Agent-Skills.md

Agent Skills: Refactor

Invoke IMMEDIATELY via python script when user requests refactoring analysis, technical debt review, or code quality improvement. Do NOT explore first - the script orchestrates exploration.

UncategorizedID: solatis/claude-config/refactor

Repository

solatis/claude-config
solatisLicense: MIT
44387

Skill Files

Browse the full folder contents for refactor.

Download Skill

Loading file tree…

skills/refactor/SKILL.md

Skill Metadata

Name
refactor
Description
Invoke IMMEDIATELY via python script when user requests refactoring analysis, technical debt review, or code quality improvement. Do NOT explore first - the script orchestrates exploration.

Refactor

When this skill activates, IMMEDIATELY invoke the script. The script IS the workflow.

Invocation

<invoke working-dir=".claude/skills/scripts" cmd="python3 -m skills.refactor.refactor --step 1 --total-steps 5 --n 10" />

| Argument | Required | Description | | --------------- | -------- | --------------------------------------------- | | --step | Yes | Current step (starts at 1) | | --total-steps | Yes | Total steps (5 for full workflow) | | --n | No | Number of categories to explore (default: 10) |

Do NOT explore or analyze first. Run the script and follow its output.

Workflow Phases

  1. Dispatch - Launch parallel Explore agents (one per category)
  2. Triage - Structure smell findings with IDs
  3. Cluster - Group smells by shared root cause
  4. Contextualize - Extract user intent, prioritize issues
  5. Synthesize - Generate actionable work items

Determining N (category count)

Default: N = 10

Adjust based on user request scope:

  • SMALL (single file, specific concern, "quick look"): N = 5
  • MEDIUM (directory, module, standard analysis): N = 10
  • LARGE (entire codebase, "thorough", "comprehensive"): N = 25

The script randomly selects N categories from the 38 available code quality categories defined in conventions/code-quality/.