Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
Replace author name with public handle (LeeOde)
What This Skill Does
A two-mode framework for writing and refactoring code, alternating between feeling-first creative exploration and structure-first engineering. Rooted in three principles: fallibilism, relational ontology, and play. Developed by reviewing 1800 lines of game code built across 10 iterative sessions in a single creative run.
Naming the two modes as distinct phases gives a concrete switching criterion that stops premature abstraction during discovery and avoids indefinite deferral of cleanup in production code.
When to use it
- Deciding when to stop prototyping and start refactoring
- Auditing a creative codebase for dead code and repeated patterns
- Organizing game jam output into maintainable structure after the jam
- Identifying hot-path performance problems after a first working build
- Naming helper functions extracted from repeated logic blocks
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: I just finished a 1500-line prototype. How do I refactor it without breaking what works?
- 1Read the entire codebase before changing anything, noting dead code, duplicates, and hot paths
- 2Move all data and configuration to the top of each file, all logic below it
- 3Extract any pattern that appears 3 or more times into a named helper function, naming it by what the pattern does
- 4Pre-index any object searches (find, filter) running inside loops using a Map
- 5Remove unused variables and commented-out code only after confirming git history preserves them
A prioritized refactoring checklist organized by risk and impact, starting with low-risk data/logic separation and ending with targeted performance changes on the hot path