Superpowers - Agent Development Workflows
Comprehensive agent development skill collection optimized for AI-assisted coding workflows. Contains 14 skills across 5 categories, prioritized by workflow stage to guide development from ideation to completion.
When to Apply
Reference these skills when:
-
Starting a new feature or project (brainstorming, writing-plans)
-
Debugging any issue (systematic-debugging)
-
Writing or modifying code (test-driven-development)
-
Managing branches and isolation (using-git-worktrees)
-
Completing and integrating work (finishing-a-development-branch)
-
Receiving or requesting code review (receiving-code-review, requesting-code-review)
Skill Categories by Workflow Stage
Stage Category Skills Priority
1 Ideation & Planning brainstorming , writing-plans
CRITICAL
2 Implementation test-driven-development , subagent-driven-development , executing-plans
CRITICAL
3 Debugging systematic-debugging
HIGH
4 Code Review requesting-code-review , receiving-code-review
HIGH
5 Completion verification-before-completion , finishing-a-development-branch
HIGH
6 Utilities using-git-worktrees , dispatching-parallel-agents , using-superpowers , writing-skills
MEDIUM
Quick Reference
- Ideation & Planning (CRITICAL)
-
brainstorming
-
Collaborative design exploration before implementation
-
writing-plans
-
Create detailed implementation plans with bite-sized tasks
- Implementation (CRITICAL)
-
test-driven-development
-
RED-GREEN-REFACTOR cycle for all code
-
subagent-driven-development
-
Execute plans with fresh subagent per task
-
executing-plans
-
Batch execution with checkpoints in separate session
- Debugging (HIGH)
- systematic-debugging
- Four-phase root cause investigation before fixes
- Code Review (HIGH)
-
requesting-code-review
-
Dispatch code reviewer subagent after implementation
-
receiving-code-review
-
Technical evaluation, not performative agreement
- Completion (HIGH)
-
verification-before-completion
-
Evidence before claims, always
-
finishing-a-development-branch
-
Merge, PR, keep, or discard options
- Utilities (MEDIUM)
-
using-git-worktrees
-
Isolated workspaces for parallel development
-
dispatching-parallel-agents
-
One agent per independent problem domain
-
using-superpowers
-
How to find and invoke skills
-
writing-skills
-
TDD applied to process documentation
Core Principles
Each skill enforces key discipline principles:
Principle Skills Iron Law
Test First test-driven-development
No production code without failing test
Root Cause First systematic-debugging
No fixes without investigation
Evidence First verification-before-completion
No claims without verification
Plan First brainstorming , writing-plans
No code without design
How to Use
Read individual skill files for detailed instructions:
brainstorming/SKILL.md test-driven-development/SKILL.md systematic-debugging/SKILL.md
Each skill file contains:
-
Overview and core principle
-
When to use (and when NOT to)
-
Step-by-step process
-
Common mistakes and red flags
-
Integration with other skills
Full Compiled Document
For the complete guide with all skills expanded: AGENTS.md