- Home
- Skills
- Solatis
- Claude Config
- Refactor
refactor_skill
- Python
538
GitHub Stars
3
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill solatis/claude-config --skill refactor- CLAUDE.md784 B
- README.md6.8 KB
- SKILL.md1.6 KB
Overview
This skill runs an automated refactoring workflow by immediately invoking the orchestration script when a user requests refactoring analysis, technical debt review, or code quality improvement. It is designed to run the scripted multi-step process without preliminary exploration and produce prioritized, actionable work items for developers. Use the script flags to control progress and scope.
How this skill works
On activation the skill immediately launches the refactor script located in the skills scripts directory and follows its step-driven workflow. The script runs five phases: dispatch parallel explorers, triage and ID smells, cluster by root cause, contextualize with user intent and priority, then synthesize actionable tasks. Control steps and category count with --step, --total-steps, and --n; do not perform any prior analysis outside the script.
When to use it
- User requests a refactoring analysis, tech-debt triage, or code quality review
- You need a reproducible, stepwise refactor workflow that yields prioritized tasks
- You want parallel exploration across multiple issue categories
- You require consistent outputs for tracking and handoff
- You need to scale exploration scope for a file, module, or whole repo
Best practices
- Invoke the script immediately on request; do not explore manually beforehand
- Set --total-steps=5 for the full workflow and advance --step according to orchestration
- Choose --n based on scope: 5 for small, 10 for medium, 25 for large
- Keep working-dir set to the skills scripts directory so the script finds its conventions and category list
- Capture and persist the script output to feed into planning or ticketing systems
Example use cases
- Quick code smell check for a single file: run with --n 5 and step through for a fast triage
- Module-level quality review: run with --n 10 and complete all 5 steps to get prioritized refactor tasks
- Comprehensive codebase debt assessment: run with --n 25 and synthesize a roadmap of work items
- Automated handoff: run the full workflow and export actionable tickets for engineering teams
FAQ
Run the orchestrator immediately from the skill scripts directory, e.g. python3 -m skills.refactor.refactor --step 1 --total-steps 5 --n 10.
Can I explore the code before running the script?
No. The script is the controlled workflow; do not explore first. Invoke the script immediately and follow its outputs and prompts.