superpowers-subagent-dev
Use when executing implementation plans with independent tasks - coordinates task execution by dispatching subagents per task with verification checkpoints, adapted for OpenClaw's isolated session model
Best use case
superpowers-subagent-dev is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Use when executing implementation plans with independent tasks - coordinates task execution by dispatching subagents per task with verification checkpoints, adapted for OpenClaw's isolated session model
Teams using superpowers-subagent-dev should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/superpowers-subagent-dev/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How superpowers-subagent-dev Compares
| Feature / Agent | superpowers-subagent-dev | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Use when executing implementation plans with independent tasks - coordinates task execution by dispatching subagents per task with verification checkpoints, adapted for OpenClaw's isolated session model
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
Related Guides
AI Agents for Coding
Browse AI agent skills for coding, debugging, testing, refactoring, code review, and developer workflows across Claude, Cursor, and Codex.
Best AI Skills for Claude
Explore the best AI skills for Claude and Claude Code across coding, research, workflow automation, documentation, and agent operations.
Cursor vs Codex for AI Workflows
Compare Cursor and Codex for AI coding workflows, repository assistance, debugging, refactoring, and reusable developer skills.
SKILL.md Source
# Superpowers Subagent 开发(OpenClaw 适配版)
## 概述
通过为每个任务 dispatch 独立 subagent 来执行计划,每个任务后进行两阶段审查:先审查规格合规,再审查代码质量。
**为什么用 subagent:** 将任务委托给专门 agent,隔离上下文。通过精确构造指令和上下文,确保它们专注并成功完成。Subagent 不应继承 session 历史——精确构造它们需要的上下文。
**核心原则:** 每个任务新鲜 subagent + 两阶段审查(规格合规 → 代码质量)
## OpenClaw 适配说明
Superpowers 原版设计基于 Claude Code 的 Task/subagent 级联模型。OpenClaw 使用 `sessions_spawn` 创建独立 session,无法做到原生级联。
**OpenClaw 适配方案:**
- 主 agent 作为 controller,协调所有工作
- 用 `sessions_spawn(mode="run")` dispatch 一次性任务 subagent
- Subagent 结果通过 session 历史或文件系统传递
- 审查在主 session 内 inline 执行(或 spawn 独立审查 session)
- 复杂审查任务用 `sessions_spawn`
## 使用条件
```
有实现计划?
→ 任务大多独立?
→ 在本 session 内工作?
→ 用 subagent-dev(这个技能)
→ 并行独立 session?
→ 用 sessions_spawn 并行 dispatch
→ 手动执行或先 brainstorming
```
## 流程
### 每次任务
1. **读取计划,提取所有任务**
- 读取计划文件一次
- 提取所有任务及其完整文本和上下文
- 创建任务列表
2. **Dispatch 实现者 subagent**
```
使用 sessions_spawn:
- mode: "run"(一次性任务)
- task: 完整任务文本 + 上下文
- cwd: 项目目录
```
3. **Subagent 提问处理**
- 如果 subagent 提问 → 回答问题,提供上下文
- 重新 dispatch 或继续
4. **Subagent 实现、测试、commit、自审**
- 实现者完成工作
- 运行测试
- Commit
- 自审
5. **规格合规审查(主 session inline 或 spawn)**
- 检查实现是否满足计划中的规格
- 发现问题 → 反馈给实现者修复 → 重新审查
- 通过 → 继续
6. **代码质量审查**
- 检查代码质量:DRY、命名、测试设计
- 发现问题 → 修复
- 通过 → 继续
7. **标记任务完成,继续下一个**
### 任务后
所有任务完成后:
- 做最终代码审查
- 调用 `superpowers-finishing-branch` 完成工作
## 模型选择
用能处理任务的最弱模型,节省成本增加速度:
| 任务类型 | 示例 | 模型 |
|---------|------|------|
| 机械实现任务 | 孤立函数、清晰规格、1-2 文件 | 快速便宜模型 |
| 集成和判断任务 | 多文件协调、模式匹配、调试 | 标准模型 |
| 架构设计和审查任务 | 需要设计判断或广泛代码库理解 | 最强模型 |
## 处理 Subagent 状态
Subagent 报告四种状态之一。适当处理:
**DONE:** 继续规格合规审查。
**DONE_WITH_CONCERNS:** 实现者完成但标记了疑虑。先读关注点再继续。如果关注点涉及正确性或范围,先解决再审查。
**NEEDS_CONTEXT:** 实现者需要未提供的信息。提供缺失上下文并重新 dispatch。
**BLOCKED:** 实现者无法完成任务。评估阻塞:
1. 如果是上下文问题,提供更多上下文并用相同模型重新 dispatch
2. 如果任务需要更多推理,用更强模型重新 dispatch
3. 如果任务太大,拆成更小的块
4. 如果计划本身有问题,上报给主人
**永远不要**忽略升级或强制相同模型重试而不做变更。
## Red Flags
**永远不要:**
- 在 main/master 分支上开始实现(没有主人明确同意)
- 跳过审查(规格合规或代码质量)
- 有未修复问题时继续
- 同时 dispatch 多个实现 subagent(冲突)
- 让 subagent 读计划文件(直接提供完整文本)
- 跳过场景设置上下文(subagent 需要理解任务在哪里_fit)
- 忽略 subagent 的提问(先回答再让他们继续)
- 接受"差不多"的规格合规(审查发现问题 = 没完成)
- 跳过审查循环(审查发现问题 = 实现者修复 = 重新审查)
- 让实现者自审替代实际审查(两者都需要)
- **在规格合规通过前开始代码质量审查**(顺序错误)
- 有公开问题时继续下一个任务
## 集成
**必需的工作流技能:**
- `superpowers-writing-plans` — 创建这个技能执行的计划
- `superpowers-finishing-branch` — 所有任务完成后的收尾
- `superpowers-tdd` — Subagent 每个任务遵循 TDD
**替代工作流:**
- `sessions_spawn` 并行 dispatch — 用于独立问题的并行调查
- 顺序执行 — 在主 session 内按批次执行Related Skills
superpowers-executing-plans
Use when executing a written implementation plan in the current session with sequential task execution and review checkpoints - for when subagent-driven mode is not available
superpowers-writing-plans
Use when you have a spec or requirements for a multi-step task, before touching code - guides writing comprehensive implementation plans with bite-sized tasks, TDD, and DRY/YAGNI principles
superpowers-verification
Use when about to claim any work is complete, fixed, passing, or successful - requires running fresh verification commands and reading actual output before making any success claims; evidence before assertions always
superpowers-tdd
Use when implementing any feature or bugfix, before writing implementation code - enforces RED-GREEN-REFACTOR cycle: write failing test first, verify it fails, write minimal code, verify it passes, then refactor
superpowers-systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior - enforces systematic four-phase debugging: root cause investigation, pattern analysis, hypothesis testing, and evidence-based fix verification
superpowers-parallel-agents
Use when facing 2 or more independent tasks that can be worked on without shared state - dispatches parallel subagents using sessions_spawn for concurrent investigation and execution, adapted for OpenClaw
superpowers-overview
Use when starting any development work or when unsure which superpowers development skill to use - provides entry point and navigation to the full superpowers skill suite for OpenClaw agents
superpowers-isolated-workspace
Use when starting feature work that needs isolation from current workspace - creates isolated git branches with clean setup and safety verification, adapted for OpenClaw environments
superpowers-finishing-branch
Use when implementation is complete, all tests pass, and you need to decide how to integrate the work - presents structured options for merge, PR, or cleanup; adapted for OpenClaw git workflow without worktrees
superpowers-brainstorming
Use before any creative work - creating features, building components, adding functionality, or modifying behavior - guides through exploration, questioning, design proposal, and spec documentation before any implementation
superpowers-requesting-code-review
Use when completing tasks, implementing major features, or before merging - dispatches code review subagent to catch issues before they cascade, adapted for OpenClaw sessions_spawn model
superpowers-receiving-code-review
Use when receiving code review feedback - requires technical verification before implementing suggestions, with reasoned pushback when feedback is technically questionable; no performative agreement