northseadl/norix-skills / agent-task-orchestration
agent-task-orchestration
VerifiedTask decomposition and multi-agent orchestration with retry, checkpoint recovery, and real-time monitoring. Mixed Codex/Claude Code engine. Parallel/sequential execution.
| File | Kind | Size |
|---|---|---|
| .gitignore | other | 14 B |
| SKILL.md | skill md | 13.6 KB |
| agents/openai.yaml | config | 281 B |
| evals/evals.json | config | 5.8 KB |
| examples/feature-task-spec.md | markdown | 3.4 KB |
| package.json | manifest | 540 B |
| references/codex-patterns.md | markdown | 4.7 KB |
| references/quality-gate.md | markdown | 2.5 KB |
| references/task-decomposition.md | markdown | 3.1 KB |
| references/trinity-protocol.md | markdown | 4.1 KB |
| scripts/dashboard.html | other | 36.5 KB |
| scripts/dispatch.mjs | script | 7.3 KB |
| scripts/lib/cli.mjs | script | 5.7 KB |
| scripts/lib/dag.mjs | script | 4.2 KB |
| scripts/lib/engines.mjs | script | 15.2 KB |
| scripts/lib/logger.mjs | script | 593 B |
| scripts/lib/manage.mjs | script | 5.7 KB |
| scripts/lib/preflight.mjs | script | 1.2 KB |
| scripts/lib/reporter.mjs | script | 5.4 KB |
| scripts/lib/retry.mjs | script | 2.3 KB |
| scripts/lib/server.mjs | script | 5.4 KB |
| scripts/lib/store.mjs | script | 8.8 KB |
examples/feature-task-spec.md
View on GitHub# 示例: iOS 订单管理移植 Task Spec
> 基于当前项目 F03(订单管理)从 Android 移植到 iOS 的真实场景
---
## 拆解结果
### T1 · 订单数据层
- **Agent**: codex-1
- **范围**: `ios/Sources/Feature/Order/Data/`
- **输入**: 无前置依赖(基础设施 I1-1 已完成)
- **交付物**:
- `OrderRepository.swift` — 订单仓库(Apollo query + 缓存策略)
- `OrderModel.swift` — 订单领域模型
- `OrderListQuery.graphql` — 从 `shared/graphql/operations/` 复用
- **验收**:
- [ ] `xcodebuild build` 通过
- [ ] Repository 单元测试覆盖 list/detail 两个路径
- **依赖**: 无
- **预估**: M (30-60min)
- **参考**: Android 实现 `feature/order/src/main/java/com/chuanghuo/operation/feature/order/data/`
---
### T2 · 订单列表 ViewModel
- **Agent**: codex-1
- **范围**: `ios/Sources/Feature/Order/Presentation/`
- **输入**: T1 的 OrderRepository
- **交付物**:
- `OrderListViewModel.swift` — 列表 VM(分页 + 筛选 + 搜索)
- `OrderFilterState.swift` — 筛选状态模型
- **验收**:
- [ ] ViewModel 单元测试: 加载/筛选/搜索/分页 4 个场景
- [ ] `xcodebuild test` 通过
- **依赖**: ← T1
- **预估**: M
---
### T3 · 订单列表 UI
- **Agent**: codex-2
- **范围**: `ios/Sources/Feature/Order/UI/`
- **输入**: T2 的 ViewModel + 设计系统组件
- **交付物**:
- `OrderListScreen.swift` — 列表页(LazyVStack + Pull-to-refresh + 分页加载)
- `OrderCard.swift` — 订单卡片组件
- `OrderFilterSheet.swift` — 筛选面板
- **验收**:
- [ ] Preview 正常渲染(空态/加载态/数据态)
- [ ] 模拟器运行可看到列表数据
- **依赖**: ← T2
- **预估**: M
---
### T4 · 订单详情全链路
- **Agent**: codex-2
- **范围**: `ios/Sources/Feature/Order/UI/` + `ios/Sources/Feature/Order/Presentation/`
- **输入**: T1 的 Repository + 设计系统组件
- **交付物**:
- `OrderDetailViewModel.swift`
- `OrderDetailScreen.swift` — 详情页(商品/物流/操作)
- `OrderStatusBadge.swift` — 状态标签组件
- **验收**:
- [ ] 从列表页点击可跳转详情页
- [ ] 发货/取消操作可触发 mutation
- [ ] `xcodebuild build` 通过
- **依赖**: ← T1 (Repository)
- **预估**: L (60-120min)
---
## 依赖关系
```
T1 (数据层) ──→ T2 (列表 VM) ──→ T3 (列表 UI)
──→ T4 (详情全链路)
无依赖立即启动: T1
T3 和 T4 可并行(仅共享 T1 产出)
```
## 监督级别
| Task | 级别 | 理由 |
|:---|:---|:---|
| T1 | L1 全自动 | 纯新文件创建,有 Android 参照 |
| T2 | L1 全自动 | ViewModel 模式固定 |
| T3 | L2 建议模式 | UI 实现需要审查布局细节 |
| T4 | L2 建议模式 | 包含 mutation 操作需审查 |
## Codex 调度命令
```bash
# 使用 Orchestrator 服务(推荐方式)
# 将上述 Task Spec 保存为 tasks/T1-order-data-layer.md 等文件后:
# 预览执行计划(打开 Dashboard 但不实际调度)
node <SKILLS_DIR>/agent-task-orchestration/scripts/dispatch.mjs ./tasks/ --dry-run
# 并行调度(T2 和 T4 同时执行)
node <SKILLS_DIR>/agent-task-orchestration/scripts/dispatch.mjs ./tasks/ \
--parallel --approval-mode full-auto --cwd .
# Dashboard 会自动打开浏览器,显示:
# - DAG 依赖图: T1 → T2+T4 → T3
# - 每个 Agent 的实时事件流
# - Token 用量和进度
```
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |