alas开源AI工作流 是 AI Skill Hub 本期精选Agent工作流之一。综合评分 7.5 分,整体质量较高。我们推荐使用将其纳入你的 AI 工具库,帮助提升工作效率。
alas是开源AI工作流,提供了一个Agent Opinionated Conductor,每个Agent,每个Worktree,只需一窗口即可使用。它支持MacOS App和Swift语言,适合于Workflow和Agentic-AI相关的项目。
alas开源AI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。
alas是开源AI工作流,提供了一个Agent Opinionated Conductor,每个Agent,每个Worktree,只需一窗口即可使用。它支持MacOS App和Swift语言,适合于Workflow和Agentic-AI相关的项目。
alas开源AI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/mrmans0n/alas cd alas # 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 alas --help # 基本运行 alas [options] <input> # 详细使用说明请查阅文档 # https://github.com/mrmans0n/alas
# alas 配置说明 # 查看配置选项 alas --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 ALAS_CONFIG="/path/to/config.yml"
Every agent. Every worktree. One window.
Run coding agents in a terminal or a native chat pane, across every worktree, and review their work — diffs, commits, PRs — without leaving the app. Then pick the whole thing up from your phone. Native macOS, terminals built on libghostty.

- Run agents two ways. A long-lived Ghostty terminal or a native chat pane over ACP — same worktree, your call. Tool calls, plans, and permission prompts render inline. Works with Claude Code, Codex, Cursor, Gemini, OpenCode, Pi, and Copilot.
- Parallel worktrees, one window. Every repo lives in the sidebar with its linked worktrees underneath. Switching is instant; terminal sessions, tabs, and scroll positions persist per worktree.

- Drive from your phone. Pair by QR, watch any session live, and take over a running desktop turn — send a prompt, stop a run, or answer a permission request from the couch. Push notifications fire when an agent needs you.
- Harness-aware. Detects Claude Code, Codex, Cursor, Gemini, and friends running in your worktrees and surfaces their state — busy, awaiting permission, awaiting input — as live badges in the sidebar. One-click hook install.
- Git that gets out of the way. File- and hunk-level staging, inline diffs, AI-drafted commit messages, and a 3-way merge editor for conflicts — all without leaving your editor.

- Review without the round-trip. GitHub PRs and GitLab MRs detected on your branch, AI-drafted titles and descriptions, and a review-evidence inspector that pulls CI logs, check results, and test output right next to the diff.
- A real editor underneath. Tree-sitter syntax for 20+ languages, LSP-powered go-to-definition, hover docs, and completions, plus Markdown source-and-preview side-by-side.

- Tabs for everything. Terminals, agent sessions, file viewers, diffs, and Markdown — open in tabs, drag to reorder, restore on relaunch.
brew install --cask mrmans0n/tap/alas
Or grab the signed DMG from Releases.
alas是一个有潜力的开源AI工作流,但仍需要进一步的开发和完善。它提供了一个Agent Opinionated Conductor,每个Agent,每个Worktree,只需一窗口即可使用,支持MacOS App和Swift语言,适合于Workflow和Agentic-AI相关的项目。
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,alas开源AI工作流 在Agent工作流赛道中表现稳健,质量良好。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 | alas |
| 原始描述 | 开源AI工作流:AI agent opinionated conductor: Every agent. Every worktree. One window.。⭐6 · Swift |
| Topics | workflowagentic-aimacos-appswift |
| GitHub | https://github.com/mrmans0n/alas |
| License | MIT |
| 语言 | Swift |
收录时间:2026-06-09 · 更新时间:2026-06-11 · 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