claude-config-advisor

Reviews and designs Claude Code project configurations including CLAUDE.md and .claude directories.

733|99|Updated Dec 19, 2025
One-click install
npx skills add https://github.com/chujianyun/skills --skill claude-config-advisor
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: claude-config-advisor
Source: https://github.com/chujianyun/skills/tree/main/skills/review/claude-config-advisor
Command: npx skills add https://github.com/chujianyun/skills --skill claude-config-advisor

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Teams and individuals often struggle to structure Claude Code project configurations: they either over-engineer with unnecessary agents, hooks, and output styles, or cram everything into one bloated CLAUDE.md. This Skill diagnoses existing configurations or plans new ones with clear responsibility boundaries.

Core Features & Use Cases

  • Configuration Review: Audits existing CLAUDE.md, .claude/skills, .claude/agents, .claude/commands, and .claude/hooks for responsibility clarity, maintainability, and over-design, then delivers a graded verdict with prioritized fixes.
  • Structure Design: Recommends a right-sized directory layout (from minimal CLAUDE.md-only setups to team collaboration structures with agents and hooks) based on project scale and goals.
  • Use Case: A developer inherits a project with a 500-line CLAUDE.md and unused .claude directories. The Skill reads the key files, identifies mixed responsibilities and over-design, and outputs a restructure plan splitting workflows into skills and commands.

Quick Start

Review my project's CLAUDE.md and .claude directory and tell me whether the configuration structure is reasonable.

Frequently Asked Questions about claude-config-advisor

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I structure a .claude directory for my project?

Start with CLAUDE.md for global rules, then add .claude/commands for repeated prompts, .claude/skills for specialized workflows, .claude/agents for role-based subagents, and .claude/hooks only for deterministic automation. Add layers incrementally based on actual needs rather than creating all directories upfront.

What should I write in a CLAUDE.md file?

CLAUDE.md should contain project-level rules: common bash commands, build and test entry points, code style constraints, and collaboration rules. Keep it short and executable, avoiding long background stories or full workflow manuals that belong in skills.

When should I use .claude/agents versus .claude/skills?

Use skills to encapsulate specialized knowledge or fixed workflows that trigger in relevant contexts. Use agents when a recurring task needs an independent role boundary, separate context from the main conversation, or different tool permissions and model settings.

What is the difference between CLAUDE.md and CLAUDE.local.md?

CLAUDE.md holds team-shared project rules and is committed to Git, while CLAUDE.local.md holds personal preferences and machine-specific settings and is typically added to .gitignore. Separating them prevents mixing shared standards with individual habits.

Why is my Claude Code configuration considered over-designed?

Over-design happens when a simple project includes many agents, hooks, or output styles without real recurring needs, or when CLAUDE.md duplicates content that belongs in skills or commands. The fix is consolidating to CLAUDE.md plus a few commands and removing low-value directories.