docs: merge dev tasks and doc requirements into unified agent brief

Single English document containing:
- Full 5-phase task breakdown (60+ tasks)
- Provider interface design and priority matrix
- API endpoint specifications
- Documentation requirements per task
- Testing requirements with examples
- Agent prompt instructions and commit rules
- Build and verification checklist

Signed-off-by: Boos4721 <boos4721@icloud.com>

plan/go-backend-development-tasks.md

@@ -1,232 +0,0 @@
-# Go 后端重构 — 开发任务书
-
-## 概述
-
-将 BoosAPI 后端从 Node.js (Fastify/TypeScript) 重构为 Go。前端 (React SPA) 保持不变。
-
-## 规模
-
-- 后端源码文件：约 350+ .ts 文件（不含测试）
-- API 端点：约 80+
-- Provider 适配器：15+ 个平台
-- Transformer 转换器：4 个协议（OpenAI、Anthropic、Gemini、Canonical）
-
-## 开发原则
-
-1. **API 兼容** — 所有端点 path/request/response 一致，前端零改动
-2. **按协议推进** — 先 OpenAI（最常用）→ Anthropic → Gemini → Canonical
-3. **测试先行** — 每个 handler 有 httptest
-4. **增量替换** — Go binary 与 Node.js 可并存，逐步切流
-5. **DB schema 兼容** — 复用现有 SQLite/MySQL/Postgres 表结构
-
----
-
-## Phase 1：脚手架 ✅ 已完成
-
-| 任务 | 状态 | 文件 |
-|------|------|------|
-| go mod init | ✅ | `go.mod` |
-| 目录结构 | ✅ | `cmd/`, `internal/`, `pkg/` |
-| 配置加载 | ✅ | `internal/config/config.go` — 全部环境变量 |
-| 数据库连接 | ✅ | `internal/db/db.go` — SQLite/MySQL/Postgres |
-| 迁移框架 | ✅ | `internal/db/migrate/000001_initial.up.sql` |
-| Health Check | ✅ | `GET /healthz` — json + DB ping |
-| 中间件 | ✅ | Logger、Auth(CORS)、Panic Recovery |
-| chi 路由分组 | ✅ | `/admin`, `/v1`, `/api` |
-| 优雅关闭 | ✅ | SIGINT/SIGTERM → graceful shutdown |
-| Dockerfile | ✅ | 多阶段构建 ~20MB |
-
----
-
-## Phase 2：Provider 核心代理（当前阶段）
-
-### P2-A：HTTP 工具层
-
-| # | 任务 | 文件 | 依赖 |
-|---|------|------|------|
-| 1 | SSE writer — `text/event-stream` 格式写入 | `pkg/httputil/sse.go` | — |
-| 2 | HTTP 流式转发 — 将上游 stream 中继到客户端 | `pkg/httputil/stream.go` | 1 |
-| 3 | HTTP 反向代理工具 — header 传递、URL 重写 | `pkg/httputil/proxy.go` | — |
-| 4 | 共享 API 类型定义 | `pkg/types/api.go` | — |
-
-### P2-B：Proxy Core 骨架
-
-| # | 任务 | 文件 | 依赖 |
-|---|------|------|------|
-| 5 | Provider 接口定义：`Provider` interface | `internal/proxycore/provider.go` | 4 |
-| 6 | 失败重试/降级策略 | `internal/proxycore/failover.go` | 5 |
-| 7 | 通道选择与负载均衡 | `internal/proxycore/channel.go` | 5 |
-| 8 | 核心编排器 `Orchestrator` | `internal/proxycore/orchestrator.go` | 5-7 |
-| 9 | 请求构建：header 注入、URL 构建 | `internal/proxycore/request.go` | — |
-
-### P2-C：OpenAI Provider
-
-| # | 任务 | 文件 | 依赖 |
-|---|------|------|------|
-| 10 | Chat Completion 非流式 | `internal/provider/openai/chat.go` | 5, 9 |
-| 11 | Chat Completion 流式 (SSE) | `internal/provider/openai/stream.go` | 1, 2, 10 |
-| 12 | Models 列表 | `internal/provider/openai/models.go` | 5 |
-| 13 | Images 生成/编辑 | `internal/provider/openai/images.go` | 5 |
-| 14 | Embeddings | `internal/provider/openai/embeddings.go` | 5 |
-| 15 | Audio (TTS/STT) | `internal/provider/openai/audio.go` | 5 |
-| 16 | Files (上传/列表/内容) | `internal/provider/openai/files.go` | 5 |
-| 17 | Responses API | `internal/provider/openai/responses.go` | 5, 9 |
-
-### P2-D：Anthropic Provider
-
-| # | 任务 | 文件 | 依赖 |
-|---|------|------|------|
-| 18 | Messages 非流式 | `internal/provider/anthropic/messages.go` | 5, 9 |
-| 19 | Messages 流式 | `internal/provider/anthropic/stream.go` | 1, 2, 18 |
-| 20 | 请求转换（OpenAI → Anthropic 格式） | `internal/transformer/anthropic/inbound.go` | — |
-| 21 | 响应转换（Anthropic → OpenAI 格式） | `internal/transformer/anthropic/outbound.go` | — |
-
-### P2-E：Gemini Provider
-
-| # | 任务 | 文件 | 依赖 |
-|---|------|------|------|
-| 22 | Generate Content 非流式 | `internal/provider/gemini/content.go` | 5, 9 |
-| 23 | Generate Content 流式 | `internal/provider/gemini/stream.go` | 1, 2, 22 |
-| 24 | 请求转换（OpenAI → Gemini） | `internal/transformer/gemini/inbound.go` | — |
-| 25 | 响应转换（Gemini → OpenAI） | `internal/transformer/gemini/outbound.go` | — |
-
-### P2-F：Canonical 协议（用于多 provider 统一处理）
-
-| # | 任务 | 文件 | 依赖 |
-|---|------|------|------|
-| 26 | Canonical Request 类型定义 | `internal/transformer/canonical/request.go` | — |
-| 27 | Canonical Response 类型定义 | `internal/transformer/canonical/response.go` | — |
-| 28 | Canonical → Provider 转换 | `internal/transformer/canonical/conversion.go` | 26-27 |
-| 29 | Provider → Canonical 转换 | `internal/transformer/canonical/normalization.go` | 26-27 |
-
-### P2-G：API 路由注册
-
-| # | 任务 | 文件 | 依赖 |
-|---|------|------|------|
-| 30 | `/v1/chat/completions` | `internal/server/handler/chat.go` | 10, 11, 18, 19, 22, 23 |
-| 31 | `/v1/models` | `internal/server/handler/models.go` | 12 |
-| 32 | `/v1/images/*` | `internal/server/handler/images.go` | 13 |
-| 33 | `/v1/embeddings` | `internal/server/handler/embeddings.go` | 14 |
-| 34 | `/v1/audio/*` | `internal/server/handler/audio.go` | 15 |
-| 35 | `/v1/files/*` | `internal/server/handler/files.go` | 16 |
-| 36 | `/v1/responses` | `internal/server/handler/responses.go` | 17 |
-| 37 | 路由统一注册 | `internal/server/router.go` | 30-36 |
-
----
-
-## Phase 3：管理 API
-
-| # | 任务 | 文件 | 依赖 |
-|---|------|------|------|
-| 38 | 管理员认证 (JWT + Bearer) | `internal/auth/jwt.go` | — |
-| 39 | 用户 CRUD | `internal/server/handler/admin/users.go` | 38 |
-| 40 | API Key 管理 | `internal/server/handler/admin/apikeys.go` | 38 |
-| 41 | 站点管理 (上游 Provider 配置) | `internal/server/handler/admin/sites.go` | 38 |
-| 42 | 账单/用量统计 | `internal/server/handler/admin/usage.go` | 38 |
-| 43 | Token 路由配置 | `internal/server/handler/admin/routing.go` | 38 |
-| 44 | 组织/团队管理 | `internal/server/handler/admin/orgs.go` | 38 |
-| 45 | 设置页面 (数据库迁移、备份等) | `internal/server/handler/admin/settings.go` | 38 |
-| 46 | 更新中心 (K3s/Helm) | `internal/server/handler/admin/update.go` | 38 |
-
----
-
-## Phase 4：增值功能
-
-| # | 任务 | 文件 | 依赖 |
-|---|------|------|------|
-| 47 | ComfyUI 反向代理 | `internal/server/handler/comfyui.go` | — |
-| 48 | Video Agent (脚本→视频 pipeline) | `internal/server/handler/video_agent.go` | — |
-| 49 | ComfyUI Agent SSE 对话 | `internal/server/handler/comfyui_agent.go` | 1 |
-| 50 | 文件存储 (proxyFileStore) | `internal/services/filestore.go` | — |
-| 51 | 搜索服务 | `internal/server/handler/search.go` | — |
-
----
-
-## Phase 5：通知 + 运维
-
-| # | 任务 | 文件 | 依赖 |
-|---|------|------|------|
-| 52 | 通知服务 (Telegram) | `internal/notify/telegram.go` | — |
-| 53 | 通知服务 (企业微信/飞书/通用 Webhook) | `internal/notify/webhook.go` | — |
-| 54 | 通知服务 (Bark/Server酱/SMTP) | `internal/notify/others.go` | — |
-| 55 | 通知节流 (冷静期合并) | `internal/notify/throttle.go` | 52-54 |
-| 56 | 桌面版后端适配 | `internal/desktop/adapter.go` | — |
-| 57 | 桌面版 runtime 状态 API | `internal/server/handler/desktop.go` | 56 |
-| 58 | OAuth Provider (Claude/Codex/Gemini) | `internal/oauth/provider.go` | — |
-| 59 | OAuth 会话存储 | `internal/oauth/session.go` | 38 |
-| 60 | 后台定时任务 (签到、模型探测、日志清理) | `internal/cron/scheduler.go` | — |
-
----
-
-## API 兼容性要求
-
-所有 Go handler 必须保持与现有 Node.js 相同的：
-
-### 请求
-- URL 路径
-- HTTP Method
-- Query 参数
-- Request Body JSON 结构
-- Header 名称
-
-### 响应
-- HTTP Status Code
-- Response Body JSON 结构
-- SSE 事件格式 (`data: {...}\n\n`)
-- Error 格式 (`{"error": {...}}`)
-- 速率限制 Header (`X-RateLimit-*`)
-
-### 验证方法
-```
-curl -s http://localhost:4000/v1/models \
-  -H "Authorization: Bearer $PROXY_TOKEN" | jq .
-# 返回格式必须与 Node.js 版一致
-```
-
----
-
-## Provider 优先级矩阵
-
-| Provider | 优先级 | 使用量 | 复杂度 | 阶段 |
-|----------|--------|--------|--------|------|
-| OpenAI | P0 | 🔴🔴🔴 | 中 | Phase 2 |
-| Anthropic | P0 | 🔴🔴🔴 | 高 (转换器) | Phase 2 |
-| Google Gemini | P1 | 🟡🟡 | 高 (转换器) | Phase 2 |
-| OneAPI/NewAPI | P1 | 🟡🟡 | 低 | Phase 3 |
-| Claude (自有平台) | P1 | 🟡🟡 | 中 | Phase 3 |
-| Codex | P2 | 🟢 | 中 | Phase 4 |
-| VoAPI/RIX/Super | P2 | 🟢 | 低 | Phase 4 |
-| Veloera/Sub2/DoneHub | P2 | 🟢 | 低 | Phase 4 |
-
----
-
-## 测试策略
-
-| 层级 | 工具 | 覆盖 |
-|------|------|------|
-| 单元测试 | `testing` | Provider 转换逻辑、工具函数 |
-| Handler 测试 | `httptest` | 每个 API 端点 |
-| 集成测试 | 自建 test suite | 完整请求→响应链路 |
-| Provider 测试 | mock HTTP server | 模拟上游返回，验证转发正确性 |
-| 兼容性测试 | 对照 Node.js 版输出 | 相同请求 → 相同响应 |
-
----
-
-## 构建与发布
-
-```bash
-# 开发构建
-CGO_ENABLED=0 go build -o boosapi ./cmd/boosapi/
-
-# 生产构建 (带版本号)
-CGO_ENABLED=0 go build -ldflags="-s -w -X main.version=$(git describe --tags --always)" \
-  -o boosapi ./cmd/boosapi/
-
-# Docker 构建
-docker build -f docker/Dockerfile -t boosapi:latest .
-
-# 运行
-./boosapi
-# 或
-docker run -p 4000:4000 boosapi:latest
-```

plan/go-backend-doc-requirements.md

@@ -1,219 +0,0 @@
-# Go 后端重构 — 文档需求
-
-## 文档目标
-
-1. **开发者上手** — 5 分钟内能本地跑起来改代码
-2. **架构可理解** — 新团队能看懂模块划分和请求链路
-3. **API 兼容参考** — 前端/客户端开发者知道怎么调用
-4. **运维可操作** — 部署、日志、监控、问题排查
-
----
-
-## 一、开发者文档
-
-### 1.1 README.md 更新
-
-| 需求 | 内容 |
-|------|------|
-| 简介 | BoosAPI 是什么，Go 重构版 |
-| 快速开始 | `go build ./cmd/boosapi` → `./boosapi` |
-| 前置依赖 | Go 1.26+，SQLite 无需安装 |
-| 环境变量 | 简表（AUTH_TOKEN、PROXY_TOKEN、PORT、DB_TYPE、DB_URL） |
-| 示例命令 | 用 `curl` 验证 `/healthz` 和 `/v1/models` |
-
-### 1.2 CONTRIBUTING.md
-
-| 需求 | 内容 |
-|------|------|
-| 分支策略 | `refactor/go-backend` → `main` |
-| 编码规范 | Go 标准 `gofmt`、`go vet`、无 lint 警告 |
-| 提交规范 | Conventional Commits：`feat:` `fix:` `refactor:` `docs:` `test:` |
-| PR 流程 | 每个 Phase 先开 PR review |
-| 测试要求 | 新 handler 必须有 httptest；改逻辑不能降低覆盖率 |
-
-### 1.3 GoDoc 注释
-
-| 需求 | 内容 |
-|------|------|
-| 每个 export 函数 | 一行注释说明作用 |
-| 每个 export 类型 | 说明用途和字段含义 |
-| Package 注释 | `package db` — 简短说明包职责 |
-| Handler 注释 | 说明对应的 API 端点和行为 |
-
-**示例：**
-```go
-// Package provider defines the Provider interface for upstream AI API backends.
-//
-// Each provider (OpenAI, Anthropic, Gemini) implements this interface to
-// support chat completions, streaming, and other capabilities through
-// a unified abstraction.
-package provider
-```
-
----
-
-## 二、架构文档
-
-### 2.1 架构概览
-
-**位置：** `docs/architecture.md`
-
-| 需求 | 内容 |
-|------|------|
-| 整体架构图 | ASCII 或 mermaid 图：前端 → 路由 → 代理核心 → Provider |
-| 请求生命周期 | 从 HTTP 入站到上游响应回来的完整路径 |
-| 模块边界 | `proxycore`、`provider`、`transformer`、`handler` 的职责 |
-| 关键数据流 | SSE 流式请求/响应、WebSocket 传输路径 |
-
-### 2.2 Provider 适配器设计
-
-**位置：** `docs/provider-architecture.md`
-
-| 需求 | 内容 |
-|------|------|
-| Provider 接口定义 | Go interface + 方法签名 |
-| Provider 注册机制 | 如何添加新的 Provider |
-| 转换器模式 | Inbound/Outbound converter 模式说明 |
-| 支持列表 | 当前已支持的 Provider 和对应基准 URL |
-
-### 2.3 数据库 Schema
-
-**位置：** `docs/database.md`
-
-| 需求 | 内容 |
-|------|------|
-| ER 图 | mermaid `erDiagram` |
-| 表结构 | 每张表的字段、类型、索引、外键 |
-| 迁移流程 | 新增字段/表的操作步骤 |
-| 兼容性 | SQLite / MySQL / PostgreSQL 差异 |
-
----
-
-## 三、API 文档
-
-### 3.1 管理 API
-
-**位置：** `docs/management-api.md`（已有，需更新）
-
-| 端点 | 方法 | 说明 |
-|------|------|------|
-| `/admin/users` | GET/POST | 用户列表/创建 |
-| `/admin/users/:id` | GET/PUT/DELETE | 用户详情/更新/删除 |
-| `/admin/api-keys` | GET/POST | API Key 管理 |
-| `/admin/sites` | GET/POST | 上游站点管理 |
-| `/admin/usage` | GET | 用量统计 |
-| `/admin/settings` | GET/PUT | 系统设置 |
-| `/admin/update-center` | GET/POST | K3s 更新中心 |
-| `/admin/backup` | POST | 手动备份 |
-
-### 3.2 代理 API（OpenAI 兼容）
-
-| 端点 | 方法 | 说明 |
-|------|------|------|
-| `/v1/chat/completions` | POST | Chat 补全（流式+非流式） |
-| `/v1/models` | GET | 模型列表 |
-| `/v1/images/generations` | POST | 图片生成 |
-| `/v1/images/edits` | POST | 图片编辑 |
-| `/v1/embeddings` | POST | 向量嵌入 |
-| `/v1/audio/transcriptions` | POST | 语音转文字 |
-| `/v1/audio/translations` | POST | 语音翻译 |
-| `/v1/audio/speech` | POST | 文字转语音 |
-| `/v1/files` | GET/POST | 文件管理 |
-| `/v1/files/:id` | GET/DELETE | 文件详情 |
-| `/v1/files/:id/content` | GET | 文件内容 |
-| `/v1/responses` | POST | Responses API |
-
-### 3.3 自定义 API
-
-| 端点 | 方法 | 说明 |
-|------|------|------|
-| `/api/comfyui/prompt` | POST | ComfyUI 工作流执行 |
-| `/api/comfyui/history` | GET | ComfyUI 历史 |
-| `/api/comfyui/view` | GET | ComfyUI 图片查看 |
-| `/api/comfyui/proxy/ui*` | GET | ComfyUI UI 代理 |
-| `/api/video-agent/*` | POST | 视频生成 Agent |
-| `/api/comfyui-agent/chat` | POST | ComfyUI 对话助手 (SSE) |
-| `/api/search` | POST | 搜索 |
-| `/api/events` | GET | SSE 事件流 |
-| `/api/auth/*` | POST | 登录/认证 |
-
----
-
-## 四、运维文档
-
-### 4.1 部署文档（已有，需更新）
-
-**位置：** `docs/deployment.md`
-
-| 需求 | 说明 |
-|------|------|
-| Docker 镜像 | 从 Go 构建改为多阶段，体积变化 |
-| 环境变量 | 所有 BoosAPI 环境变量（更新 BOOSAPI_*） |
-| Docker Compose | 示例更新为 Go 版本镜像 |
-| 健康检查 | `/healthz` 端点使用说明 |
-| 资源要求 | Go binary 内存/CPU 需求 |
-
-### 4.2 运维手册（已有，需更新）
-
-**位置：** `docs/operations.md`
-
-| 需求 | 说明 |
-|------|------|
-| 日志 | Go slog 输出格式与日志级别 |
-| 监控 | 健康检查、prometheus/metrics |
-| 故障排查 | 常见错误和排查步骤 |
-| 备份恢复 | SQLite/MySQL/Postgres 备份命令 |
-
-### 4.3 CLI 参数
-
-| 参数 | 说明 |
-|------|------|
-| `-h` / `--help` | 帮助 |
-| `-version` | 版本号 |
-| `-config` | 配置文件路径（可选） |
-
----
-
-## 五、GoDoc 发布（可选）
-
-| 需求 | 工具 |
-|------|------|
-| 在线 GoDoc | `pkg.go.dev/github.com/cita-777/boosapi` |
-| 本地生成 | `go doc ./...` |
-
----
-
-## 文档模板
-
-每个 Go package 的 README 模板：
-
-`````markdown
-# package `<name>`
-
-`package <name>` 提供了...
-
-## 用法
-
-```go
-import "github.com/cita-777/boosapi/internal/<path>"
-```
-
-## 类型
-
-### `<TypeName>`
-
-```go
-type TypeName struct {
-    Field1 Type `json:"field1"`
-    Field2 Type `json:"field2"`
-}
-```
-
-## 函数
-
-### `<FuncName>`
-
-```go
-func FuncName(param Type) (ResultType, error)
-```
-`````

plan/go-backend-rewrite.md

@@ -0,0 +1,746 @@
+# BoosAPI Go Backend Rewrite — Agent Development Brief
+
+You are a senior Go developer tasked with rewriting the BoosAPI backend from Node.js/Fastify/TypeScript to Go. The frontend (React SPA) and database schema remain unchanged.
+
+## Repository
+
+- **Branch:** `refactor/go-backend`
+- **Remote:** `gitea` (`https://git.boos.lat/boos/BoosAPI.git`)
+- **Go version:** 1.26
+- **Module:** `github.com/cita-777/boosapi`
+
+## Golden Rules
+
+1. **API compatibility first.** Every endpoint path, HTTP method, request body, response JSON, SSE event format, and error shape must match the existing Node.js implementation exactly. Frontend must work with zero changes.
+2. **Test before claim.** Every handler must have an `httptest` test. Every provider must have a mock HTTP server test. Run `go test ./...` and `go vet ./...` before committing.
+3. **One phase at a time.** Complete all tasks in the current phase before moving to the next.
+4. **Phase 1 is done.** Do not touch Phase 1 files unless fixing a bug.
+5. **Commit convention:** `feat:` for new functionality, `fix:` for bugs, `refactor:` for restructuring, `docs:` for documentation, `test:` for tests. Always sign off: `Signed-off-by: Boos4721 <boos4721@icloud.com>`
+6. **Go standard:** `gofmt`, `go vet`, no lint warnings. Use stdlib where possible.
+
+---
+
+## Architecture
+
+```
+                        ┌──────────────────┐
+                        │   React SPA      │
+                        │   (unchanged)    │
+                        └────────┬─────────┘
+                                 │ HTTP / SSE / WebSocket
+                        ┌────────▼─────────┐
+                        │  Go Backend       │
+                        │  (net/http + chi) │
+                        │                   │
+                        │  ┌───────────┐   │
+                        │  │ chi Router │   │
+                        │  └─────┬─────┘   │
+                        │        │          │
+                        │  ┌─────▼─────┐   │
+                        │  │   Proxy   │   │
+                        │  │   Core    │   │
+                        │  └─────┬─────┘   │
+                        │        │          │
+                        │  ┌─────▼─────┐   │
+                        │  │ Providers │   │
+                        │  │ (interfce)│   │
+                        │  └─────┬─────┘   │
+                        │        │          │
+                        │  ┌─────▼─────┐   │
+                        │  │ DB Layer  │   │
+                        │  │ (sql.DB)  │   │
+                        │  └───────────┘   │
+                        └──────────────────┘
+```
+
+### Directory Layout
+
+```
+cmd/boosapi/main.go           # Entry point (Phase 1 done, do not touch)
+internal/
+  config/config.go            # Env var loader (Phase 1 done)
+  db/db.go                    # DB connection + migration runner (Phase 1 done)
+  db/migrate/                 # SQL migration files (Phase 1 done)
+  server/
+    router.go                 # Route registration (extend as you add handlers)
+    middleware/middleware.go   # Auth, CORS, Logger, Recover (Phase 1 done)
+    handler/                  # HTTP handlers
+      health.go               # GET /healthz (Phase 1 done)
+      chat.go                 # POST /v1/chat/completions
+      models.go               # GET /v1/models
+      images.go               # POST /v1/images/*
+      embeddings.go           # POST /v1/embeddings
+      audio.go                # POST /v1/audio/*
+      files.go                # GET/POST /v1/files/*
+      responses.go            # POST /v1/responses
+      comfyui.go              # ComfyUI reverse proxy
+      video_agent.go          # Video generation agent
+      comfyui_agent.go        # ComfyUI conversational agent (SSE)
+      search.go               # POST /api/search
+      admin/                  # Admin handlers
+        users.go
+        apikeys.go
+        sites.go
+        usage.go
+        routing.go
+        orgs.go
+        settings.go
+        update.go
+  proxycore/
+    provider.go               # Provider interface
+    orchestrator.go           # Core orchestrator (route, failover, fallback)
+    channel.go                # Channel selection + load balancing
+    request.go                # Request building (headers, URL)
+    failover.go               # Retry + degradation strategy
+  provider/
+    openai/
+      chat.go                 # Chat completion (non-streaming)
+      stream.go               # Chat completion (SSE streaming)
+      models.go               # Models list
+      images.go               # Image generation/edits
+      embeddings.go           # Embeddings
+      audio.go                # TTS/STT
+      files.go                # File upload/list/content
+      responses.go            # Responses API
+    anthropic/
+      messages.go             # Messages (non-streaming)
+      stream.go               # Messages (SSE streaming)
+    gemini/
+      content.go              # Generate content (non-streaming)
+      stream.go               # Generate content (SSE streaming)
+  transformer/
+    canonical/
+      request.go              # Canonical request type
+      response.go             # Canonical response type
+      conversion.go           # Canonical ↔ Provider conversion
+    anthropic/
+      inbound.go              # OpenAI → Anthropic request conversion
+      outbound.go             # Anthropic → OpenAI response conversion
+    gemini/
+      inbound.go              # OpenAI → Gemini request conversion
+      outbound.go             # Gemini → OpenAI response conversion
+  auth/
+    jwt.go                    # JWT generation + validation
+    session.go                # Session store
+  notify/
+    telegram.go               # Telegram bot notification
+    webhook.go                # Webhook (WeChat/Feishu/generic)
+    others.go                 # Bark, ServerChan, SMTP
+    throttle.go               # Cooldown merging
+  oauth/
+    provider.go               # OAuth provider interface
+    session.go                # OAuth session storage
+    codex.go                  # Codex OAuth
+    claude.go                 # Claude OAuth
+    gemini.go                 # Gemini OAuth
+  cron/
+    scheduler.go              # Background job scheduler
+  desktop/
+    adapter.go                # Electron desktop backend adapter
+    runtime.go                # Desktop runtime status
+  services/
+    filestore.go              # Proxy file storage
+    search.go                 # Search service
+pkg/
+  httputil/
+    sse.go                    # SSE writer utilities
+    stream.go                 # HTTP streaming relay
+    proxy.go                  # HTTP reverse proxy tools
+  types/
+    api.go                    # Shared API types
+```
+
+---
+
+## Phase 2: Provider Core Proxy
+
+### P2-A: HTTP Utility Layer
+
+#### Task 1: SSE Writer (`pkg/httputil/sse.go`)
+Create a helper for Server-Sent Events over `http.ResponseWriter`.
+
+```go
+package httputil
+
+// SSEWriter wraps http.ResponseWriter for SSE streaming.
+type SSEWriter struct {
+    w    http.ResponseWriter
+    flusher http.Flusher
+    done  <-chan struct{} // for client disconnect detection
+}
+
+func NewSSEWriter(w http.ResponseWriter) (*SSEWriter, error)
+func (s *SSEWriter) WriteEvent(eventType string, data []byte) error
+func (s *SSEWriter) WriteJSON(eventType string, v interface{}) error
+func (s *SSEWriter) Done() <-chan struct{}
+```
+
+- Set `Content-Type: text/event-stream`, `Cache-Control: no-cache`, `Connection: keep-alive`
+- Each event: `data: {json}\n\n`
+- Must detect client disconnection via `r.Context().Done()`
+- Must flush after each event
+
+#### Task 2: HTTP Stream Relay (`pkg/httputil/stream.go`)
+Relay an upstream SSE/streaming response to a downstream `SSEWriter`.
+
+```go
+// RelayStream copies an upstream streaming response to w. It handles
+// chunked transfer from an *http.Response body and forwards data.
+func RelayStream(ctx context.Context, dst *SSEWriter, src *http.Response) error
+```
+
+- Copy bytes from `src.Body` to `dst`
+- Stop when context is cancelled (client disconnect)
+- Close `src.Body` when done
+
+#### Task 3: HTTP Proxy Utilities (`pkg/httputil/proxy.go`)
+```go
+// CopyHeaders copies headers from src to dst.
+func CopyHeaders(dst, src http.Header)
+
+// ForwardedHeaderSet builds the X-Forwarded-* headers.
+func ForwardedHeaderSet(r *http.Request, host string) http.Header
+```
+
+#### Task 4: Shared Types (`pkg/types/api.go`)
+Define shared request/response types used across the application:
+- `ChatRequest`, `ChatMessage`, `ChatChoice`
+- `StreamChoice`, `Delta`
+- `Model`, `ModelList`
+- `ErrorResponse`, `Usage`
+- `ImageRequest`, `ImageResponse`
+- `EmbeddingRequest`, `EmbeddingResponse`
+- `AudioRequest`, `AudioResponse`
+- `FileObject`, `FileList`
+- `APIResponse`, `PaginatedResponse`
+- SSE event types: `TextEvent`, `ImageEvent`, `CharacterPrompts`, `SceneEvent`, `TTSEvent`, `WorkflowEvent`, `DoneEvent`
+
+### P2-B: Proxy Core Skeleton
+
+#### Task 5: Provider Interface (`internal/proxycore/provider.go`)
+Define the core `Provider` interface:
+
+```go
+// Provider defines the interface for upstream AI API backends.
+type Provider interface {
+    // Name returns the provider identifier (e.g., "openai", "anthropic").
+    Name() string
+
+    // ChatCompletion sends a non-streaming chat completion request.
+    ChatCompletion(ctx context.Context, req *ChatRequest) (*ChatResponse, error)
+
+    // ChatCompletionStream sends a streaming chat completion request.
+    // Returns a channel of StreamEvents.
+    ChatCompletionStream(ctx context.Context, req *ChatRequest) (<-chan StreamEvent, error)
+
+    // Models returns the list of available models from this provider.
+    Models(ctx context.Context) ([]Model, error)
+}
+```
+
+Start minimal. Only add methods that are actually implemented. The interface will grow as providers are added.
+
+#### Task 6: Failover Strategy (`internal/proxycore/failover.go`)
+```go
+// FailoverPolicy defines how to handle upstream failures.
+type FailoverPolicy struct {
+    MaxRetries    int
+    RetryDelay    time.Duration
+    FallbackOrder []string // provider names in fallback order
+}
+
+// ShouldRetry determines if a request should be retried based on the error.
+func ShouldRetry(err error, attempt int) bool
+```
+
+#### Task 7: Channel Selection (`internal/proxycore/channel.go`)
+```go
+// ChannelSelector picks the best upstream channel for a given model request.
+type ChannelSelector struct {
+    // ...
+}
+
+func (s *ChannelSelector) Select(ctx context.Context, model string, preferProvider string) (*ProviderChannel, error)
+```
+
+- Weighted random selection based on site priority and weight
+- Skip disabled/cooldown sites
+- Return error if no channel is available
+
+#### Task 8: Core Orchestrator (`internal/proxycore/orchestrator.go`)
+```go
+// Orchestrator manages the full lifecycle of a proxy request.
+type Orchestrator struct {
+    selector   *ChannelSelector
+    failover   *FailoverPolicy
+    registry   *ProviderRegistry
+}
+
+func NewOrchestrator(selector *ChannelSelector, failover *FailoverPolicy, registry *ProviderRegistry) *Orchestrator
+
+// Execute handles a proxy request: select channel, call provider, handle failover.
+func (o *Orchestrator) Execute(ctx context.Context, req *ProxyRequest) (*ProxyResponse, error)
+
+// ExecuteStream handles a streaming proxy request.
+func (o *Orchestrator) ExecuteStream(ctx context.Context, req *ProxyRequest) (<-chan StreamEvent, error)
+```
+
+#### Task 9: Request Builder (`internal/proxycore/request.go`)
+```go
+// BuildUpstreamRequest constructs an HTTP request to the upstream provider.
+// Injects headers (x-boosapi-*, authorization, user-agent, content-type).
+// Handles URL construction, query parameters, body serialization.
+func BuildUpstreamRequest(ctx context.Context, channel *ProviderChannel, body []byte) (*http.Request, error)
+```
+
+- Inject `Content-Type: application/json`
+- Set `Authorization: Bearer <channel API key>`
+- Pass through `x-boosapi-*` headers from downstream
+- Handle streaming vs non-streaming content type
+
+### P2-C: OpenAI Provider
+
+#### Task 10: Chat Completion Non-Streaming (`internal/provider/openai/chat.go`)
+```go
+type Provider struct {
+    baseURL    string
+    httpClient *http.Client
+}
+
+func (p *Provider) ChatCompletion(ctx context.Context, req *ChatRequest) (*ChatResponse, error)
+```
+
+- POST `{baseURL}/v1/chat/completions`
+- Map `ChatRequest` → OpenAI request JSON
+- Parse OpenAI response JSON → `ChatResponse`
+- Handle errors (4xx, 5xx, rate limits)
+
+#### Task 11: Chat Completion Streaming (`internal/provider/openai/stream.go`)
+```go
+func (p *Provider) ChatCompletionStream(ctx context.Context, req *ChatRequest) (<-chan StreamEvent, error)
+```
+
+- POST `{baseURL}/v1/chat/completions` with `stream: true`
+- Parse SSE `data: {...}` lines
+- Each line → `StreamEvent`
+- On `[DONE]` → close channel
+- Use `httputil.SSEWriter` for downstream forwarding
+
+#### Task 12: Models (`internal/provider/openai/models.go`)
+```go
+func (p *Provider) Models(ctx context.Context) ([]Model, error)
+```
+
+- GET `{baseURL}/v1/models`
+- Parse response, return model list
+
+#### Task 13: Images (`internal/provider/openai/images.go`)
+```go
+func (p *Provider) GenerateImage(ctx context.Context, req *ImageRequest) (*ImageResponse, error)
+func (p *Provider) EditImage(ctx context.Context, req *ImageEditRequest) (*ImageResponse, error)
+```
+
+- POST `{baseURL}/v1/images/generations`
+- POST `{baseURL}/v1/images/edits` (multipart form)
+
+#### Task 14: Embeddings (`internal/provider/openai/embeddings.go`)
+```go
+func (p *Provider) CreateEmbedding(ctx context.Context, req *EmbeddingRequest) (*EmbeddingResponse, error)
+```
+
+#### Task 15: Audio (`internal/provider/openai/audio.go`)
+```go
+func (p *Provider) CreateTranscription(ctx context.Context, req *AudioRequest) (*AudioResponse, error)
+func (p *Provider) CreateTranslation(ctx context.Context, req *AudioRequest) (*AudioResponse, error)
+func (p *Provider) CreateSpeech(ctx context.Context, req *TTSRequest) (*http.Response, error)
+```
+
+- Transcriptions/Translations: multipart form POST
+- Speech: JSON POST, return raw audio binary
+
+#### Task 16: Files (`internal/provider/openai/files.go`)
+```go
+func (p *Provider) ListFiles(ctx context.Context) ([]FileObject, error)
+func (p *Provider) UploadFile(ctx context.Context, file *FileUpload) (*FileObject, error)
+func (p *Provider) GetFile(ctx context.Context, fileID string) (*FileObject, error)
+func (p *Provider) GetFileContent(ctx context.Context, fileID string) (io.ReadCloser, error)
+func (p *Provider) DeleteFile(ctx context.Context, fileID string) error
+```
+
+#### Task 17: Responses API (`internal/provider/openai/responses.go`)
+```go
+type ResponsesProvider struct {
+    // wraps Provider for POST /v1/responses
+}
+func (p *ResponsesProvider) CreateResponse(ctx context.Context, req *ResponsesRequest) (*ResponsesResponse, error)
+```
+
+### P2-D: Anthropic Provider
+
+#### Task 18: Messages Non-Streaming (`internal/provider/anthropic/messages.go`)
+Implement Anthropic Messages API format.
+
+```go
+type Provider struct {
+    baseURL    string
+    httpClient *http.Client
+}
+
+func (p *Provider) Messages(ctx context.Context, req *AnthropicMessagesRequest) (*AnthropicMessagesResponse, error)
+```
+
+- POST `{baseURL}/v1/messages`
+- Anthropic-specific request/response types
+- Handle `x-api-key` header auth
+
+#### Task 19: Messages Streaming (`internal/provider/anthropic/stream.go`)
+```go
+func (p *Provider) MessagesStream(ctx context.Context, req *AnthropicMessagesRequest) (<-chan AnthropicStreamEvent, error)
+```
+
+- POST with `stream: true`
+- Parse Anthropic SSE format (`event: message_start`, `event: content_block_delta`, etc.)
+- Convert to unified StreamEvent
+
+### P2-E: Gemini Provider
+
+#### Task 22: Generate Content Non-Streaming (`internal/provider/gemini/content.go`)
+Implement Google Gemini GenerateContent API.
+
+```go
+type Provider struct {
+    baseURL    string
+    httpClient *http.Client
+}
+
+func (p *Provider) GenerateContent(ctx context.Context, req *GeminiRequest) (*GeminiResponse, error)
+```
+
+- POST `{baseURL}/v1beta/models/{model}:generateContent`
+- Gemini-specific request/response types
+- API key as query param
+
+#### Task 23: Generate Content Streaming (`internal/provider/gemini/stream.go`)
+```go
+func (p *Provider) GenerateContentStream(ctx context.Context, req *GeminiRequest) (<-chan GeminiStreamEvent, error)
+```
+
+### P2-F: Canonical Protocol Transformers
+
+#### Task 26: Canonical Types (`internal/transformer/canonical/request.go`, `response.go`)
+Define provider-agnostic request/response types that all providers convert to/from:
+
+```go
+type CanonicalRequest struct {
+    Model    string
+    Messages []CanonicalMessage
+    Stream   bool
+    MaxTokens int
+    Temperature float64
+    // ... other common parameters
+}
+
+type CanonicalResponse struct {
+    ID      string
+    Model   string
+    Choices []CanonicalChoice
+    Usage   *Usage
+}
+```
+
+#### Task 27: Conversions (`internal/transformer/canonical/conversion.go`)
+```go
+// ToOpenAI converts a canonical request to OpenAI format.
+func ToOpenAI(req *CanonicalRequest) (*OpenAIChatRequest, error)
+
+// FromOpenAIResponse converts an OpenAI response to canonical format.
+func FromOpenAIResponse(resp *OpenAIChatResponse) (*CanonicalResponse, error)
+
+// ToAnthropic, FromAnthropicResponse, ToGemini, FromGeminiResponse...
+```
+
+### P2-G: API Route Registration
+
+#### Task 30-37: Handlers + Router
+
+Create handler files under `internal/server/handler/` and register routes in `internal/server/router.go`.
+
+**Chat handler** (`handler/chat.go`):
+```go
+type ChatHandler struct {
+    orchestrator *proxycore.Orchestrator
+}
+
+func (h *ChatHandler) ServeHTTP(w http.ResponseWriter, r *http.Request)
+```
+
+- Parse `ChatRequest` from body
+- If `stream: true` → use SSE streaming path
+- If `stream: false` → use non-streaming path
+- Return OpenAI-compatible response format
+
+**Router** (`server/router.go`):
+```go
+func RegisterRoutes(r chi.Router, deps *Dependencies) {
+    // /v1 routes (proxy token protected)
+    r.Route("/v1", func(r chi.Router) {
+        r.Use(middleware.ProxyToken(deps.Config.ProxyToken))
+        r.Post("/chat/completions", deps.ChatHandler.ServeHTTP)
+        r.Get("/models", deps.ModelsHandler.ServeHTTP)
+        r.Post("/images/generations", deps.ImagesHandler.Generate)
+        // ...
+    })
+
+    // /api routes
+    r.Route("/api", func(r chi.Router) {
+        r.Post("/comfyui-agent/chat", deps.ComfyUIAgentHandler.ServeHTTP)
+        // ...
+    })
+
+    // /admin routes (auth token protected)
+    r.Route("/admin", func(r chi.Router) {
+        r.Use(middleware.AuthToken(deps.Config.AuthToken))
+        // ...
+    })
+}
+```
+
+---
+
+## Phase 3: Management API
+
+#### Task 38: JWT Auth (`internal/auth/jwt.go`)
+Implement JWT token generation and validation for admin sessions.
+
+```go
+func GenerateToken(secret string, userID string, role string) (string, error)
+func ValidateToken(secret string, tokenString string) (*Claims, error)
+```
+
+#### Task 39-46: Admin Handlers
+
+Each handler follows this pattern:
+```go
+type UsersHandler struct {
+    db *sql.DB
+}
+
+func (h *UsersHandler) List(w http.ResponseWriter, r *http.Request)  // GET /admin/users
+func (h *UsersHandler) Get(w http.ResponseWriter, r *http.Request)   // GET /admin/users/{id}
+func (h *UsersHandler) Create(w http.ResponseWriter, r *http.Request) // POST /admin/users
+func (h *UsersHandler) Update(w http.ResponseWriter, r *http.Request) // PUT /admin/users/{id}
+func (h *UsersHandler) Delete(w http.ResponseWriter, r *http.Request) // DELETE /admin/users/{id}
+```
+
+Required admin endpoints:
+| Route | Handler | SQL Query |
+|-------|---------|-----------|
+| `GET /admin/users` | `UsersHandler.List` | `SELECT * FROM users ORDER BY created_at DESC` |
+| `POST /admin/users` | `UsersHandler.Create` | `INSERT INTO users ...` |
+| `GET /admin/users/{id}` | `UsersHandler.Get` | `SELECT * FROM users WHERE id = ?` |
+| `PUT /admin/users/{id}` | `UsersHandler.Update` | `UPDATE users SET ... WHERE id = ?` |
+| `DELETE /admin/users/{id}` | `UsersHandler.Delete` | `DELETE FROM users WHERE id = ?` |
+| `GET /admin/api-keys` | `APIKeysHandler.List` | `SELECT * FROM api_keys` |
+| `POST /admin/api-keys` | `APIKeysHandler.Create` | `INSERT INTO api_keys ...` |
+| `GET /admin/sites` | `SitesHandler.List` | `SELECT * FROM sites ORDER BY priority DESC` |
+| `POST /admin/sites` | `SitesHandler.Create` | `INSERT INTO sites ...` |
+| `PUT /admin/sites/{id}` | `SitesHandler.Update` | `UPDATE sites SET ... WHERE id = ?` |
+| `DELETE /admin/sites/{id}` | `SitesHandler.Delete` | `DELETE FROM sites WHERE id = ?` |
+| `GET /admin/usage` | `UsageHandler.Summary` | `SELECT ... FROM usage_logs GROUP BY ...` |
+| `GET /admin/settings` | `SettingsHandler.Get` | Application settings |
+| `PUT /admin/settings` | `SettingsHandler.Update` | Application settings |
+
+---
+
+## Phase 4: Value-Added Features
+
+#### Task 47: ComfyUI Proxy (`internal/server/handler/comfyui.go`)
+Reverse proxy for ComfyUI (port 8188). Key endpoints:
+- `POST /api/comfyui/prompt` → `POST /prompt`
+- `GET /api/comfyui/history` → `GET /history`
+- `GET /api/comfyui/view` → `GET /view`
+- `GET /api/comfyui/proxy/ui*` → `GET /*` (with HTML rewriting)
+
+#### Task 48: Video Agent (`internal/server/handler/video_agent.go`)
+One-shot script → video pipeline. Uses the gateway's own `/v1/chat/completions` and `/v1/images/generations` and `/v1/audio/speech`.
+
+Endpoints:
+- `POST /api/video-agent/analyze`
+- `POST /api/video-agent/character-prompts`
+- `POST /api/video-agent/generate-character`
+- `POST /api/video-agent/scene-prompt`
+- `POST /api/video-agent/generate-tts`
+- `POST /api/video-agent/full-workflow`
+
+#### Task 49: ComfyUI Agent SSE Chat (`internal/server/handler/comfyui_agent.go`)
+Conversational agent using SSE streaming. The LLM generates structured SSE events directly.
+
+- `POST /api/comfyui-agent/chat` — receives user message, streams agent response
+- System prompt instructs LLM to emit SSE-formatted output
+- Event types: `text`, `character_prompts`, `image`, `scene`, `tts`, `comfyui_workflow`, `done`
+
+---
+
+## Phase 5: Notifications + Operations
+
+#### Task 52-55: Notification Service (`internal/notify/`)
+Multi-channel notification dispatch with cooldown merging.
+
+| Channel | Config key | Delivery |
+|---------|-----------|----------|
+| Telegram | `telegramEnabled`, `telegramBotToken`, `telegramChatId` | Bot API |
+| WeChat Work | `wecom*` | Webhook |
+| Feishu | `feishu*` | Webhook |
+| Generic Webhook | `webhookUrl` | JSON POST |
+| Bark | `bark*` | HTTP (iOS) |
+| ServerChan | `serverchan*` | HTTP |
+| SMTP | `smtp*` | Email |
+
+#### Task 58-59: OAuth (`internal/oauth/`)
+Provider OAuth flows for Claude, Codex, Gemini:
+
+```go
+// OAuthProvider handles the OAuth flow for an AI platform.
+type OAuthProvider interface {
+    AuthURL(state string) string
+    ExchangeCode(ctx context.Context, code string) (*OAuthToken, error)
+    RefreshToken(ctx context.Context, refreshToken string) (*OAuthToken, error)
+}
+```
+
+---
+
+## Documentation Requirements
+
+Each task must include appropriate documentation:
+
+### Code-Level Documentation
+- Every exported Go function must have a doc comment
+- Every exported type must have a doc comment
+- Every package must have a `package` doc comment
+
+Example:
+```go
+// Package provider defines the Provider interface for upstream AI API backends.
+//
+// Each provider (OpenAI, Anthropic, Gemini) implements this interface to
+// support chat completions, streaming, and other capabilities through
+// a unified abstraction.
+package provider
+```
+
+### API Documentation
+When adding a new handler, document the endpoint in the handler file:
+
+```go
+// POST /v1/chat/completions
+//
+// Creates a model response for the given chat conversation.
+// Supports both streaming (SSE) and non-streaming responses.
+//
+// Request: ChatRequest
+// Response (non-streaming): ChatResponse
+// Response (streaming): SSE events of StreamChoice
+//
+// Errors:
+//   400 - Invalid request body
+//   401 - Missing or invalid API key
+//   429 - Rate limited
+//   500 - Upstream error
+```
+
+### User-Facing Documentation
+Update the relevant doc files under `docs/` when adding significant functionality:
+- `docs/deployment.md` — Environment variables, Docker setup
+- `docs/operations.md` — Log format, health checks, troubleshooting
+- `docs/management-api.md` — Admin API reference
+- `docs/architecture.md` — Architecture overview (create if missing)
+
+---
+
+## Testing Requirements
+
+Every handler must have a test:
+
+```go
+func TestChatHandler_NonStreaming(t *testing.T) {
+    // Setup mock server for upstream
+    upstream := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+        w.WriteHeader(http.StatusOK)
+        json.NewEncoder(w).Encode(openai.ChatResponse{...})
+    }))
+    defer upstream.Close()
+
+    // Setup handler with mock upstream URL
+    handler := NewChatHandler(&Config{UpstreamBaseURL: upstream.URL})
+
+    // Create request
+    body := `{"model":"gpt-4","messages":[{"role":"user","content":"hi"}]}`
+    req := httptest.NewRequest("POST", "/v1/chat/completions", strings.NewReader(body))
+    req.Header.Set("Authorization", "Bearer test-token")
+    rec := httptest.NewRecorder()
+
+    // Serve
+    handler.ServeHTTP(rec, req)
+
+    // Assert
+    assert.Equal(t, http.StatusOK, rec.Code)
+    var resp ChatResponse
+    json.Unmarshal(rec.Body.Bytes(), &resp)
+    assert.Equal(t, "gpt-4", resp.Model)
+}
+```
+
+### Test Coverage Targets
+- Provider packages: 80%+ (mock HTTP server)
+- Handler packages: 80%+ (httptest)
+- Transformer packages: 90%+ (pure function tests)
+- Utility packages: 90%+
+- Overall project: 70%+
+
+---
+
+## Provider Implementation Priority
+
+| Provider | Priority | Complexity | When |
+|----------|----------|-----------|------|
+| OpenAI (Chat, Models) | P0 | Medium | Phase 2-A |
+| OpenAI (Images, Embeddings, Audio, Files) | P0 | Low-Medium | Phase 2-A |
+| Anthropic Messages | P0 | High (different auth, streaming format) | Phase 2-B |
+| Google Gemini | P1 | High (different API structure) | Phase 2-C |
+| OneAPI / NewAPI | P1 | Low (OpenAI-compatible) | Phase 3 |
+| Claude Platform | P1 | Medium | Phase 3 |
+| Codex | P2 | Medium | Phase 4 |
+| VoAPI / RIX / Super | P2 | Low | Phase 4 |
+| Veloera / Sub2 / DoneHub | P2 | Low | Phase 4 |
+
+---
+
+## Build & Run
+
+```bash
+# Build
+CGO_ENABLED=0 go build -o boosapi ./cmd/boosapi/
+
+# Run
+AUTH_TOKEN=admin PROXY_TOKEN=proxy ./boosapi
+
+# Test
+go test ./... -v -count=1
+
+# Vet
+go vet ./...
+
+# Docker
+docker build -f docker/Dockerfile -t boosapi:latest .
+docker run -p 4000:4000 -e AUTH_TOKEN=admin -e PROXY_TOKEN=proxy boosapi:latest
+```
+
+## Before Committing
+
+1. `go vet ./...` — zero warnings
+2. `go test ./...` — all tests pass
+3. `gofmt -s -w .` — code is formatted
+4. `git diff --cached` — review for secrets (no real IPs, tokens, passwords)
+5. Commit message follows convention: `feat:`, `fix:`, `refactor:`, `docs:`, `test:`
+6. Sign off: `Signed-off-by: Boos4721 <boos4721@icloud.com>`