P.A.C.E. v2.0 complexity routing engine. Analyzes task complexity and routes to optimal path (A/B/C/D) with effort level mapping. v8.0: Added Path D for Agent Teams, effort parameter integration.
Scanned 9/5/2026
Install to Claude Code
npx -y skills add WenJunDuan/Rlues --skill phase-router --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Phase Router?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/wenjunduan-phase-router-88c82048)More formats (shields.io, HTML) on the badges page.
---
name: phase-router
description: |
P.A.C.E. v2.0 complexity routing engine. Analyzes task complexity
and routes to optimal path (A/B/C/D) with effort level mapping.
v8.0: Added Path D for Agent Teams, effort parameter integration.
---
# Phase Router (P.A.C.E. v2.0)
## 路由决策树
```
任务输入
│
├─ 单文件 AND <30行变更?
│ → Path A (effort=low)
│
├─ 2-10文件 OR 30-500行?
│ → Path B (effort=medium)
│
├─ >10文件 OR >500行 OR 跨模块?
│ → Path C (effort=high)
│
└─ 架构级 AND 可并行拆分 AND (>3独立模块 OR 用户--team)?
→ Path D (effort=max, Agent Teams)
```
## Path A - Quick Fix
```yaml
effort: low
workflow: R1 → E → R2
cunzhi: 仅结束时
duration: 30-60 分钟
todo: 单条任务,直接写入 doing.md
```
## Path B - Planned Development
```yaml
effort: medium
workflow: R1 → I → P → E → R2
cunzhi: plan 确认 + 结束确认
duration: 2-8 小时
todo: 分解为 2-5 个子任务
model_route: 按任务类型选择 (见 model-router skill)
```
## Path C - System Development
```yaml
effort: high (关键阶段 max)
workflow: 完整九步
cunzhi: 每阶段关键节点
duration: 数天
todo: 按九步阶段分解,每步独立任务
model_route: Claude Opus 优先 (推理深度)
```
## Path D - Agent Teams Orchestration
```yaml
effort: max
workflow: Lead 分析 → 拆分子任务 → 并行执行 → 合并 → 验证
cunzhi: Lead 分配前确认 + 合并后确认
duration: 数小时-数天 (但并行加速)
requires: CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1
todo: 按 teammate 分组
```
### Path D 触发条件 (满足 ≥2 项)
- 涉及 >3 个独立模块/目录
- 前端 + 后端 + 测试可并行
- 预估 >4 小时顺序执行
- 存在独立的研究/调研子任务
- 用户显式 `--team`
### Path D 回退
Agent Teams 不稳定或失败 → 降级到 Path C 顺序执行。
## effort 与 Adaptive Thinking 映射
| P.A.C.E. | effort | 模型行为 |
|:---|:---|:---|
| Path A | low | 可能跳过思考,极速响应 |
| Path B | medium | 适度思考 |
| Path C | high | 几乎总是深度思考 (默认) |
| Path D | max | 最高推理深度,峰值能力 |
## 评估指标
```yaml
file_count: 涉及文件数
code_lines: 预估变更行数
architecture_impact: none / minor / major
module_count: 涉及独立模块数
parallelizable: true / false
```
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!