core-tester

The core-planner skill is a strategic agent that decomposes complex tasks into actionable execution plans with dependency analysis and timeline estimation. It's designed for developers needing structured breakdowns before implementation, working alongside coding and testing skills. Key capabilities include task orchestration, risk assessment, and resource allocation using workspace tools.

The claude-reflection skill enables Claude to autonomously learn from user interactions by detecting corrections, capturing preferences, and extracting patterns. It automatically triggers to persist knowledge across sessions using a Reflect-Abstract-Generalize-Store loop. Developers can manually trigger it with `/claude-reflection` or review captured learnings in `~/.claude/memory/learnings.yaml`.

The core-reviewer skill performs automated code reviews focusing on quality, security, and maintainability. It analyzes code for best practices, performance issues, and documentation quality using tools like security audits and GitHub repository analysis. Use this skill when you need comprehensive code quality assurance before merging or deploying changes.

Core Researcher is a specialized Claude agent for deep technical investigation and information gathering in software projects. It excels at pattern analysis, documentation research, and knowledge synthesis using tools like code analysis, web search, and repository exploration. Use this skill when you need thorough research on dependencies, codebases, or technical topics before implementation.