petekp/claude-code-setup
Overview
This skill audits an existing skill ecosystem and delivers actionable recommendations to improve agent-friendliness, consistency, and maintainability. It inventories global and project-level skills, evaluates individual skill quality, and produces a prioritized report with remediation steps. Use it to get a clear health score and concrete fixes for immediate wins.
How this skill works
The auditor scans skill directories to build an inventory, extracts metadata and size metrics, and detects supporting files like references, scripts, and assets. It runs per-skill checks on metadata quality, content length, progressive disclosure, actionability, and output contracts. Finally, it analyzes the collection for coverage gaps, overlaps, naming consistency, and organizational issues, and generates a prioritized report with grades and remediation steps.
When to use it
- After adding several new skills to check for overlap and consistency
- Quarterly or before major releases to maintain ecosystem health
- When skills feel ineffective or produce inconsistent outputs
- Before sharing or publishing a skill collection to external users
- When you need a prioritized list of concrete improvements
Best practices
- Keep skill descriptions concise and include clear "use when" triggers
- Split long content into reference files and keep manifests short
- Define explicit output contracts so related skills interoperate
- Use consistent naming and trigger patterns across similar skills
- Script repeatable operations to reduce manual steps and drift
Example use cases
- Audit a mixed set of global and project-scoped skills to find duplication
- Generate a health report with grades and a prioritized remediation plan
- Identify skills that are too long and recommend moving details to references
- Standardize trigger phrases and output formats across a domain
- Detect missing automation and recommend scripts for common tasks
FAQ
A quick inventory can complete in minutes; a full audit with per-skill grading and a written report typically takes longer depending on the number of skills, often under an hour for small collections and a few hours for large ones.
Will the auditor change my files?
No—audits are read-only. The output is a report with actionable recommendations; you apply changes manually or via separate automation.
17 skills
This skill generates a focused bootstrap prompt to preserve essential context for continuing work in a new Claude Code session.
This skill applies startup execution wisdom to product, strategy, and business decisions to improve prioritization and go-to-market outcomes.
This skill helps you design and implement world-class SwiftUI interfaces with accessibility, performance, and Apple-like polish across platforms.
This skill guides you through a full manual testing workflow for Claude HUD, including reset, onboarding, core checks, and debugging steps.
This skill helps you fetch and summarize Open Source repository docs with DeepWiki, revealing structure, APIs, and explanations.
This skill guides test-driven development with red-green-refactor loops, enabling you to write integration tests first and drive minimal code changes.
This skill analyzes exported social media data to generate comprehensive personality profiles for AI personalization and self-reflection.
This skill helps you generate temporary debugging interfaces that reveal internal state, transitions, and data flow for complex systems.
This skill audits your skill ecosystem for consistency and best practices, delivering concrete improvements and a ready to act improvement plan.
This skill helps you write clear inline code comments and file headers that explain why, not just what, for humans and AI.
This skill helps you create a concise CLAUDE.md with WHAT-WHY-HOW context to guide Claude across projects.
This skill provides direct, actionable UI/UX critiques focused on clarity, hierarchy, accessibility, and craft to improve design reviews and ship-ready
This skill helps you apply professional typography principles to create readable, balanced interfaces with clear hierarchy and effective measure.
This skill starts development servers with intelligent port management, auto-detects environment, handles conflicts, and opens the browser for a seamless local
This skill helps you integrate Tool UI components into React apps quickly by discovering, installing, scaffolding, and validating runtime wiring.
This skill helps you develop React components with accessible, composable patterns, forwardRef, and consistent APIs for reliable architecture.
Master multi-agent orchestration using Claude Code's TeammateTool and Task system. Use when coordinating multiple agents, running parallel code reviews, creating pipeline workflows with dependencies, building self-organizing task queues, or any task benefiting from divide-and-conquer patterns.
