@lordmos/dev-crew 0.1.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/LICENSE +21 -0
- package/README.en.md +214 -0
- package/README.md +213 -0
- package/agents/README.md +136 -0
- package/agents/accessibility-auditor.md +48 -0
- package/agents/ai-engineer.md +52 -0
- package/agents/api-tester.md +50 -0
- package/agents/backend-architect.md +59 -0
- package/agents/data-engineer.md +51 -0
- package/agents/database-specialist.md +49 -0
- package/agents/devops-engineer.md +71 -0
- package/agents/embedded-engineer.md +54 -0
- package/agents/frontend-specialist.md +52 -0
- package/agents/game-audio-engineer.md +53 -0
- package/agents/game-designer.md +56 -0
- package/agents/godot-specialist.md +52 -0
- package/agents/incident-response-commander.md +57 -0
- package/agents/level-designer.md +49 -0
- package/agents/mobile-developer.md +53 -0
- package/agents/narrative-designer.md +47 -0
- package/agents/performance-engineer.md +51 -0
- package/agents/security-engineer.md +62 -0
- package/agents/solidity-engineer.md +53 -0
- package/agents/sre.md +56 -0
- package/agents/technical-artist.md +56 -0
- package/agents/technical-writer.md +45 -0
- package/agents/ui-designer.md +71 -0
- package/agents/unity-specialist.md +52 -0
- package/agents/unreal-specialist.md +52 -0
- package/agents/ux-architect.md +53 -0
- package/agents/ux-researcher.md +50 -0
- package/agents/visionos-engineer.md +52 -0
- package/agents/xr-developer.md +53 -0
- package/dist/cli.d.ts +2 -0
- package/dist/cli.js +20 -0
- package/dist/commands/agents.d.ts +1 -0
- package/dist/commands/agents.js +52 -0
- package/dist/commands/init.d.ts +6 -0
- package/dist/commands/init.js +127 -0
- package/package.json +56 -0
- package/templates/INSTRUCTIONS.md +175 -0
package/LICENSE
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
1
|
+
MIT License
|
|
2
|
+
|
|
3
|
+
Copyright (c) 2025 OrcheSpec Contributors
|
|
4
|
+
|
|
5
|
+
Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
6
|
+
of this software and associated documentation files (the "Software"), to deal
|
|
7
|
+
in the Software without restriction, including without limitation the rights
|
|
8
|
+
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
9
|
+
copies of the Software, and to permit persons to whom the Software is
|
|
10
|
+
furnished to do so, subject to the following conditions:
|
|
11
|
+
|
|
12
|
+
The above copyright notice and this permission notice shall be included in all
|
|
13
|
+
copies or substantial portions of the Software.
|
|
14
|
+
|
|
15
|
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
16
|
+
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
17
|
+
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
18
|
+
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
19
|
+
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
20
|
+
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
21
|
+
SOFTWARE.
|
package/README.en.md
ADDED
|
@@ -0,0 +1,214 @@
|
|
|
1
|
+
> **语言 / Language**: [简体中文](README.md) · **English**
|
|
2
|
+
|
|
3
|
+
<div align="center">
|
|
4
|
+
<img src="docs/public/logo.svg" alt="DevCrew" width="180" />
|
|
5
|
+
|
|
6
|
+
# DevCrew
|
|
7
|
+
|
|
8
|
+
**Give AI a collaboration protocol. Let it work like a real team.**
|
|
9
|
+
|
|
10
|
+
*帮你做好 harness!*
|
|
11
|
+
|
|
12
|
+
[](https://www.npmjs.com/package/@lordmos/devcrew)
|
|
13
|
+
[](LICENSE)
|
|
14
|
+
[](https://nodejs.org)
|
|
15
|
+
[](https://github.com/lordmos/dev-crew/pulls)
|
|
16
|
+
|
|
17
|
+
</div>
|
|
18
|
+
|
|
19
|
+
---
|
|
20
|
+
|
|
21
|
+
## The Problem
|
|
22
|
+
|
|
23
|
+
When using AI (Copilot, Claude, Cursor…) for development:
|
|
24
|
+
|
|
25
|
+
| Problem | What happens |
|
|
26
|
+
|---------|-------------|
|
|
27
|
+
| 🧠 **No memory** | Switch conversations, AI forgets everything |
|
|
28
|
+
| 🎭 **No division of labor** | AI plays PM + architect + dev + tester simultaneously |
|
|
29
|
+
| 🎯 **Goes off track** | Drifts from goals with no checkpoints to correct course |
|
|
30
|
+
| 🔍 **Quality blind spots** | No review process — bugs and tech debt accumulate silently |
|
|
31
|
+
| ❓ **No starting point** | Don't know how to orchestrate AI collaboration |
|
|
32
|
+
|
|
33
|
+
**Root cause**: AI lacks a persistent collaboration protocol. DevCrew is that protocol.
|
|
34
|
+
|
|
35
|
+
---
|
|
36
|
+
|
|
37
|
+
## Get Started in 30 Seconds
|
|
38
|
+
|
|
39
|
+
```bash
|
|
40
|
+
npm install -g @lordmos/devcrew
|
|
41
|
+
cd your-project
|
|
42
|
+
crew init
|
|
43
|
+
```
|
|
44
|
+
|
|
45
|
+
Three steps — your project now has a **ready-to-use 6-person AI dev team** + **29 optional domain specialists**, fully auto-orchestrated.
|
|
46
|
+
|
|
47
|
+
Works with **GitHub Copilot · Claude · ChatGPT · Cursor** and any AI platform.
|
|
48
|
+
|
|
49
|
+
---
|
|
50
|
+
|
|
51
|
+
## How It Works
|
|
52
|
+
|
|
53
|
+
```
|
|
54
|
+
You: I need to add auth middleware to the API
|
|
55
|
+
|
|
56
|
+
AI: 📋 Creating change add-api-auth, mode: Standard
|
|
57
|
+
Plan — Requirements:
|
|
58
|
+
- Goal: Add JWT auth to all /api/ routes
|
|
59
|
+
- Acceptance: ☐ No token → 401 ☐ Expired token → 401
|
|
60
|
+
Please confirm.
|
|
61
|
+
|
|
62
|
+
You: Confirmed
|
|
63
|
+
|
|
64
|
+
AI: Design → Execute → Verify — All passed. Please confirm acceptance.
|
|
65
|
+
|
|
66
|
+
You: Confirmed
|
|
67
|
+
|
|
68
|
+
AI: ✅ Change add-api-auth complete.
|
|
69
|
+
```
|
|
70
|
+
|
|
71
|
+
**You only confirmed twice** (requirements + results). Everything else was automatic.
|
|
72
|
+
|
|
73
|
+
---
|
|
74
|
+
|
|
75
|
+
## What `crew init` Creates
|
|
76
|
+
|
|
77
|
+
```
|
|
78
|
+
your-project/
|
|
79
|
+
├── INSTRUCTIONS.md ← AI behavior instructions (core file)
|
|
80
|
+
├── devcrew.yaml ← Project config (modes, specialists)
|
|
81
|
+
└── devcrew/
|
|
82
|
+
└── specs/ ← Shared specifications
|
|
83
|
+
```
|
|
84
|
+
|
|
85
|
+
AI reads `INSTRUCTIONS.md` and automatically assembles the team, following the PDEVI workflow.
|
|
86
|
+
|
|
87
|
+
---
|
|
88
|
+
|
|
89
|
+
## Core Concepts
|
|
90
|
+
|
|
91
|
+
### PDEVI Workflow
|
|
92
|
+
|
|
93
|
+
```mermaid
|
|
94
|
+
flowchart LR
|
|
95
|
+
P["Plan"] --> D["Design"]
|
|
96
|
+
D --> E["Execute"]
|
|
97
|
+
E --> V["Verify"]
|
|
98
|
+
V -->|Fail| I["Iterate"]
|
|
99
|
+
I -->|Design issue| D
|
|
100
|
+
I -->|Implementation issue| E
|
|
101
|
+
V -->|Pass| Done["✅ Done"]
|
|
102
|
+
```
|
|
103
|
+
|
|
104
|
+
Three modes for every scenario:
|
|
105
|
+
|
|
106
|
+
| Mode | Flow | Best for |
|
|
107
|
+
|------|------|----------|
|
|
108
|
+
| **Standard** | P → D → E → V → I | New features, refactoring |
|
|
109
|
+
| **Express** | P → E → V | Bug fixes |
|
|
110
|
+
| **Prototype** | P → D → E | Quick prototyping |
|
|
111
|
+
|
|
112
|
+
### Built-in Team (6 Roles)
|
|
113
|
+
|
|
114
|
+
| Role | Responsibility |
|
|
115
|
+
|------|---------------|
|
|
116
|
+
| 🎯 **PjM** Project Manager | Orchestration, mode inference, stage transitions |
|
|
117
|
+
| 📋 **PdM** Product Manager | Requirements analysis, PRD import, acceptance criteria |
|
|
118
|
+
| 🏗️ **Architect** | Tech decisions, task decomposition, dependency analysis |
|
|
119
|
+
| 💻 **Implementer** | Code generation, refactoring, dependency management |
|
|
120
|
+
| 🧪 **Tester** | Test execution, acceptance checks, coverage |
|
|
121
|
+
| 👀 **Reviewer** | Code review, security scanning, best practices |
|
|
122
|
+
|
|
123
|
+
Role switching is fully automatic — no manual assignment needed.
|
|
124
|
+
|
|
125
|
+
### Domain Specialists (29)
|
|
126
|
+
|
|
127
|
+
Beyond the core team, **29 domain specialists** across 10 fields, activated on demand:
|
|
128
|
+
|
|
129
|
+
> 🎮 Game Dev (8) · 🎨 UI/UX (3) · 🔒 Security (1) · ⚙️ DevOps (3) · 🧪 Testing (3) · 💻 Engineering (5) · 📊 Data (2) · 🤖 AI/ML (1) · 🌐 Web3 (1) · 🥽 Spatial Computing (2)
|
|
130
|
+
|
|
131
|
+
```yaml
|
|
132
|
+
# devcrew.yaml
|
|
133
|
+
specialists:
|
|
134
|
+
- game-designer
|
|
135
|
+
- security-engineer
|
|
136
|
+
```
|
|
137
|
+
|
|
138
|
+
```bash
|
|
139
|
+
crew agents # List all available specialists
|
|
140
|
+
```
|
|
141
|
+
|
|
142
|
+
> See the full [Specialist Directory](agents/README.md)
|
|
143
|
+
|
|
144
|
+
---
|
|
145
|
+
|
|
146
|
+
## Skill Commands
|
|
147
|
+
|
|
148
|
+
| Command | Purpose |
|
|
149
|
+
|---------|---------|
|
|
150
|
+
| `/crew:init` | Initialize workspace |
|
|
151
|
+
| `/crew:plan <name>` | Create a change and start working |
|
|
152
|
+
| `/crew:status` | Check current progress |
|
|
153
|
+
| `/crew:explore` | Discuss / analyze (no code changes) |
|
|
154
|
+
| `/crew:release` | Archive completed changes |
|
|
155
|
+
|
|
156
|
+
> Natural language works too — "show me the progress" = `/crew:status`
|
|
157
|
+
|
|
158
|
+
---
|
|
159
|
+
|
|
160
|
+
## Use Cases
|
|
161
|
+
|
|
162
|
+
| Scenario | You say | DevCrew does |
|
|
163
|
+
|----------|---------|-------------|
|
|
164
|
+
| Greenfield | "I have an idea, build from scratch" | Init → guide requirements → Standard |
|
|
165
|
+
| Existing PRD | "Here's the PRD, execute it" | Import PRD → refine → Standard |
|
|
166
|
+
| Mid-project | "Code exists, help me continue" | Scan code → establish baseline → Standard |
|
|
167
|
+
| Brainstorm | "Let's discuss the approach" | `/crew:explore` (no code changes) |
|
|
168
|
+
| Bug fix | "There's a bug, fix it fast" | Express mode |
|
|
169
|
+
| Refactor | "This code needs refactoring" | Standard (full workflow) |
|
|
170
|
+
| Prototype | "Build a quick prototype first" | Prototype mode |
|
|
171
|
+
| Learn codebase | "Help me understand this code" | `/crew:explore` (code analysis) |
|
|
172
|
+
|
|
173
|
+
---
|
|
174
|
+
|
|
175
|
+
## Architecture
|
|
176
|
+
|
|
177
|
+
```
|
|
178
|
+
┌─────────────────────────────────────────────┐
|
|
179
|
+
│ Tool Layer │
|
|
180
|
+
│ ┌──────────────┐ ┌──────────────────────┐ │
|
|
181
|
+
│ │ CLI │ │ Agent Skill │ │
|
|
182
|
+
│ │ crew init │ │ INSTRUCTIONS.md │ │
|
|
183
|
+
│ │ crew agents │ │ /crew: commands │ │
|
|
184
|
+
│ └──────────────┘ └──────────────────────┘ │
|
|
185
|
+
├─────────────────────────────────────────────┤
|
|
186
|
+
│ Protocol Layer (core, zero tool dependency) │
|
|
187
|
+
│ Directory conventions · File formats · │
|
|
188
|
+
│ PDEVI workflow · Communication rules │
|
|
189
|
+
└─────────────────────────────────────────────┘
|
|
190
|
+
```
|
|
191
|
+
|
|
192
|
+
> Even without CLI, manually placing `INSTRUCTIONS.md` works. CLI just makes it easier.
|
|
193
|
+
|
|
194
|
+
---
|
|
195
|
+
|
|
196
|
+
## Documentation
|
|
197
|
+
|
|
198
|
+
| Doc | Description |
|
|
199
|
+
|-----|-------------|
|
|
200
|
+
| [User Manual](docs/USER-MANUAL.md) | Detailed guide for 8 scenarios |
|
|
201
|
+
| [Best Practices](docs/examples/) | Scenario walkthrough examples |
|
|
202
|
+
| [Specialists](agents/README.md) | 29 specialists · 10 domains |
|
|
203
|
+
|
|
204
|
+
## Contributing
|
|
205
|
+
|
|
206
|
+
Contributions welcome! See [CONTRIBUTING.md](CONTRIBUTING.md).
|
|
207
|
+
|
|
208
|
+
## License
|
|
209
|
+
|
|
210
|
+
[MIT](LICENSE)
|
|
211
|
+
|
|
212
|
+
## Credits
|
|
213
|
+
|
|
214
|
+
Domain specialists adapted from the [agency-agents-zh](https://github.com/jnMetaCode/agency-agents-zh) project.
|
package/README.md
ADDED
|
@@ -0,0 +1,213 @@
|
|
|
1
|
+
> **语言 / Language**: **简体中文** · [English](README.en.md)
|
|
2
|
+
|
|
3
|
+
<div align="center">
|
|
4
|
+
<img src="docs/public/logo.svg" alt="DevCrew" width="180" />
|
|
5
|
+
|
|
6
|
+
# DevCrew
|
|
7
|
+
|
|
8
|
+
**给 AI 一套协作协议,让它像真正的团队一样工作。**
|
|
9
|
+
|
|
10
|
+
*帮你做好 harness!*
|
|
11
|
+
|
|
12
|
+
[](https://www.npmjs.com/package/@lordmos/devcrew)
|
|
13
|
+
[](LICENSE)
|
|
14
|
+
[](https://nodejs.org)
|
|
15
|
+
[](https://github.com/lordmos/dev-crew/pulls)
|
|
16
|
+
|
|
17
|
+
</div>
|
|
18
|
+
|
|
19
|
+
---
|
|
20
|
+
|
|
21
|
+
## 痛点
|
|
22
|
+
|
|
23
|
+
用 AI(Copilot、Claude、Cursor…)辅助开发时,你是否遇到过:
|
|
24
|
+
|
|
25
|
+
| 问题 | 表现 |
|
|
26
|
+
|------|------|
|
|
27
|
+
| 🧠 **无记忆** | 换个对话窗口,AI 忘了之前做了什么 |
|
|
28
|
+
| 🎭 **无分工** | AI 同时充当 PM + 架构师 + 开发 + 测试,顾此失彼 |
|
|
29
|
+
| 🎯 **会跑偏** | 做着做着偏离目标,没有检查点纠正 |
|
|
30
|
+
| 🔍 **质量盲区** | 没有审查环节,bug 和技术债悄悄积累 |
|
|
31
|
+
| ❓ **不知从何开始** | 面对新项目,不知道如何编排 AI 协作 |
|
|
32
|
+
|
|
33
|
+
**根因**:AI 缺少一套持久化的协作协议。DevCrew 就是那套协议。
|
|
34
|
+
|
|
35
|
+
---
|
|
36
|
+
|
|
37
|
+
## 30 秒上手
|
|
38
|
+
|
|
39
|
+
```bash
|
|
40
|
+
npm install -g @lordmos/devcrew
|
|
41
|
+
cd your-project
|
|
42
|
+
crew init
|
|
43
|
+
```
|
|
44
|
+
|
|
45
|
+
三步完成——你的项目现在拥有一个**开箱即用的 6 人 AI 开发团队** + **29 位可选领域专家**,全部自动编排。
|
|
46
|
+
|
|
47
|
+
适用于 **GitHub Copilot · Claude · ChatGPT · Cursor** 及任何 AI 平台。
|
|
48
|
+
|
|
49
|
+
---
|
|
50
|
+
|
|
51
|
+
## 它是怎么工作的
|
|
52
|
+
|
|
53
|
+
```
|
|
54
|
+
你: 我要给 API 加认证中间件
|
|
55
|
+
|
|
56
|
+
AI: 📋 创建变更 add-api-auth,模式: Standard
|
|
57
|
+
Plan — 需求整理:
|
|
58
|
+
- 目标: 为所有 /api/ 路由添加 JWT 认证
|
|
59
|
+
- 验收标准: ☐ 未携带 token 返回 401 ☐ 过期 token 返回 401
|
|
60
|
+
请确认。
|
|
61
|
+
|
|
62
|
+
你: 确认
|
|
63
|
+
|
|
64
|
+
AI: Design → Execute → Verify — 全部通过。请确认验收。
|
|
65
|
+
|
|
66
|
+
你: 确认
|
|
67
|
+
|
|
68
|
+
AI: ✅ 变更 add-api-auth 完成。
|
|
69
|
+
```
|
|
70
|
+
|
|
71
|
+
**你只确认了两次**(需求 + 结果),其余全部自动。
|
|
72
|
+
|
|
73
|
+
---
|
|
74
|
+
|
|
75
|
+
## `crew init` 做了什么
|
|
76
|
+
|
|
77
|
+
```
|
|
78
|
+
your-project/
|
|
79
|
+
├── INSTRUCTIONS.md ← AI 行为指令(核心文件)
|
|
80
|
+
├── devcrew.yaml ← 项目配置(模式、专家选择)
|
|
81
|
+
└── devcrew/
|
|
82
|
+
└── specs/ ← 共享规约
|
|
83
|
+
```
|
|
84
|
+
|
|
85
|
+
AI 读取 `INSTRUCTIONS.md` 后,自动建立团队,按 PDEVI 流程协作。
|
|
86
|
+
|
|
87
|
+
---
|
|
88
|
+
|
|
89
|
+
## 核心概念
|
|
90
|
+
|
|
91
|
+
### PDEVI 工作流
|
|
92
|
+
|
|
93
|
+
```mermaid
|
|
94
|
+
flowchart LR
|
|
95
|
+
P["Plan"] --> D["Design"]
|
|
96
|
+
D --> E["Execute"]
|
|
97
|
+
E --> V["Verify"]
|
|
98
|
+
V -->|不通过| I["Iterate"]
|
|
99
|
+
I -->|方案问题| D
|
|
100
|
+
I -->|实现问题| E
|
|
101
|
+
V -->|通过| Done["✅ 完成"]
|
|
102
|
+
```
|
|
103
|
+
|
|
104
|
+
三种模式,覆盖所有场景:
|
|
105
|
+
|
|
106
|
+
| 模式 | 流程 | 适用 |
|
|
107
|
+
|------|------|------|
|
|
108
|
+
| **Standard** | P → D → E → V → I | 新功能、重构 |
|
|
109
|
+
| **Express** | P → E → V | Bug 修复 |
|
|
110
|
+
| **Prototype** | P → D → E | 快速原型 |
|
|
111
|
+
|
|
112
|
+
### 内建团队(6 人)
|
|
113
|
+
|
|
114
|
+
| 角色 | 职责 |
|
|
115
|
+
|------|------|
|
|
116
|
+
| 🎯 **PjM** 项目经理 | 调度编排、模式推断、阶段推进 |
|
|
117
|
+
| 📋 **PdM** 产品经理 | 需求梳理、PRD 导入、验收标准 |
|
|
118
|
+
| 🏗️ **Architect** 架构师 | 技术选型、任务分解、依赖分析 |
|
|
119
|
+
| 💻 **Implementer** 开发 | 代码生成、重构、依赖安装 |
|
|
120
|
+
| 🧪 **Tester** 测试 | 测试执行、验收检查、覆盖率 |
|
|
121
|
+
| 👀 **Reviewer** 审查 | 规范检查、安全扫描、最佳实践 |
|
|
122
|
+
|
|
123
|
+
角色切换完全自动,无需手动分配。
|
|
124
|
+
|
|
125
|
+
### 领域专家(29 位)
|
|
126
|
+
|
|
127
|
+
核心团队之外,**29 位领域专家**覆盖 10 个领域,按需激活:
|
|
128
|
+
|
|
129
|
+
> 🎮 游戏开发(8)· 🎨 UI/UX(3)· 🔒 安全(1)· ⚙️ DevOps(3)· 🧪 测试(3)· 💻 工程(5)· 📊 数据(2)· 🤖 AI/ML(1)· 🌐 Web3(1)· 🥽 空间计算(2)
|
|
130
|
+
|
|
131
|
+
```yaml
|
|
132
|
+
# devcrew.yaml
|
|
133
|
+
specialists:
|
|
134
|
+
- game-designer
|
|
135
|
+
- security-engineer
|
|
136
|
+
```
|
|
137
|
+
|
|
138
|
+
```bash
|
|
139
|
+
crew agents # 查看所有可用专家
|
|
140
|
+
```
|
|
141
|
+
|
|
142
|
+
> 完整列表见 [领域专家目录](agents/README.md)
|
|
143
|
+
|
|
144
|
+
---
|
|
145
|
+
|
|
146
|
+
## Skill 指令
|
|
147
|
+
|
|
148
|
+
| 指令 | 用途 |
|
|
149
|
+
|------|------|
|
|
150
|
+
| `/crew:init` | 初始化工作区 |
|
|
151
|
+
| `/crew:plan <名称>` | 创建变更并开始工作 |
|
|
152
|
+
| `/crew:status` | 查看当前进度 |
|
|
153
|
+
| `/crew:explore` | 讨论 / 分析(不改代码) |
|
|
154
|
+
| `/crew:release` | 归档已完成变更 |
|
|
155
|
+
|
|
156
|
+
> 自然语言同样有效——"帮我看看进度" = `/crew:status`
|
|
157
|
+
|
|
158
|
+
---
|
|
159
|
+
|
|
160
|
+
## 使用场景
|
|
161
|
+
|
|
162
|
+
| 场景 | 你说 | DevCrew 做 |
|
|
163
|
+
|------|------|------------|
|
|
164
|
+
| 从零开始 | "有个想法,从零构建" | 初始化 → 引导需求 → Standard |
|
|
165
|
+
| 已有 PRD | "需求文档在这,执行吧" | 导入 PRD → 提炼 → Standard |
|
|
166
|
+
| 中途接入 | "代码已有,帮我续上" | 扫描代码 → 建基线 → Standard |
|
|
167
|
+
| 头脑风暴 | "讨论一下方案" | `/crew:explore`(不改代码) |
|
|
168
|
+
| Bug 修复 | "有个 bug,快修" | Express 模式 |
|
|
169
|
+
| 代码重构 | "这段代码要重构" | Standard 完整流程 |
|
|
170
|
+
| 快速原型 | "先做个原型验证" | Prototype 模式 |
|
|
171
|
+
| 学习代码库 | "帮我理解这段代码" | `/crew:explore`(分析代码) |
|
|
172
|
+
|
|
173
|
+
---
|
|
174
|
+
|
|
175
|
+
## 架构
|
|
176
|
+
|
|
177
|
+
```
|
|
178
|
+
┌─────────────────────────────────────────────┐
|
|
179
|
+
│ 工具层 │
|
|
180
|
+
│ ┌──────────────┐ ┌──────────────────────┐ │
|
|
181
|
+
│ │ CLI │ │ Agent Skill │ │
|
|
182
|
+
│ │ crew init │ │ INSTRUCTIONS.md │ │
|
|
183
|
+
│ │ crew agents │ │ /crew: 指令 │ │
|
|
184
|
+
│ └──────────────┘ └──────────────────────┘ │
|
|
185
|
+
├─────────────────────────────────────────────┤
|
|
186
|
+
│ 协议层(核心,零工具依赖) │
|
|
187
|
+
│ 目录约定 · 文件格式 · PDEVI · 通信规则 │
|
|
188
|
+
└─────────────────────────────────────────────┘
|
|
189
|
+
```
|
|
190
|
+
|
|
191
|
+
> 即使不装 CLI,手动放入 `INSTRUCTIONS.md` 也能工作。CLI 只是让流程更方便。
|
|
192
|
+
|
|
193
|
+
---
|
|
194
|
+
|
|
195
|
+
## 文档
|
|
196
|
+
|
|
197
|
+
| 文档 | 说明 |
|
|
198
|
+
|------|------|
|
|
199
|
+
| [用户手册](docs/USER-MANUAL.md) | 8 种场景详细指南 |
|
|
200
|
+
| [最佳实践](docs/examples/) | 场景串联示例 |
|
|
201
|
+
| [领域专家](agents/README.md) | 29 位专家 · 10 个领域 |
|
|
202
|
+
|
|
203
|
+
## 贡献
|
|
204
|
+
|
|
205
|
+
欢迎参与!详见 [CONTRIBUTING.md](CONTRIBUTING.md)。
|
|
206
|
+
|
|
207
|
+
## 许可证
|
|
208
|
+
|
|
209
|
+
[MIT](LICENSE)
|
|
210
|
+
|
|
211
|
+
## 致谢
|
|
212
|
+
|
|
213
|
+
领域专家部分基于 [agency-agents-zh](https://github.com/jnMetaCode/agency-agents-zh) 项目改编。
|
package/agents/README.md
ADDED
|
@@ -0,0 +1,136 @@
|
|
|
1
|
+
# 领域专家目录
|
|
2
|
+
|
|
3
|
+
> 共 **29** 位专家 · **10** 个领域包
|
|
4
|
+
|
|
5
|
+
DevCrew 内建 6 个核心角色(PjM/PdM/Architect/Implementer/Tester/Reviewer)。领域专家是**可选的补充角色**——为特定项目类型提供专业知识,增强核心团队的领域深度。
|
|
6
|
+
|
|
7
|
+
## 使用方式
|
|
8
|
+
|
|
9
|
+
```yaml
|
|
10
|
+
# devcrew.yaml
|
|
11
|
+
specialists:
|
|
12
|
+
- game-designer # 按需激活
|
|
13
|
+
- ui-designer
|
|
14
|
+
- security-engineer
|
|
15
|
+
```
|
|
16
|
+
|
|
17
|
+
PjM 在 `/crew:init` 时读取配置,加载对应专家 prompt,在 PDEVI 的相关阶段自动参与。每个专家独立,可自由组合。
|
|
18
|
+
|
|
19
|
+
---
|
|
20
|
+
|
|
21
|
+
## 专家速查表
|
|
22
|
+
|
|
23
|
+
### 🎮 游戏开发(8)
|
|
24
|
+
|
|
25
|
+
| 专家 | 文件 | 阶段 | 核心能力 |
|
|
26
|
+
|------|------|------|---------|
|
|
27
|
+
| 游戏设计师 | [`game-designer.md`](game-designer.md) | P·D·V | GDD、游戏循环、经济平衡 |
|
|
28
|
+
| 关卡设计师 | [`level-designer.md`](level-designer.md) | D·E·V | 关卡节奏、空间叙事、难度曲线 |
|
|
29
|
+
| 叙事设计师 | [`narrative-designer.md`](narrative-designer.md) | P·D·V | 剧情架构、对话系统、分支叙事 |
|
|
30
|
+
| 技术美术 | [`technical-artist.md`](technical-artist.md) | D·E·V | Shader、性能美术、管线工具 |
|
|
31
|
+
| 游戏音频工程师 | [`game-audio-engineer.md`](game-audio-engineer.md) | D·E·V | 自适应音频、Wwise/FMOD、空间音效 |
|
|
32
|
+
| Unity 专家 | [`unity-specialist.md`](unity-specialist.md) | D·E·V | DOTS/ECS、URP/HDRP、Addressables |
|
|
33
|
+
| Godot 专家 | [`godot-specialist.md`](godot-specialist.md) | D·E·V | GDScript/C#、Scene 树、信号系统 |
|
|
34
|
+
| Unreal 专家 | [`unreal-specialist.md`](unreal-specialist.md) | D·E·V | Blueprint/C++、GAS、Nanite/Lumen |
|
|
35
|
+
|
|
36
|
+
### 🎨 UI/UX 设计(3)
|
|
37
|
+
|
|
38
|
+
| 专家 | 文件 | 阶段 | 核心能力 |
|
|
39
|
+
|------|------|------|---------|
|
|
40
|
+
| UI 设计师 | [`ui-designer.md`](ui-designer.md) | D·E·V | 设计系统、组件库、响应式 |
|
|
41
|
+
| UX 架构师 | [`ux-architect.md`](ux-architect.md) | P·D·V | 信息架构、用户流程、可用性 |
|
|
42
|
+
| UX 研究员 | [`ux-researcher.md`](ux-researcher.md) | P·D·V | 用研方法、可用性测试、数据洞察 |
|
|
43
|
+
|
|
44
|
+
### 🔒 安全(1)
|
|
45
|
+
|
|
46
|
+
| 专家 | 文件 | 阶段 | 核心能力 |
|
|
47
|
+
|------|------|------|---------|
|
|
48
|
+
| 安全工程师 | [`security-engineer.md`](security-engineer.md) | D·E·V | 威胁建模、OWASP、安全审查 |
|
|
49
|
+
|
|
50
|
+
### ⚙️ DevOps & 可靠性(3)
|
|
51
|
+
|
|
52
|
+
| 专家 | 文件 | 阶段 | 核心能力 |
|
|
53
|
+
|------|------|------|---------|
|
|
54
|
+
| DevOps 工程师 | [`devops-engineer.md`](devops-engineer.md) | D·E·V | CI/CD、容器化、IaC、监控 |
|
|
55
|
+
| SRE | [`sre.md`](sre.md) | D·E·V | SLO/错误预算、混沌工程、容量规划 |
|
|
56
|
+
| 事件响应指挥官 | [`incident-response-commander.md`](incident-response-commander.md) | E·V | 事件分级、根因分析、Runbook |
|
|
57
|
+
|
|
58
|
+
### 🧪 测试 & QA(3)
|
|
59
|
+
|
|
60
|
+
| 专家 | 文件 | 阶段 | 核心能力 |
|
|
61
|
+
|------|------|------|---------|
|
|
62
|
+
| API 测试专家 | [`api-tester.md`](api-tester.md) | D·E·V | 契约测试、自动化、性能基线 |
|
|
63
|
+
| 性能工程师 | [`performance-engineer.md`](performance-engineer.md) | D·E·V | 负载测试、性能基线、瓶颈分析 |
|
|
64
|
+
| 无障碍审计师 | [`accessibility-auditor.md`](accessibility-auditor.md) | D·E·V | WCAG、键盘/屏幕阅读器、对比度 |
|
|
65
|
+
|
|
66
|
+
### 💻 工程专项(5)
|
|
67
|
+
|
|
68
|
+
| 专家 | 文件 | 阶段 | 核心能力 |
|
|
69
|
+
|------|------|------|---------|
|
|
70
|
+
| 前端专家 | [`frontend-specialist.md`](frontend-specialist.md) | D·E·V | React/Vue、状态管理、性能优化 |
|
|
71
|
+
| 后端架构师 | [`backend-architect.md`](backend-architect.md) | D·E·V | API 设计、数据建模、分布式架构 |
|
|
72
|
+
| 移动开发专家 | [`mobile-developer.md`](mobile-developer.md) | D·E·V | iOS/Android/跨平台、离线优先 |
|
|
73
|
+
| 嵌入式工程师 | [`embedded-engineer.md`](embedded-engineer.md) | D·E·V | 固件/RTOS、外设驱动、低功耗 |
|
|
74
|
+
| 技术写作专家 | [`technical-writer.md`](technical-writer.md) | P·D·V | API 文档、用户指南、术语一致性 |
|
|
75
|
+
|
|
76
|
+
### 📊 数据 & 数据库(2)
|
|
77
|
+
|
|
78
|
+
| 专家 | 文件 | 阶段 | 核心能力 |
|
|
79
|
+
|------|------|------|---------|
|
|
80
|
+
| 数据工程师 | [`data-engineer.md`](data-engineer.md) | D·E·V | ETL/ELT、数据管道、数据质量 |
|
|
81
|
+
| 数据库专家 | [`database-specialist.md`](database-specialist.md) | D·E·V | 数据建模、查询优化、高可用 |
|
|
82
|
+
|
|
83
|
+
### 🤖 AI/ML(1)
|
|
84
|
+
|
|
85
|
+
| 专家 | 文件 | 阶段 | 核心能力 |
|
|
86
|
+
|------|------|------|---------|
|
|
87
|
+
| AI 工程师 | [`ai-engineer.md`](ai-engineer.md) | D·E·V | 模型训练、RAG/LLM、MLOps |
|
|
88
|
+
|
|
89
|
+
### 🌐 Web3(1)
|
|
90
|
+
|
|
91
|
+
| 专家 | 文件 | 阶段 | 核心能力 |
|
|
92
|
+
|------|------|------|---------|
|
|
93
|
+
| Solidity 工程师 | [`solidity-engineer.md`](solidity-engineer.md) | D·E·V | 智能合约、Gas 优化、安全审计 |
|
|
94
|
+
|
|
95
|
+
### 🥽 空间计算(2)
|
|
96
|
+
|
|
97
|
+
| 专家 | 文件 | 阶段 | 核心能力 |
|
|
98
|
+
|------|------|------|---------|
|
|
99
|
+
| visionOS 工程师 | [`visionos-engineer.md`](visionos-engineer.md) | D·E·V | RealityKit、SwiftUI、空间交互 |
|
|
100
|
+
| XR 开发专家 | [`xr-developer.md`](xr-developer.md) | D·E·V | Unity/WebXR、VR/AR/MR、多平台 |
|
|
101
|
+
|
|
102
|
+
---
|
|
103
|
+
|
|
104
|
+
## 推荐领域包
|
|
105
|
+
|
|
106
|
+
| 场景 | 推荐专家组合 | YAML 示例 |
|
|
107
|
+
|------|------------|----------|
|
|
108
|
+
| **Web 全栈** | frontend + backend + ui-designer + security | `[frontend-specialist, backend-architect, ui-designer, security-engineer]` |
|
|
109
|
+
| **游戏 (Unity)** | game-designer + unity + level-designer + ui | `[game-designer, unity-specialist, level-designer, ui-designer]` |
|
|
110
|
+
| **移动应用** | mobile + ui-designer + ux-architect + api-tester | `[mobile-developer, ui-designer, ux-architect, api-tester]` |
|
|
111
|
+
| **云原生/微服务** | backend + devops + sre + security | `[backend-architect, devops-engineer, sre, security-engineer]` |
|
|
112
|
+
| **数据平台** | data-engineer + database + backend | `[data-engineer, database-specialist, backend-architect]` |
|
|
113
|
+
| **AI 应用** | ai-engineer + backend + performance | `[ai-engineer, backend-architect, performance-engineer]` |
|
|
114
|
+
| **嵌入式/IoT** | embedded + security + technical-writer | `[embedded-engineer, security-engineer, technical-writer]` |
|
|
115
|
+
| **安全合规** | security + accessibility + api-tester + sre | `[security-engineer, accessibility-auditor, api-tester, sre]` |
|
|
116
|
+
|
|
117
|
+
---
|
|
118
|
+
|
|
119
|
+
## 专家与核心团队的关系
|
|
120
|
+
|
|
121
|
+
```mermaid
|
|
122
|
+
graph TD
|
|
123
|
+
PjM["PjM 调度"]
|
|
124
|
+
Core["核心团队<br>PdM · Architect · Implementer · Tester · Reviewer"]
|
|
125
|
+
Spec["领域专家<br>(按配置激活)"]
|
|
126
|
+
|
|
127
|
+
PjM -->|编排| Core
|
|
128
|
+
PjM -->|按阶段调度| Spec
|
|
129
|
+
Spec -->|领域知识补充| Core
|
|
130
|
+
```
|
|
131
|
+
|
|
132
|
+
**规则**:
|
|
133
|
+
- 领域专家**补充**核心角色,不替代
|
|
134
|
+
- PjM 在 PDEVI 对应阶段调度专家参与
|
|
135
|
+
- 专家产出合并到 proposal.md / design.md(不创建独立文件)
|
|
136
|
+
- 未配置专家时,核心团队独立完成所有工作(零影响)
|
|
@@ -0,0 +1,48 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: 无障碍审计师
|
|
3
|
+
category: testing
|
|
4
|
+
stages: [design, execute, verify]
|
|
5
|
+
source: agency-agents-zh/testing/testing-accessibility-auditor.md
|
|
6
|
+
---
|
|
7
|
+
|
|
8
|
+
# 领域专家:无障碍审计师
|
|
9
|
+
|
|
10
|
+
你是一位资深无障碍审计师。你确保产品对所有用户可用——包括视障、听障、运动障碍和认知障碍用户。你以 WCAG 标准为基准,以用户体验为目标。
|
|
11
|
+
|
|
12
|
+
## 在 PDEVI 中的职责
|
|
13
|
+
|
|
14
|
+
### Design 阶段 → 补充 design.md
|
|
15
|
+
|
|
16
|
+
**无障碍标准**
|
|
17
|
+
| 维度 | 目标等级 | 关键要求 |
|
|
18
|
+
|------|---------|---------|
|
|
19
|
+
| 感知性 | WCAG AA | 文本替代、字幕、对比度 ≥4.5:1 |
|
|
20
|
+
| 可操作性 | WCAG AA | 键盘可达、触控 ≥44px、无闪烁 |
|
|
21
|
+
| 可理解性 | WCAG AA | 清晰标签、可预测行为、错误建议 |
|
|
22
|
+
| 健壮性 | WCAG AA | 语义 HTML、ARIA 正确使用 |
|
|
23
|
+
|
|
24
|
+
### Execute 阶段 → 辅助 Implementer
|
|
25
|
+
|
|
26
|
+
- 审查 HTML 语义结构(heading 层级/landmark/列表)
|
|
27
|
+
- 验证 ARIA 属性使用正确性(role/state/property)
|
|
28
|
+
- 检查键盘导航流(Tab 顺序/焦点管理/快捷键)
|
|
29
|
+
- 验证屏幕阅读器兼容性(VoiceOver/NVDA/JAWS)
|
|
30
|
+
- 检查颜色对比度和色盲友好性
|
|
31
|
+
|
|
32
|
+
### Verify 阶段 → 补充验证标准
|
|
33
|
+
|
|
34
|
+
- [ ] 自动化扫描通过(axe-core/Lighthouse 0 错误)?
|
|
35
|
+
- [ ] 纯键盘可完成所有核心任务?
|
|
36
|
+
- [ ] 屏幕阅读器朗读内容有意义?
|
|
37
|
+
- [ ] 色彩对比度 ≥4.5:1(正常文本)/ ≥3:1(大文本)?
|
|
38
|
+
- [ ] 焦点指示器清晰可见?
|
|
39
|
+
- [ ] 动画可关闭(prefers-reduced-motion)?
|
|
40
|
+
- [ ] 表单错误提示关联到对应字段?
|
|
41
|
+
|
|
42
|
+
## 关键规则
|
|
43
|
+
|
|
44
|
+
1. **语义 HTML 优先于 ARIA**:能用原生元素就不加 ARIA
|
|
45
|
+
2. **键盘是底线**:如果键盘不能操作,就是不可访问的
|
|
46
|
+
3. **自动化只发现 30% 问题**:自动工具 + 手动测试 + 辅助技术实测
|
|
47
|
+
4. **对比度不是建议是要求**:低于标准 = 不达标
|
|
48
|
+
5. **无障碍是功能不是优化**:和其他功能一样必须测试和维护
|
|
@@ -0,0 +1,52 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: AI 工程师
|
|
3
|
+
category: ai-ml
|
|
4
|
+
stages: [design, execute, verify]
|
|
5
|
+
source: agency-agents-zh/engineering/engineering-ai-engineer.md
|
|
6
|
+
---
|
|
7
|
+
|
|
8
|
+
# 领域专家:AI 工程师
|
|
9
|
+
|
|
10
|
+
你是一位资深 AI/ML 工程师。你将机器学习模型从实验带到生产——涵盖模型选型、训练管道、推理优化、LLM 应用和 MLOps。
|
|
11
|
+
|
|
12
|
+
## 在 PDEVI 中的职责
|
|
13
|
+
|
|
14
|
+
### Design 阶段 → 补充 design.md
|
|
15
|
+
|
|
16
|
+
**AI/ML 架构**
|
|
17
|
+
| 维度 | 决策 | 理由 |
|
|
18
|
+
|------|------|------|
|
|
19
|
+
| 问题类型 | [分类/生成/检索/推荐] | [业务目标] |
|
|
20
|
+
| 模型方案 | [预训练微调/RAG/从头训练/API 调用] | [数据量/精度] |
|
|
21
|
+
| 框架 | [PyTorch/TensorFlow/HuggingFace/LangChain] | [生态/需求] |
|
|
22
|
+
| 推理部署 | [API 服务/边端/批处理] | [延迟/成本] |
|
|
23
|
+
| 评估指标 | [准确率/F1/BLEU/自定义] | [业务含义] |
|
|
24
|
+
|
|
25
|
+
**LLM 应用设计**(如适用)
|
|
26
|
+
- Prompt 工程策略(Few-shot/CoT/Tool-use)
|
|
27
|
+
- RAG 架构(向量库选型/分块策略/检索排序)
|
|
28
|
+
- 安全护栏(内容过滤/幻觉检测/成本控制)
|
|
29
|
+
|
|
30
|
+
### Execute 阶段 → 辅助 Implementer
|
|
31
|
+
|
|
32
|
+
- 搭建训练/微调管道和实验追踪(MLflow/W&B)
|
|
33
|
+
- 实现特征工程和数据预处理管道
|
|
34
|
+
- 构建推理服务(模型服务化/缓存/批处理)
|
|
35
|
+
- 实现 RAG 系统(嵌入/索引/检索/生成)
|
|
36
|
+
- 编写评估套件和 A/B 测试框架
|
|
37
|
+
|
|
38
|
+
### Verify 阶段 → 补充验证标准
|
|
39
|
+
|
|
40
|
+
- [ ] 模型指标达到基线(离线评估通过)?
|
|
41
|
+
- [ ] 推理延迟和吞吐量达标?
|
|
42
|
+
- [ ] 模型大小和内存在部署环境预算内?
|
|
43
|
+
- [ ] 边界情况和对抗样本测试通过?
|
|
44
|
+
- [ ] 模型版本化和回滚机制就绪?
|
|
45
|
+
|
|
46
|
+
## 关键规则
|
|
47
|
+
|
|
48
|
+
1. **先基线后优化**:用最简单的方案建立基线,再迭代
|
|
49
|
+
2. **数据质量 > 模型复杂度**:垃圾数据无法用更大的模型拯救
|
|
50
|
+
3. **实验可复现**:固定种子/版本化数据/记录超参数
|
|
51
|
+
4. **LLM 不是万能的**:能用规则解决的不用模型
|
|
52
|
+
5. **成本要可控**:监控 Token 消耗/GPU 时间,设置预算告警
|