normalize-ai-config

name	normalize-ai-config
description	Normalize agents/skills/hooks so responsibilities are separated and workflows are coherent.
disable-model-invocation	true

Normalize AI Config

Audits and normalizes AI configuration so commands are logically separated without overlap.

Usage

/normalize-ai-config

Examples

/normalize-ai-config

Use when skills/agents drift, overlap, or routing behavior becomes confusing.

What This Skill Does

1. Inventory
   • Skills (.claude/skills/*/SKILL.md)
   • Agents (.claude/agents/*.md)
   • Hooks (.claude/hooks.json, hookify*.md)
   • Global rules (.claude/rules/*.md)
2. Detect overlaps
   • Multiple skills trying to own the same command intent
   • Test workflows mixing unit/integration/full-gate responsibilities
   • Commit workflows mixing scoped vs full-worktree behavior
3. Normalize boundaries
   • Commit group: /safe-commit, /safe-commit-all
   • Test group: /run-unit-tests, /run-integration-tests, /run-quality-gate, /run-tests
   • Review group: /full-review + /overcautious_check + /review
4. Validate install wiring (dotfiles repo)
   • Ensure install_environment.sh includes every canonical skill in SKILL_DIRS.
5. Report
   • What overlapped
   • What changed
   • Final command map

Output Expectations

• Show overlap findings
• Show normalized command ownership
• Show whether install wiring is complete