setup |
---
Scaffold — Set up the .ai/ directory and configure IDEs
rosetta scaffoldThe scaffold command detects your project stack (Next.js, React Vite, FastAPI, etc.) and generates appropriate IDE configuration files (CLAUDE.md, .cursor/rules/*.mdc, etc.).
Options: - --yes - Skip all interactive prompts, use defaults (for CI/CD automation) - --ide <name> - Specify IDE(s) to generate (for example: claude, vscode, cursor, windsurf, copilot, antigravity, replit) - --stack <stack> - Override detected stack (for example: next.js, react-vite, node-api) - --dry-run - Preview what would be generated without writing files
Examples:
rosetta init # Interactive: detect stack, select IDEs, confirm
rosetta init --yes # Non-interactive: use detected stack, defaults
rosetta init --ide claude --ide windsurf # Generate for specific IDEs
rosetta init --stack node-api --ide vscode --ide copilot
rosetta init --dry-run # Preview without writingrosetta scaffold --auto-ideateThe --auto-ideate flag analyzes your project and automatically copies a framing prompt to your clipboard. Paste this into your AI agent to start a guided workshop that implements custom skills for your codebase.
Scaffold — Full zero-config setup (Brain + Memory + IDE Configs)
rosetta scaffoldscaffold command detects your stack, initializes the .ai/ directory, creates your 3-layer memory system, and generates native IDE rules (CLAUDE.md, .cursor/rules/, etc.). Note: rosetta init is an alias for scaffold.
Options: - --yes - Skip all prompts, use defaults - --auto-ideate - Generate ideation template and copy to clipboard (Recommended) - --ide <name> - Specify IDE(s) to generate (e.g. claude, cursor, windsurf) - --stack <stack> - Override detected stack (e.g. next.js, react-vite, node-api)
Sync — Verify IDE wrappers or regenerate them from templates
rosetta sync --regenerate-wrappers.ai/master-skill.md. Use this if your wrappers get out of sync or if you manually edit the master spec.
Watch — Monitor .ai/master-skill.md and auto-regenerate all IDE wrappers on change
rosetta watchCatalog — Browse available skills in the catalog ```bash rosetta catalog
Install skills directly from any Git repository with rosetta install. Rosetta validates the skill, records provenance, and tracks it in your manifest.
```bash
rosetta install https://github.com/org/api-auth
rosetta install https://github.com/community/ppt-gen --global
**Installation features:**
- Validates the repository contains a `SKILL.md` file with required frontmatter (`name`, `description`)
- Checks skill name validity (lowercase alphanumeric and hyphens)
- **Automatically translates** skills into each detected IDE's native format (frontmatter, path, file extension)
- Copies the skill to `.rosetta/skills/<name>` in your project (or `~/.rosetta/skills/` for global)
- Preserves the `.git` directory and adds an `upstream` remote pointing to the source repository
- Records installation details in `.rosetta/skills/manifest.json`, including source URL, commit hash, install date, and all translated paths
- Supports `--force` to overwrite an existing installation
- Supports `--ide <name>` to install for a single specific IDE
- Supports `--dry-run` to preview changes
**Multi-IDE output example:**Get your project AI-ready in one step:
npx rosettablueprint scaffold --auto-ideateWhat happens next: 1. Detection: Rosetta identifies your frameworks, languages, and test setup. 2. Brain Creation: A structured .ai/ directory is created (Memory, Logs, Master Spec). 3. IDE Sync: Native rules are written to your editor (e.g., .cursor/rules/ or CLAUDE.md). 4. Ideation: A framed prompt is copied to your clipboard. 5. Chat: Paste into your AI chat, answer 3 questions, and get custom skills implemented.
---
One command → Rules injected → Start chatting.
Rosetta is a CLI tool that builds a Global Brain for your project. It automatically detects your stack, initializes a 3-layer memory system, and translates your conventions into native rules for any IDE (Claude Code, Cursor, Windsurf, Copilot, and more).
The Rosetta Story: You shouldn't have to learn how.cursorrulesorCLAUDE.mdwork. Runrosetta scaffoldand your AI agent instantly becomes a senior engineer on your specific codebase.
<a href="https://github.com/RajanChavada/RajanChavada/Rosetta/actions"><img src="https://img.shields.io/badge/version-0.5.6-blue.svg" alt="Version"></a> <a href="https://github.com/RajanChavada/Rosetta/blob/main/LICENSE"><img src="https://img.shields.io/badge/license-MIT-blue.svg" alt="License"></a> <a href="https://github.com/RajanChavada/Rosetta"><img src="https://img.shields.io/badge/status-stable-green.svg" alt="Status"></a>
---
| Command | Description |
|---|---|
migrate | Interactive wizard for existing repos |
migrate --source <path> | Migrate from a custom folder (e.g. agentic-corder/) |
migrate-from-cursor | Convert .cursorrules to .ai/ |
migrate-from-claude | Convert CLAUDE.md to .ai/ |
---
Templates that Rosetta scaffolds into other projects. When installed or created on a multi-IDE project, skills are automatically translated into each IDE's native format:
| Skill | Stack | Domain |
|---|---|---|
node-express-postgres | Node.js, Express, PostgreSQL | Backend APIs |
frontend-react-next | React, Next.js | Frontend apps |
testing-full-pyramid | Testing frameworks | Quality assurance |
data-ml-project | Data Science, ML | Analytics & ML |
AI Skill Hub 为第三方内容聚合平台，本页面信息基于公开数据整理，不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后，再部署至生产环境，并做好必要的安全评估。
✅ MIT 协议 — 最宽松的开源协议之一，可自由商用、修改、分发，仅需保留版权声明。
经综合评估，Rosetta Agent工作流 在Agent工作流赛道中表现稳健，有一定参考价值。如果你已有明确的使用需求，可以直接上手体验；如果还在评估阶段，建议对比同类工具后再做决策。
| 原始名称 | Rosetta |
| 原始描述 | 开源AI工作流：CLI tool that scaffolds and syncs agentic AI rules/skills across distinct and fr。⭐3 · HTML |
| Topics | workflowhtml |
| GitHub | https://github.com/RajanChavada/Rosetta |
| License | MIT |
| 语言 | HTML |
收录时间：2026-05-14 · 更新时间：2026-05-16 · License：MIT · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型，复制安装指令后粘贴到对应客户端