1. Software Development

Reviews technical proposals and architectural decisions against complexity dimensions, advocating for the simplest viable approach to avoid unnecessary complexity.

Enables Test-Driven Development for features and bug fixes, following red-green-refactor cycle and avoiding testing anti-patterns.

Enables efficient software development by dispatching independent subagents for each task with code review between iterations to ensure quality and fast delivery.

Requires running verification commands and confirming output before claiming work completion, commit, or PR creation to ensure code quality.

Systematically processes external code review feedback for responding to PR comments, technical suggestions, and reviewer comments in collaborative development.

Systematically identifies code smells and provides targeted refactoring recipes to enhance maintainability during development and code reviews.

Generates structured templates for creating new Claude Code skills, enabling efficient development of custom skill modules.

Systematic debugging framework using four phases (root cause, pattern analysis, hypothesis testing, implementation) to resolve complex bugs and concurrency issues.

Provides on-demand command-line access to MCP servers without configuration changes, enabling quick testing and exploration during development.

Debugs memory leaks, profiles usage, and optimizes allocations in C++ and Python to resolve OOM errors and improve application performance.

Context for modifying Codako game engine's world operator, simulation logic, and rule evaluation code.

Manages game world lifecycle operations including creation, loading, saving, sharing, and fetching via API and editor logic in Codako.

Performs comprehensive code reviews to identify bugs, security vulnerabilities, and adherence to best practices in pull requests and codebases.

Retrieves universal coding patterns from a vector database to assist in complex development tasks and unfamiliar problems, auto-invoking based on context.

Creates event-triggered automation hooks for Claude Code, enabling custom pre/post tool execution logic in the ~/.claude/hooks/ directory.

Enables creation of specialized skills for Claude, extending its capabilities with custom methodologies and workflows.

Runs code quality checks including linting, formatting, type checking, and tests for the lazyopencode project.

Retrieves universal coding patterns from a vector database using two-stage retrieval, auto-invoking for complex tasks or upon user command '--recall'.

Provides a structured nested file organization for test cases, improving test management and execution efficiency in development workflows.

Retrieves universal coding patterns from file-based memory to assist in complex coding tasks and unfamiliar problems, triggered by specific user commands.

Automates code review for correctness, security, regression risks, and test coverage with severity prioritization (P0/P1/P2).

Guides feature implementation from Jira tickets/PRDs using Research-Plan-Implement workflow with quality gates to prevent hallucinations, overengineering, and underengineering.

Conducts background performance audits during development, identifying inefficiencies, memory leaks, and widget rebuilds, and reports critical issues to the main agent.

Facilitates domain modeling through Event Storming sessions to extract, name, and review business events, commands, aggregates, and read models.