claude_dev
37579ebdcd
- docs/PRD.md - 产品需求文档（移到根目录） - docs/design/01-design-v0.1.md - 系统设计文档（移到子目录） Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
751 lines
21 KiB
Markdown
751 lines
21 KiB
Markdown
# sanguo_moziplus_v3 设计文档 v0.1
|
||
|
||
**项目名称**: sanguo_moziplus_v3
|
||
**版本**: v0.1
|
||
**创建日期**: 2026-06-29
|
||
**状态**: Draft
|
||
|
||
---
|
||
|
||
## 1. 架构概述
|
||
|
||
### 1.1 系统架构图
|
||
|
||
```
|
||
┌─────────────────────────────────────────────────────────────────┐
|
||
│ Claude Code CLI │
|
||
│ (用户交互层) │
|
||
└─────────────────────────────┬───────────────────────────────────┘
|
||
│
|
||
┌─────────────────────────────▼───────────────────────────────────┐
|
||
│ CLAUDE.md (规则层) │
|
||
│ - 协作规范 │
|
||
│ - 角色分工 │
|
||
│ - Git 规范 │
|
||
└─────────────────────────────┬───────────────────────────────────┘
|
||
│
|
||
┌─────────────────────────────▼───────────────────────────────────┐
|
||
│ Superpowers (能力层) │
|
||
│ - Planning Skill │
|
||
│ - Review Skill │
|
||
│ - Debug Skill │
|
||
└─────────────────────────────┬───────────────────────────────────┘
|
||
│
|
||
┌─────────────────────────────▼───────────────────────────────────┐
|
||
│ CCB (通信层) │
|
||
│ ┌─────────┐ ┌─────────┐ ┌─────────┐ ┌─────────┐ │
|
||
│ │Bridge │ │Bridge │ │Bridge │ │Bridge │ │
|
||
│ │Codex │ │Gemini │ │GPT │ │Custom │ │
|
||
│ └────┬────┘ └────┬────┘ └────┬────┘ └────┬────┘ │
|
||
└───────┼────────────┼────────────┼────────────┼───────────────────┘
|
||
│ │ │ │
|
||
▼ ▼ ▼ ▼
|
||
┌───────┐ ┌───────┐ ┌───────┐ ┌───────┐
|
||
│Codex │ │Gemini │ │GPT │ │Custom │
|
||
│API │ │API │ │API │ │API │
|
||
└───────┘ └───────┘ └───────┘ └───────┘
|
||
|
||
┌─────────────────────────────────────────────────────────────────┐
|
||
│ 任务存储层 (SQLite) │
|
||
│ - tasks (任务表) │
|
||
│ - collaborations (协作记录表) │
|
||
│ - costs (成本表) │
|
||
└─────────────────────────────────────────────────────────────────┘
|
||
```
|
||
|
||
### 1.2 核心设计原则
|
||
|
||
1. **分层解耦**: 规则层、能力层、通信层清晰分离
|
||
2. **可扩展**: 桥接器插件化，支持新增 AI 模型
|
||
3. **可配置**: 通过配置文件定义角色和协作模式
|
||
4. **成本优先**: Claude 只做决策，体力活交给便宜 AI
|
||
|
||
---
|
||
|
||
## 2. 核心组件设计
|
||
|
||
### 2.1 CCB 桥接器 (Bridge)
|
||
|
||
#### 2.1.1 桥接器接口
|
||
|
||
```python
|
||
class BridgeAgent(ABC):
|
||
"""桥接器抽象基类"""
|
||
|
||
@abstractmethod
|
||
async def ask(self, prompt: str, context: dict) -> BridgeResponse:
|
||
"""向 AI 发送请求"""
|
||
pass
|
||
|
||
@abstractmethod
|
||
def get_cost(self) -> TokenCost:
|
||
"""获取成本统计"""
|
||
pass
|
||
|
||
@abstractmethod
|
||
def get_capabilities(self) -> List[str]:
|
||
"""获取能力列表"""
|
||
pass
|
||
```
|
||
|
||
#### 2.1.2 具体实现
|
||
|
||
**Codex Bridge**:
|
||
```python
|
||
class CodexBridge(BridgeAgent):
|
||
def __init__(self, api_key: str, endpoint: str):
|
||
self.api_key = api_key
|
||
self.endpoint = endpoint
|
||
|
||
async def ask(self, prompt: str, context: dict) -> BridgeResponse:
|
||
# 调用 Codex API
|
||
response = await self._call_codex(prompt, context)
|
||
return BridgeResponse(
|
||
content=response.content,
|
||
tokens_used=response.tokens,
|
||
cost=response.tokens * CODEX_PRICE
|
||
)
|
||
```
|
||
|
||
**Gemini Bridge**:
|
||
```python
|
||
class GeminiBridge(BridgeAgent):
|
||
def __init__(self, api_key: str, model: str = "gemini-pro"):
|
||
self.api_key = api_key
|
||
self.model = model
|
||
|
||
async def ask(self, prompt: str, context: dict) -> BridgeResponse:
|
||
# 调用 Gemini API
|
||
response = await self._call_gemini(prompt, context)
|
||
return BridgeResponse(
|
||
content=response.content,
|
||
tokens_used=response.tokens,
|
||
cost=response.tokens * GEMINI_PRICE
|
||
)
|
||
```
|
||
|
||
### 2.2 任务编排器 (Orchestrator)
|
||
|
||
#### 2.2.1 任务模型
|
||
|
||
```python
|
||
from enum import Enum
|
||
from datetime import datetime
|
||
from typing import Optional, List, Dict, Any
|
||
|
||
class TaskType(Enum):
|
||
PLAN = "plan" # 规划任务
|
||
CODE_BACKEND = "code_backend" # 后端编码
|
||
CODE_FRONTEND = "code_frontend" # 前端编码
|
||
REVIEW = "review" # 代码审查
|
||
VERIFY = "verify" # 验收测试
|
||
|
||
class TaskStatus(Enum):
|
||
PENDING = "pending"
|
||
IN_PROGRESS = "in_progress"
|
||
COMPLETED = "completed"
|
||
FAILED = "failed"
|
||
BLOCKED = "blocked"
|
||
|
||
class AgentRole(Enum):
|
||
CLAUDE = "claude" # PM / 架构师
|
||
CODEX = "codex" # 后端开发
|
||
GEMINI = "gemini" # 前端开发
|
||
GPT = "gpt" # 通用助手
|
||
|
||
@dataclass
|
||
class TokenCost:
|
||
model: str
|
||
input_tokens: int
|
||
output_tokens: int
|
||
total_cost: float
|
||
|
||
@property
|
||
def total_tokens(self) -> int:
|
||
return self.input_tokens + self.output_tokens
|
||
|
||
@dataclass
|
||
class Task:
|
||
id: str
|
||
type: TaskType
|
||
title: str
|
||
description: str
|
||
assigned_to: AgentRole
|
||
status: TaskStatus = TaskStatus.PENDING
|
||
dependencies: List[str] = field(default_factory=list)
|
||
input: Dict[str, Any] = field(default_factory=dict)
|
||
output: Dict[str, Any] = field(default_factory=dict)
|
||
cost: Optional[TokenCost] = None
|
||
created_at: datetime = field(default_factory=datetime.now)
|
||
completed_at: Optional[datetime] = None
|
||
error_message: Optional[str] = None
|
||
|
||
def can_execute(self, completed_tasks: set) -> bool:
|
||
"""检查任务是否可以执行（依赖已完成）"""
|
||
return all(dep in completed_tasks for dep in self.dependencies)
|
||
```
|
||
|
||
#### 2.2.2 编排器实现
|
||
|
||
```python
|
||
class TaskOrchestrator:
|
||
"""任务编排器"""
|
||
|
||
def __init__(self, db_path: str):
|
||
self.db = sqlite3.connect(db_path)
|
||
self.bridges: Dict[AgentRole, BridgeAgent] = {}
|
||
self._init_bridges()
|
||
|
||
def _init_bridges(self):
|
||
"""初始化桥接器"""
|
||
config = self._load_config()
|
||
if config.get("codex", {}).get("enabled"):
|
||
self.bridges[AgentRole.CODEX] = CodexBridge(...)
|
||
if config.get("gemini", {}).get("enabled"):
|
||
self.bridges[AgentRole.GEMINI] = GeminiBridge(...)
|
||
|
||
async def create_task(self, task: Task) -> str:
|
||
"""创建任务"""
|
||
await self.db.execute(
|
||
"INSERT INTO tasks ...",
|
||
task.to_dict()
|
||
)
|
||
return task.id
|
||
|
||
async def execute_task(self, task_id: str) -> Task:
|
||
"""执行任务"""
|
||
task = await self._get_task(task_id)
|
||
|
||
# 检查依赖
|
||
if not task.can_execute(await self._get_completed_tasks()):
|
||
raise TaskBlockedError("Dependencies not met")
|
||
|
||
# 获取桥接器
|
||
bridge = self.bridges.get(task.assigned_to)
|
||
if not bridge:
|
||
raise BridgeNotFoundError(f"No bridge for {task.assigned_to}")
|
||
|
||
# 执行任务
|
||
task.status = TaskStatus.IN_PROGRESS
|
||
await self._update_task(task)
|
||
|
||
try:
|
||
response = await bridge.ask(
|
||
prompt=task.description,
|
||
context=task.input
|
||
)
|
||
|
||
task.output = response.content
|
||
task.cost = response.cost
|
||
task.status = TaskStatus.COMPLETED
|
||
task.completed_at = datetime.now()
|
||
|
||
except Exception as e:
|
||
task.status = TaskStatus.FAILED
|
||
task.error_message = str(e)
|
||
|
||
await self._update_task(task)
|
||
return task
|
||
```
|
||
|
||
### 2.3 成本追踪器 (CostTracker)
|
||
|
||
```python
|
||
class CostTracker:
|
||
"""成本追踪器"""
|
||
|
||
async def record_cost(self, task_id: str, cost: TokenCost):
|
||
"""记录成本"""
|
||
await self.db.execute(
|
||
"INSERT INTO costs (task_id, model, input_tokens, output_tokens, total_cost) VALUES (?, ?, ?, ?, ?)",
|
||
(task_id, cost.model, cost.input_tokens, cost.output_tokens, cost.total_cost)
|
||
)
|
||
|
||
async def get_cost_report(self, project_id: str) -> CostReport:
|
||
"""获取成本报告"""
|
||
rows = await self.db.execute(
|
||
"SELECT model, SUM(input_tokens), SUM(output_tokens), SUM(total_cost) FROM costs WHERE project_id = ? GROUP BY model",
|
||
(project_id,)
|
||
)
|
||
|
||
return CostReport(
|
||
by_model=[{r[0]: {"tokens": r[1]+r[2], "cost": r[3]}} for r in rows],
|
||
total=sum(r[3] for r in rows)
|
||
)
|
||
```
|
||
|
||
---
|
||
|
||
## 3. 数据库设计
|
||
|
||
### 3.1 表结构
|
||
|
||
#### tasks 表
|
||
|
||
```sql
|
||
CREATE TABLE tasks (
|
||
id TEXT PRIMARY KEY,
|
||
type TEXT NOT NULL,
|
||
title TEXT NOT NULL,
|
||
description TEXT NOT NULL,
|
||
assigned_to TEXT NOT NULL,
|
||
status TEXT NOT NULL,
|
||
dependencies TEXT, -- JSON array
|
||
input TEXT, -- JSON object
|
||
output TEXT, -- JSON object
|
||
cost_model TEXT,
|
||
cost_input_tokens INTEGER,
|
||
cost_output_tokens INTEGER,
|
||
cost_total REAL,
|
||
created_at TEXT NOT NULL,
|
||
completed_at TEXT,
|
||
error_message TEXT
|
||
);
|
||
|
||
CREATE INDEX idx_tasks_status ON tasks(status);
|
||
CREATE INDEX idx_tasks_assigned ON tasks(assigned_to);
|
||
```
|
||
|
||
#### costs 表
|
||
|
||
```sql
|
||
CREATE TABLE costs (
|
||
id INTEGER PRIMARY KEY AUTOINCREMENT,
|
||
task_id TEXT NOT NULL,
|
||
model TEXT NOT NULL,
|
||
input_tokens INTEGER NOT NULL,
|
||
output_tokens INTEGER NOT NULL,
|
||
total_cost REAL NOT NULL,
|
||
recorded_at TEXT NOT NULL,
|
||
FOREIGN KEY (task_id) REFERENCES tasks(id)
|
||
);
|
||
|
||
CREATE INDEX idx_costs_model ON costs(model);
|
||
CREATE INDEX idx_costs_task ON costs(task_id);
|
||
```
|
||
|
||
#### collaborations 表
|
||
|
||
```sql
|
||
CREATE TABLE collaborations (
|
||
id INTEGER PRIMARY KEY AUTOINCREMENT,
|
||
project_id TEXT NOT NULL,
|
||
task_id TEXT NOT NULL,
|
||
from_agent TEXT NOT NULL,
|
||
to_agent TEXT NOT NULL,
|
||
message TEXT NOT NULL,
|
||
created_at TEXT NOT NULL,
|
||
FOREIGN KEY (task_id) REFERENCES tasks(id)
|
||
);
|
||
```
|
||
|
||
---
|
||
|
||
## 4. 协作模式设计
|
||
|
||
### 4.1 串行审查模式 (Superpowers)
|
||
|
||
```
|
||
┌─────────────┐
|
||
│ Implementer │
|
||
│ (Codex) │
|
||
└──────┬──────┘
|
||
│ task_completed
|
||
▼
|
||
┌─────────────┐
|
||
│Spec Reviewer│
|
||
│ (Gemini) │
|
||
└──────┬──────┘
|
||
│ spec_passed
|
||
▼
|
||
┌──────────────────┐
|
||
│ Code Quality │
|
||
│ Reviewer │
|
||
│ (Gemini) │
|
||
└──────┬───────────┘
|
||
│ quality_passed
|
||
▼
|
||
┌─────────────┐
|
||
│ Next Task │
|
||
└─────────────┘
|
||
```
|
||
|
||
**实现要点**:
|
||
- Implementer 完成任务后报告状态
|
||
- Spec Reviewer 先检查实现与需求一致性
|
||
- 通过后再进行 Code Quality Review
|
||
- 任一环节失败则返回 Implementer
|
||
|
||
### 4.2 对抗辩论模式 (TradingAgents)
|
||
|
||
```
|
||
┌──────────────────┐
|
||
│ 4 Analysts (并行) │
|
||
└────┬────────┬────┘
|
||
│ │
|
||
▼ ▼
|
||
┌─────────┐ ┌─────────┐
|
||
│ Bull │ │ Bear │
|
||
└────┬────┘ └────┬────┘
|
||
│ │
|
||
└─────┬─────┘
|
||
▼
|
||
┌─────────┐
|
||
│Research │
|
||
│ Manager │
|
||
└────┬────┘
|
||
▼
|
||
┌─────────┐
|
||
│ Trader │
|
||
└─────────┘
|
||
```
|
||
|
||
**实现要点**:
|
||
- 多个 Analyst 并行分析
|
||
- Bull 和 Bear 交替辩论（最多 N 轮）
|
||
- Research Manager 裁决并综合
|
||
- 最终由 Trader 做决策
|
||
|
||
### 4.3 三方共识模式 (oh-my-claudecode)
|
||
|
||
```
|
||
┌─────────┐
|
||
│ Planner │
|
||
└────┬────┘
|
||
│ plan_ready
|
||
▼
|
||
┌───────────┐
|
||
│ Architect │
|
||
└────┬──────┘
|
||
│ architecture_approved
|
||
▼
|
||
┌─────────┐
|
||
│ Critic │
|
||
└────┬────┘
|
||
│ consensus_reached
|
||
▼
|
||
┌─────────────┐
|
||
│ Execute │
|
||
└─────────────┘
|
||
```
|
||
|
||
**实现要点**:
|
||
- Planner 先做规划
|
||
- Architect 审查架构
|
||
- Critic 提出批判
|
||
- 三方达成共识后才执行
|
||
|
||
---
|
||
|
||
## 5. 配置设计
|
||
|
||
### 5.1 CLAUDE.md 模板
|
||
|
||
```markdown
|
||
# {项目名称}
|
||
|
||
## 工作模式
|
||
|
||
Superpowers + AI 协作
|
||
|
||
## 角色分工
|
||
|
||
### Claude (我) — 架构师 / 项目经理
|
||
- 需求分析、架构设计、任务拆分
|
||
- 使用 Superpowers 进行规划、审查、调试
|
||
- 代码审核、最终验收、Git 提交管理
|
||
- **绝对不亲自编写代码**
|
||
|
||
### Codex — 后端开发
|
||
- 服务端代码、API、数据库、Migration
|
||
- 单元测试、集成测试
|
||
- 通过 `ask codex "..."` 调用
|
||
|
||
### Gemini — 前端开发
|
||
- 前端组件、页面、样式、交互逻辑
|
||
- 代码审查、安全审计
|
||
- 通过 `ask gemini "..."` 调用
|
||
|
||
## 协作规范
|
||
|
||
1. Claude 不写代码，所有编码任务委派给 Codex/Gemini
|
||
2. 后端任务 → Codex，前端任务 → Gemini
|
||
3. 代码审查 → Gemini
|
||
4. 验收测试 → Claude
|
||
|
||
## Git 规范
|
||
|
||
- 分支命名: `feature/<task-name>`
|
||
- 提交格式: `<type>: <description>`
|
||
- 合并策略: Squash merge
|
||
```
|
||
|
||
### 5.2 配置文件 (config.yaml)
|
||
|
||
```yaml
|
||
# 项目配置
|
||
project:
|
||
name: "sanguo_moziplus_v3"
|
||
description: "多 AI 桥接协作系统"
|
||
|
||
# 桥接器配置
|
||
bridges:
|
||
codex:
|
||
enabled: true
|
||
endpoint: "http://localhost:8080"
|
||
api_key_env: "CODEX_API_KEY"
|
||
model: "codex-4"
|
||
max_tokens: 4000
|
||
|
||
gemini:
|
||
enabled: true
|
||
endpoint: "http://localhost:8081"
|
||
api_key_env: "GEMINI_API_KEY"
|
||
model: "gemini-pro"
|
||
max_tokens: 4000
|
||
|
||
gpt:
|
||
enabled: false
|
||
endpoint: "https://api.openai.com/v1"
|
||
api_key_env: "OPENAI_API_KEY"
|
||
model: "gpt-4"
|
||
max_tokens: 4000
|
||
|
||
# 协作模式配置
|
||
collaboration:
|
||
default_mode: "sequential" # sequential, debate, consensus
|
||
max_retries: 3
|
||
timeout_seconds: 300
|
||
|
||
# 成本配置
|
||
cost:
|
||
tracking_enabled: true
|
||
alert_threshold: 10.0 # 美元
|
||
models:
|
||
claude:
|
||
input_price: 0.003 # per 1k tokens
|
||
output_price: 0.015
|
||
codex:
|
||
input_price: 0.0001
|
||
output_price: 0.0002
|
||
gemini:
|
||
input_price: 0.0001
|
||
output_price: 0.0001
|
||
gpt:
|
||
input_price: 0.03
|
||
output_price: 0.06
|
||
|
||
# 数据库配置
|
||
database:
|
||
path: "~/.moziplus-v3/tasks.db"
|
||
backup_enabled: true
|
||
backup_interval_hours: 24
|
||
```
|
||
|
||
---
|
||
|
||
## 6. API 设计
|
||
|
||
### 6.1 Claude Code 集成 API
|
||
|
||
```python
|
||
class CCBClient:
|
||
"""CCB 客户端，供 Claude Code 调用"""
|
||
|
||
async def ask(self, agent: str, task: str, context: dict = None) -> dict:
|
||
"""
|
||
委派任务给指定 AI
|
||
|
||
Args:
|
||
agent: AI 名称 (codex, gemini, gpt)
|
||
task: 任务描述
|
||
context: 上下文信息
|
||
|
||
Returns:
|
||
{
|
||
"task_id": "...",
|
||
"status": "pending",
|
||
"estimated_cost": 0.01
|
||
}
|
||
"""
|
||
pass
|
||
|
||
async def status(self, task_id: str) -> dict:
|
||
"""查询任务状态"""
|
||
pass
|
||
|
||
async def cost(self, project_id: str = None) -> dict:
|
||
"""获取成本统计"""
|
||
pass
|
||
|
||
async def config(self, key: str = None, value: Any = None) -> dict:
|
||
"""配置管理"""
|
||
pass
|
||
```
|
||
|
||
### 6.2 命令行接口
|
||
|
||
```bash
|
||
# 委派任务
|
||
moziplus ask codex "实现用户认证 API"
|
||
|
||
# 查询状态
|
||
moziplus status <task_id>
|
||
|
||
# 成本报告
|
||
moziplus cost
|
||
|
||
# 配置管理
|
||
moziplus config set collaboration.default_mode debate
|
||
```
|
||
|
||
---
|
||
|
||
## 7. 实现计划
|
||
|
||
### 7.1 Phase 1: 桥接器实现 (Week 1)
|
||
|
||
- [ ] 实现 BridgeAgent 抽象类
|
||
- [ ] 实现 CodexBridge
|
||
- [ ] 实现 GeminiBridge
|
||
- [ ] 实现 GPTBridge
|
||
- [ ] 单元测试
|
||
|
||
### 7.2 Phase 2: 任务编排器 (Week 1-2)
|
||
|
||
- [ ] 实现任务模型
|
||
- [ ] 实现数据库层
|
||
- [ ] 实现任务编排器
|
||
- [ ] 实现成本追踪器
|
||
- [ ] 集成测试
|
||
|
||
### 7.3 Phase 3: Claude Code 集成 (Week 2)
|
||
|
||
- [ ] 实现 CCBClient
|
||
- [ ] 实现命令行接口
|
||
- [ ] 实现 CLAUDE.md 加载
|
||
- [ ] 端到端测试
|
||
|
||
### 7.4 Phase 4: 协作模式 (Week 3)
|
||
|
||
- [ ] 实现串行审查模式
|
||
- [ ] 实现对抗辩论模式
|
||
- [ ] 实现三方共识模式
|
||
- [ ] 模式测试
|
||
|
||
---
|
||
|
||
## 8. 技术栈
|
||
|
||
| 组件 | 技术 | 说明 |
|
||
|------|------|------|
|
||
| **语言** | Python 3.11+ | 主要开发语言 |
|
||
| **数据库** | SQLite | 任务存储 |
|
||
| **异步** | asyncio | 异步任务处理 |
|
||
| **配置** | PyYAML | 配置文件解析 |
|
||
| **CLI** | Click / Typer | 命令行接口 |
|
||
| **测试** | pytest | 单元测试和集成测试 |
|
||
| **API 调用** | httpx | HTTP 客户端 |
|
||
|
||
---
|
||
|
||
## 9. 部署架构
|
||
|
||
```
|
||
┌─────────────────────────────────────────────────┐
|
||
│ 开发环境 │
|
||
│ ~/.openclaw/sanguo_projects/sanguo_moziplus_v3 │
|
||
└─────────────────────────────────────────────────┘
|
||
|
||
┌─────────────────────────────────────────────────┐
|
||
│ 安装环境 │
|
||
│ ~/.sanguo_projects/sanguo_moziplus_v3 │
|
||
└─────────────────────────────────────────────────┘
|
||
|
||
┌─────────────────────────────────────────────────┐
|
||
│ 数据目录 │
|
||
│ ~/.moziplus-v3/ │
|
||
│ - tasks.db │
|
||
│ - config.yaml │
|
||
│ - logs/ │
|
||
└─────────────────────────────────────────────────┘
|
||
```
|
||
|
||
---
|
||
|
||
## 10. 与 v2 的集成
|
||
|
||
### 10.1 兼容性设计
|
||
|
||
- v3 可独立运行
|
||
- 可选择性集成 v2 的黑板架构
|
||
- 共享数据库 schema（部分表）
|
||
|
||
### 10.2 迁移路径
|
||
|
||
```
|
||
v2 用户 → v3 迁移步骤：
|
||
1. 安装 v3
|
||
2. 配置桥接器
|
||
3. 更新 CLAUDE.md
|
||
4. 运行迁移脚本
|
||
5. 验证功能
|
||
```
|
||
|
||
---
|
||
|
||
## 11. 测试策略
|
||
|
||
### 11.1 单元测试
|
||
|
||
- 桥接器测试（Mock API）
|
||
- 任务模型测试
|
||
- 成本计算测试
|
||
|
||
### 11.2 集成测试
|
||
|
||
- 端到端任务流程
|
||
- 多 AI 协作场景
|
||
- 成本追踪验证
|
||
|
||
### 11.3 压力测试
|
||
|
||
- 并发任务处理
|
||
- 大量任务调度
|
||
- 长时间运行稳定性
|
||
|
||
---
|
||
|
||
## 12. 监控和日志
|
||
|
||
### 12.1 日志策略
|
||
|
||
```python
|
||
import logging
|
||
|
||
logger = logging.getLogger("moziplus-v3")
|
||
|
||
# 任务日志
|
||
logger.info(f"Task {task_id} assigned to {agent}")
|
||
|
||
# 成本日志
|
||
logger.info(f"Cost: {model} {tokens} tokens = ${cost}")
|
||
|
||
# 错误日志
|
||
logger.error(f"Task {task_id} failed: {error}")
|
||
```
|
||
|
||
### 12.2 监控指标
|
||
|
||
- 任务成功率
|
||
- 平均任务耗时
|
||
- 成本趋势
|
||
- 桥接器可用性
|
||
|
||
---
|
||
|
||
**文档版本**: v0.1
|
||
**最后更新**: 2026-06-29
|
||
**作者**: Claude Dev
|
||
**审核状态**: 待审核
|