Superpowers
Paste the following prompt into your AI chat to install this skill:
Please install @user_14606527/superpowers-skill according to the guide at https://skillhub.cn/install/skillhub.md.
About this skill
Specific Problems Addressed
Software development often struggles with unclear requirements, iterative design changes, and poor code maintainability, leading to low-quality outputs and inefficiencies. Superpowers addresses this by providing a structured, repeatable workflow for AI agents, ensuring each development phase has defined checkpoints and documentation, thereby reducing chaos and enhancing code maintainability and project traceability.
Core Workflow
Superpowers enforces a mandatory 7-phase workflow, with each phase producing specific outputs:
- Brainstorming: Automatically triggers at the start of any task, clarifying real user needs through questioning, exploring alternatives, and generating a design document saved as design.md.
- Design Validation: Ensures the design is complete and feasible by presenting chunks for user review, confirming clarity, and updating the design based on feedback.
- Implementation Planning: Decomposes the design into fine-grained tasks (each 2-5 minutes) with exact file paths, code snippets, and verification steps, following YAGNI and DRY principles, outputting plan.md.
- Test-Driven Development (TDD): Enforces the Red-Green-Refactor cycle during implementation—write failing tests first (Red), then minimal code to pass tests (Green), then refactor to keep code clean.
- Subagent-Driven Development: Uses subagents for parallel or sequential task execution, creating fresh subagents per task with two-stage reviews (spec compliance and code quality), and setting human checkpoints during batch execution.
- Code Review: Reviews code against the plan, reporting issues by severity (Critical, Major, Minor), with critical issues fixed before proceeding, outputting review.md.
- Finishing Development Branch: After all tasks, verifies all tests pass, presents options for merging, creating PRs, keeping branches, or discarding, cleans up the workspace, and generates a final report.
Core Principles and Considerations
The methodology is grounded in principles: Test-Driven Development, Systematic over Ad-Hoc decision-making, Complexity Reduction, and Evidence over Claims. It is suited for projects requiring high code quality and maintainability, but note that it is a mandatory workflow that activates upon detecting development activities. Key considerations include handling subagent deviations from the plan, test failures, or requirement changes by returning to earlier phases. The upfront design phase may add initial time investment, but it supports long-term project health.
Use Cases
- When developing a new feature, use Superpowers for brainstorming to clarify needs through questioning, explore design options, and generate a design document for team reference.
- During code implementation, enforce the test-driven development process: write failing tests first, then minimal code to pass tests, and refactor code to ensure quality.
- For complex projects, use subagent-driven development to assign tasks in parallel to multiple subagents, conduct two-stage code reviews (spec compliance and quality), and set human checkpoints.
- Upon project completion, verify all tests pass in the finishing development branch phase, generate a final report, and decide on code merging, PR creation, or workspace cleanup.
Best For
- Engineers using AI agents for software development who need to ensure code quality and maintainability by following a structured workflow.
- Project managers seeking a systematic approach to manage development processes, producing design documents and plans for progress tracking.
- Test engineers focused on test-driven development practices, requiring automated test integration and quality verification mechanisms.
- Technical architects ensuring consistency between design validation and implementation planning to prevent development from deviating from original requirements.
Related Skills
Supports inline SVG/HTML rendering and Mermaid code output, covering common technical diagram creation like architecture and flow diagrams.
A P8-level programming expert skill that follows a six-step closed-loop workflow, covering the full development lifecycle including code generation, review, and debugging for complex software projects.
Intelligently transforms natural language requests into editable professional online diagrams, supporting various types like flowcharts, architecture diagrams, etc.
A Chinese localization plugin for GitHub Desktop that automatically detects the installation directory and performs text string replacements to localize the interface, supporting loose and asar modes with backup and rollback features.