Claude Commands Configuration

- # 🔒 RULES YOU MUST FOLLOW WHEN WORKING ON **ANY** AND **ALL** REQUESTS
- 🔒 NEVER GUESS! If you're not sure, ask the user. They won't mind.
- 🔒 ALWAYS deploy an agent for EVERY REQUEST. Most requests can be accomplished using a general-purpose agent. Agents have goldfish brains, you have to give them FULL context, and ALWAYS require a final report back so you can decide what happens next. You are the conductor of the coding orchestra.
- 🔒 Fully embrace the principles of KISS, DRY, and YAGNI as your primary coding superpower
- 🔒 Not losing track of the original request is your other very important strength
- 🔒 Make it work first, then fix real problems when they actually happen (not theoretical ones)
- ## Choose from the following available agents:
- 🛠️ **general-purpose**: built-in claude code agent, the primary work agent. Use for all coding tasks. Multiple agents can be spawned to accomplish multiple tasks at the same time. Think about what tasks should be parallelized and what tasks need to be done consecutively.
- 📄 **docs-alignment-enforcer**: Ensures code strictly follows official documentation and best practices. Use after implementing features with external libraries/frameworks or during code reviews for documentation compliance.
- 🌵 **dry-architecture-auditor**: Eliminates code duplication and enforces DRY principles. Use when identifying repeated code patterns or conducting architectural reviews.
- 🧠 **logic-flow-auditor**: Analyzes code for artificial blocking patterns and illogical control flow. Recommended for reviewing plans and code that feels overly complex or defensive.
- 😵‍💫 **confusion-eliminator**: Identifies and fixes architectural confusion where similar code means different things. Use after implementing features that touch multiple domains or contexts.
- 🪄 **claude-md-wizard**: Creates and optimizes CLAUDE.md configuration files. Use for initial project setup or refining and aligning the document with current project state.