By icefrag
Unlock git worktree-orchestrated development workflows: create isolated branches for parallel TDD tasks, dispatch subagents for execution and reviews, generate Spring Boot integration tests and JaCoCo coverage, produce Mermaid tech designs and MySQL models, evolve coding rules from reviews, and sync remote updates.
npx claudepluginhub icefrag/nbl-superpowers --plugin nbl.superpowersYou MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
Multi-source deep research using firecrawl and exa MCPs. Searches the web, synthesizes findings, and delivers cited reports with source attribution. Use when the user wants thorough research on any topic with evidence and citations.
管理 rules/common/ 目录下规则文件的编辑,确保版本追踪和依赖一致性。 触发条件:用户请求修改规则、修改编码规范。
Use after any code review, error correction, or convention violation to capture the lesson as a persistent rule. Triggers when the user invokes /evolve, or when they express intent like "we should remember this", "make sure this doesn't happen again", "add a rule for this", or "let's note this down". Also use when a recurring mistake pattern emerges across the conversation.
Use when you have a written implementation plan to execute in a separate session with review checkpoints
Use when implementation is complete, all tests pass, and you need to decide how to integrate the work - guides completion of development work by presenting structured options for merge, PR, or cleanup
从 GitHub 仓库安装最新规则文件到本地 ~/.claude/rules/ 目录。 触发条件:用户请求安装规则、从远程更新本地规则。
Use when writing Java Spring Boot integration tests that perform real database CRUD operations with MockMvc and MyBatis-Plus
Use when executing implementation plans with parallelizable tasks - analyzes dependencies, groups by level, executes independent tasks concurrently
Use when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation
Java Web 死代码清理和重构专家。安全识别并移除死代码,每一步都进行测试验证。 触发条件:用户请求清理死代码、重构优化、移除未使用代码。
Use when completing tasks, implementing major features, or before merging to verify work meets requirements
一键安装自定义状态行脚本到 ~/.claude,显示模型、Git状态、上下文使用率、成本和worktree信息
Use when executing implementation plans with independent tasks in the current session
根据需求描述生成技术设计文档。包含需求背景、架构设计、技术设计、API设计、模型设计等章节。 触发条件:用户请求技术方案、技术设计文档、API设计、数据库设计等。
测试覆盖率分析,识别缺口,生成缺失的测试以达到80%+覆盖率。 触发条件:用户请求测试覆盖率分析、提升测试覆盖率、编写测试。
Use when implementing any feature or bugfix, before writing implementation code
Use when starting feature work that needs isolation from current workspace or before executing implementation plans - creates isolated git worktrees with smart directory selection and safety verification
Use when you have a spec or requirements for a multi-step task, before touching code
Use when creating new skills, editing existing skills, or verifying skills work before deployment
基于官方 superpowers 技能体系的扩展,重点增强了多代理并行开发和隔离工作区能力。
同时整合了 everything-claude-code 项目中精选的实用技能,如 /refactor-clean 死代码清理、/test-coverage 测试覆盖率分析
需求澄清(nbl.brainstorming)
→ 输出设计文档
→ 详细计划(nbl.writing-plans)
→ 创建隔离工作区(nbl.using-git-worktrees)
if 任务存在依赖关系:
→ subAgent 顺序执行(nbl.subagent-driven-development)
else:
→ 多 subAgent 并行执行(nbl.parallel-subagent-driven-development)
→ 代码审查(nbl.requesting-code-review)
→ 处理反馈(nbl.receiving-code-review)
→ 人工审核确认
→ 合并到主分支
→ 清理 worktree(nbl.finishing-a-development-branch)
在 Claude Code 中执行以下命令安装此插件:
# 添加插件市场
/plugin marketplace add https://github.com/icefrag/nbl-superpowers
# 安装插件
/plugin install nbl.superpowers@nbl.superpowers
按开发阶段排列:
| Skill | 描述 | 阶段 |
|---|---|---|
| nbl.brainstorming | 需求澄清和设计文档生成 | 📝 需求 |
| nbl.writing-plans | 分解任务生成详细执行计划 | 📋 规划 |
| nbl.using-git-worktrees | 创建 Git worktree 隔离工作区 | ⚙️ 准备 |
| nbl.executing-plans | 主 Agent 直接执行简单任务 | ▶️ 执行 |
| nbl.subagent-driven-development | SubAgent 串行执行任务 | ▶️ 执行 |
| nbl.parallel-subagent-driven-development | SubAgent 并行执行多个任务 | ▶️ 执行 |
| nbl.requesting-code-review | 请求代码审查 | 🔍 审查 |
| nbl.receiving-code-review | 处理代码审查反馈 | 🔍 审查 |
| nbl.finishing-a-development-branch | 合并清理,完成开发分支 | 🎬 收尾 |
这些是可独立使用的工具技能:
| Skill | 描述 | 触发场景 |
|---|---|---|
| nbl.refactor-clean | Java Web 死代码清理和重构专家 | 清理未使用代码、重构优化 |
| nbl.test-coverage | 测试覆盖率分析,生成缺失测试 | 提升测试覆盖率 |
| nbl.tech-design | 根据需求生成技术设计文档 | 技术方案、API 设计、数据库设计 |
| nbl.deep-research | 多源深度网络研究 | 需要调研收集信息 |
| nbl.status-line | 自定义 Claude Code 状态栏 | 安装显示模型 / Git / 上下文 / 成本 / worktree 信息 |
nbl.status-line 是一个自定义状态栏脚本,安装后会在 Claude Code 每次响应前显示:
[Haiku 4.5] 📁 nbl.superpowers | 🌿 main clean
██░░░░░░░░ 15% | $0.12 | ⏱️ 0m 50s
Worktrees:
1. feature-auth → fix-login +2~1?3
2. feature-api → main clean
显示内容:模型名称、项目名、Git 分支及状态(+staged ~modified ?untracked)、上下文使用率进度条、费用累计、会话耗时,以及所有 worktree 列表。
skills/
├── nbl.brainstorming/ # 需求澄清和设计
├── nbl.writing-plans/ # 详细执行计划
├── nbl.using-git-worktrees/ # Git worktree 隔离工作区
├── nbl.executing-plans/ # 主 Agent 直接执行
├── nbl.subagent-driven-development/ # SubAgent 串行执行
├── nbl.parallel-subagent-driven-development/ # SubAgent 并行执行
├── nbl.requesting-code-review/ # 请求代码审查
├── nbl.receiving-code-review/ # 处理代码审查反馈
├── nbl.finishing-a-development-branch/ # 完成开发分支
├── nbl.refactor-clean/ # Java Web 死代码清理
├── nbl.test-coverage/ # 测试覆盖率分析
├── nbl.tech-design/ # 技术设计文档生成
├── nbl.deep-research/ # 多源深度研究
├── nbl.status-line/ # 自定义状态栏
├── nbl.writing-skills/ # 技能开发工具
└── nbl.test-driven-development/ # 测试驱动开发
rules/
└── common/ # 开发规范规则集(示例管理,需手动拷贝到 ~/.claude/rules/ 生效)
| 特性 | 说明 |
|---|---|
| 物理隔离 | Git worktree 级别的隔离,多个任务完全不干扰 |
| 并行开发 | 多 subAgent 同时执行多个独立任务,充分利用 Claude Code 能力 |
| 安全审核 | 代码在 worktree 开发完成,人工审核后才合并到主分支 |
| 多会话支持 | 支持同时打开多个 Claude Code 会话并行处理多个需求 |
| 兼容官方 | 所有技能遵循官方 superpowers 设计原则,学习成本低 |
| 生态整合 | 整合了 everything-claude-code 项目中精选的实用技能,如 nbl.refactor-clean 死代码清理、nbl.tech-design 技术文档生成、nbl.test-coverage 测试覆盖率分析等 |
遵循原项目许可证,扩展部分遵循相同协议。
Complete Superpowers skills library by Jesse Vincent (obra/superpowers), packaged for GitHub Copilot CLI. Includes TDD, systematic debugging, brainstorming, plan writing, subagent-driven development, code review, and more.
Share bugs, ideas, or general feedback.
Superpowers Plus core skills library for Claude Code: planning, execution routing, TDD, debugging, and collaboration workflows
Agentic development framework for Claude Code — disciplined workflow routing, TDD enforcement, safety hooks, systematic debugging, and code review
Access thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.
Design fluency for frontend development. 1 skill with 23 commands (/impeccable polish, /impeccable audit, /impeccable critique, etc.) and curated anti-pattern detection.
Behavioral guidelines to reduce common LLM coding mistakes, derived from Andrej Karpathy's observations on LLM coding pitfalls