双 Agent 协作开发系统
基于 OpenCode 的双 Agent 协作框架，实现产品经理与开发的分离式协作。
安装
pip install opencode-collaboration
快速开始
1. 初始化项目
# 创建新项目
oc-collab init myproject
cd myproject
2. 创建需求文档
编辑 docs/01-requirements/requirements_v*.md 文件，定义需求。
3. 评审并签署
# Agent 2 (开发) 评审需求
oc-collab review requirements --comment "技术方案可行"
# Agent 1 (产品) 签署确认
oc-collab signoff requirements --comment "同意实现"
4. 推进到下一阶段
# 推进到设计阶段
oc-collab advance -p design
常用命令
| 命令 | 说明 |
|---|---|
oc-collab status |
查看项目状态和待办摘要 (v2.2.3) |
oc-collab .a |
显示当前关联的项目信息 (v2.2.3) |
oc-collab todowrite |
创建待办任务 (v2.2.3) |
oc-collab todoedit |
编辑待办任务 (v2.2.3) |
oc-collab todo |
查看待办事项 |
oc-collab todowrite --auto-check |
创建TODO，自动检查参数并携带上下文 (v2.2.6) |
oc-collab skill search --keywords <kw> |
搜索Skill文档 (v2.2.6) |
oc-collab skill slice <skill> |
查看Skill特定切片 (v2.2.6) |
oc-collab skill enforce |
检查Skill加载状态 (v2.2.6) |
oc-collab advance -p <phase> |
推进到指定阶段 |
oc-collab signoff <stage> |
签署确认 |
oc-collab history |
查看协作历史 |
oc-collab compliance check |
合规检查 (v2.2.2) |
oc-collab git sync |
Git同步 (v2.2.2) |
oc-collab rules init [--force] |
初始化框架规则 (v2.2.8) |
oc-collab rules status |
检查规则初始化状态 (v2.2.8) |
oc-collab rules init --auto-load |
自动加载默认规则 (v2.2.9) |
oc-collab compliance check |
合规检查 (v2.2.2) |
oc-collab compliance report |
生成合规报告 (v2.2.9) |
oc-collab compliance violations |
查看违规记录 (v2.2.9) |
oc-collab deploy check-docs |
检查部署文档同步 (v2.2.9) |
oc-collab deploy sync-docs |
同步部署文档 (v2.2.9) |
oc-collab webhook notify test |
测试Webhook通知功能 (v2.2.8) |
oc-collab startup check |
启动时检查未读TODO (v2.2.10) |
oc-collab todo list |
显示TODO列表 (v2.2.10) |
oc-collab todo list --unread |
仅显示未读TODO (v2.2.10) |
| `oc-collab todo list --agent <1 | 2>` |
oc-collab todo mark-read <id> |
标记TODO为已读 (v2.2.10) |
oc-collab todo stats |
TODO统计信息 (v2.2.10) |
oc-collab skill check |
检查Skill加载状态 (v2.2.10) |
oc-collab skill status |
显示Skill合规状态 (v2.2.10) |
oc-collab skill verify <action> |
验证操作前Skill (v2.2.10) |
协作流程
需求评审 → 设计评审 → 开发完成 → 测试通过 → 部署上线
↓ ↓ ↓ ↓ ↓
签署 签署 签署 签署 完成
每个阶段必须双方签署后才能进入下一阶段
阶段说明
| 阶段 | 说明 |
|---|---|
requirements |
需求评审阶段 |
design |
设计评审阶段 |
development |
开发阶段 |
testing |
测试阶段 |
deployment |
部署阶段 |
项目结构
myproject/
├── docs/
│ ├── 01-requirements/ # 需求文档
│ ├── 02-design/ # 设计文档
│ └── 03-test/ # 测试报告
├── state/
│ └── project_state.yaml # 项目状态
└── src/ # 源代码
版本历史
| 版本 | 日期 | 功能 |
|---|---|---|
| v2.5.1 | 2026-03-23 | Patch: test-results API路径修复为/api/v1/test-results，符合REQ v1.3规范 |
| v2.5.0 | 2026-03-23 | 双Git仓库配置管理(Gitee+Gitea)、Git配置/推送/提交API、测试结果接收API、错误隔离 |
| v2.4.2 | 2026-02-26 | HTTP API服务 (FastAPI, 端口8001)、ITodoStorage接口抽象、TODO CRUD接口 |
| v2.3.2 | 2026-02-19 | SQLite存储 + 实时通知交互 + 监听守护进程 + Agent TODO隔离 |
| v2.2.12 | 2026-02-15 | 部署自动化CLI：完整部署流程、预览模式、版本管理、PyPI/Git推送、发布验证 |
| v2.2.11 | 2026-02-14 | Agent独立TODO编号、Skill强制执行增强、StateReceiver状态接收器 |
| v2.2.10 | 2026-02-14 | TodoQueueManager + Agent Startup Checker + CLI Enhancements |
| v2.2.9 | 2026-02-14 | StateNotifier集成（todowrite/signoff/phase_advance）、自动Bug检测、Agent Compliance CLI准入、规则自动加载 |
| v2.2.8 | 2026-02-13 | Webhook基础架构：EventDispatcher、StateNotifier骨架、HMAC验证 |
| v2.2.7 | 2026-02-10 | Skill内容准确性测试、覆盖率统计、事件监听与崩溃恢复 |
| v2.2.6 | 2026-02-09 | 智能辅助：todowrite自动检查、TODO上下文携带、冲突检测；Skill关键词检索、切片查看、强制查找 |
| v2.2.5 | 2026-02-09 | 文件Owner机制、signoff修复、需求分析流程、评审机制优化 |
| v2.2.4 | 2026-02-08 | Skill强制加载检查、Git提交前签署验证、需求文档完整性检查 |
| v2.2.3 | 2026-02-08 | Agent体验优化：.a项目信息、任务同步、状态摘要 |
| v2.2.2 | 2026-02-07 | 协作规范强制执行、Git同步集成 |
| v2.2.1 | 2026-02-07 | 双代理认知免疫系统、动态Checklist |
| v2.2.0 | 2026-02-06 | 初始版本 |
子系统集成
如需将 oc-collab 集成到其他子系统（如 PM-Agent），请查阅：
- 能力使用指南:
docs/07-research/RESEARCH_oc-collab_capabilities_for_PM-Agent.md - 核心架构:
docs/00-architecture/CORE_ARCHITECTURE.md
详细文档
完整使用指南请参见项目 GitHub 仓库： https://github.com/legoland01/dual-agent-collaboration-system
Author: liuzhen License: MIT
1 maintainer
Metadata
1 maintainer
Release files for opencode-collaboration 2.5.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| opencode_collaboration-2.5.1.tar.gz | 493.1 kB | Without provenance | Details for opencode_collaboration-2.5.1.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| opencode_collaboration-2.5.1-py3-none-any.whl | Python 3 | none | any | Details for opencode_collaboration-2.5.1-py3-none-any.whl |
Total release size: 836.6 kB
Release files / opencode_collaboration-2.5.1.tar.gz
| Download URL | opencode_collaboration-2.5.1.tar.gz |
|---|---|
| Size | 493.1 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
12a38c3e4d14fea180b7142be152cb98176d2cc206af0b75e3ea9232de9388e2
|
|
BLAKE2b-256 checksum How to use checksums |
0484a4e850c59e572b3b589e779b8900db3d5510120d61a17ab730fc121b705a
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.9.6
|
Release files / opencode_collaboration-2.5.1-py3-none-any.whl
| Download URL | opencode_collaboration-2.5.1-py3-none-any.whl |
|---|---|
| Size | 343.5 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
84680ae7d90c3fecf9bc5bbe259783409dd5c8ea9022767b3f81ce1054b2c97d
|
|
BLAKE2b-256 checksum How to use checksums |
447e20bfc9e4f3915fa0c55249a6109acb115aed0e3d78f7df4327e640545806
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.9.6
|
Security for opencode-collaboration 2.5.1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.
