Is Test-Driven Development (TDD) mandatory?

Yes, the iteration-workflow skill enforces TDD by requiring failing tests to be written and verified before the implementation code is added.

How do I trigger the iteration workflow?

You can activate it using the command `/superpowers-iterate:iterate ` or by simply asking Claude to 'follow the iteration workflow'.

How does it handle state management during long tasks?

The skill maintains an `iteration-state.json` file in the `.agents` directory, tracking the progress of every phase to allow for recovery and persistent context.

Does this skill require external model access?

By default, Full mode uses Codex via MCP for high-reasoning reviews. However, Lite mode (`--lite`) is available to run reviews using Claude subagents exclusively.

What are the 9 phases of the iteration workflow?

The workflow follows a strict sequence: Brainstorm, Plan, Plan Review, Implement, Review, Test, Simplify, Final Review, and Codex Final.

Superpowers Iteration Workflow

Name: Superpowers Iteration Workflow
Author: kenkenmain

bykenkenmain

Productivity & Workflow

Orchestrates a comprehensive 9-phase development lifecycle featuring automated planning, TDD-driven implementation, and multi-model code reviews.

About

This skill provides a rigorous, enterprise-grade orchestration layer for Claude Code, guiding the AI through a structured 9-phase development loop. It manages the entire lifecycle from initial brainstorming and TDD-focused planning to parallel subagent implementation and multi-layered quality assurance using both Claude and Codex models. By enforcing strict exit criteria and state management, it ensures that complex features are built with high architectural integrity, consistent test coverage, and minimal technical debt.

Key Features

Multi-model verification system utilizing Codex for high-reasoning reviews and Claude for rapid iterations.
0 GitHub stars
TDD-first implementation strategy requiring mandatory failing tests before code production.
Native LSP integration to leverage deep code intelligence and real-time diagnostics during the dev cycle.
Parallel subagent dispatching for research, planning, and sequential implementation tasks.
9-Phase structured lifecycle: Brainstorm, Plan, Review, Implement, Test, Simplify, and Final Audit.

Use Cases

Building complex new features that require architectural planning and high test coverage.
Standardizing team development patterns through a repeatable, AI-driven SDLC.
Executing large-scale refactors where code simplification and regression testing are critical.

About

Key Features

Multi-model verification system utilizing Codex for high-reasoning reviews and Claude for rapid iterations.
0 GitHub stars
TDD-first implementation strategy requiring mandatory failing tests before code production.
Native LSP integration to leverage deep code intelligence and real-time diagnostics during the dev cycle.
Parallel subagent dispatching for research, planning, and sequential implementation tasks.
9-Phase structured lifecycle: Brainstorm, Plan, Review, Implement, Test, Simplify, and Final Audit.

Use Cases

Building complex new features that require architectural planning and high test coverage.
Standardizing team development patterns through a repeatable, AI-driven SDLC.
Executing large-scale refactors where code simplification and regression testing are critical.