AI Collaboration System / 智能多 Agent 协作编排框架
Project description
AI Collaboration System / 智能协作系统
面向 Codex / Claude / Gemini 的终端多 Agent 协作编排框架。
本框架将真实开发流程中的规划、实现、审查与交付环节拆分至不同模型,并提供可观测、可回放、可降级的协作机制。
核心特性
- 主控优先:先规划再执行,避免单 Agent 盲目推进
- 配置驱动:模型与角色通过配置文件管理,无硬编码依赖
- 过程可见:基于 tmux 的协作模式支持实时交互、日志记录与会话回放
- 故障可控:针对命令不可用、权限受限、子 Agent 失败等场景提供显式降级策略
统一命令接口
项目采用单一命令入口:ai-collab
- 执行任务:
ai-collab "<task>" - 管理操作:
ai-collab <subcommand>(如init、status、config)
查看完整执行参数:
ai-collab run --help
安装方式
从源码安装
git clone https://github.com/skyhua0224/ai-collab.git
cd ai-collab
python3 -m pip install -e .
从 PyPI 安装
pip install ai-collab
快速开始
1. 初始化配置
ai-collab init
ai-collab status
2. 执行协作任务
ai-collab "设计并实现一个带鉴权的 REST API,并给出测试与发布建议"
3. 启用 tmux 可视化协作
ai-collab \
--provider codex \
--execution-mode tmux \
--tmux-target inline \
--tmux-prewarm-subagents \
"实现一个包含前端与后端的最小业务功能,并完成审查"
常用子命令
| 命令 | 功能说明 |
|---|---|
ai-collab init |
初始化配置文件模板 |
ai-collab status |
查看当前配置与 Agent 可用性状态 |
ai-collab detect |
检测协作需求并生成编排建议 |
ai-collab monitor |
手动启动 tmux 协作工作区 |
ai-collab config |
管理配置项 |
ai-collab select |
根据任务复杂度选择模型策略 |
可观测性与故障排查
日志目录结构
.ai-collab/logs/<session>/
常用排查命令
# 实时查看日志
tail -f .ai-collab/logs/<session>/*.log
# 检测任务编排建议(JSON 格式)
ai-collab detect "<task>" --output json
# 模拟执行(不实际运行)
ai-collab --dry-run --output json "<task>"
文档导航
开发与验证
# 运行测试
pytest -q
# 构建分发包
python3 -m build
# 验证分发包
python3 -m twine check dist/*
开源协议
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
ai_collab-0.1.3.tar.gz
(81.8 kB
view details)
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
ai_collab-0.1.3-py3-none-any.whl
(66.1 kB
view details)
File details
Details for the file ai_collab-0.1.3.tar.gz.
File metadata
- Download URL: ai_collab-0.1.3.tar.gz
- Upload date:
- Size: 81.8 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.12.3
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
64e40bba90b52ad73e779abbc0b67574cb6352622ce2e0b6c468810a6e0ec714
|
|
| MD5 |
22fe5273748bc24d470c8f720a4e456b
|
|
| BLAKE2b-256 |
65f4b3ab41e897319ed7f19aa44acc2740175bda7c4b0342912bb63fbe420930
|
File details
Details for the file ai_collab-0.1.3-py3-none-any.whl.
File metadata
- Download URL: ai_collab-0.1.3-py3-none-any.whl
- Upload date:
- Size: 66.1 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.12.3
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
a8bee648dca7444f25e3ce4260dc220cb3becc5de99c37f8eaf8c8fb4eedd5ae
|
|
| MD5 |
767984559a2d07568b5c645bdc924e61
|
|
| BLAKE2b-256 |
be6ca74ed43cf21a7778b66a8c87d3aa7200644f894cf74c4d0fe46828f69d8f
|